Court rulemaking, democratic legitimacy, and procedural efficacy, The

process of making process: Court rulemaking, democratic legitimacy, and procedural efficacy, The

Bone, Robert G


The ideal of nationally uniform procedural rules promulgated by the Supreme Court after consideration by expert committees-commonly known as “court rulemaking”-has been the cornerstone of civil rulemaking in the federal courts since adoption of the Rules Enabling Act in 1934.1 Yet today the court rulemaking model is under siege. Many critics question the democratic legitimacy of what they see as a politically unaccountable process, and call for more public participation and an expanded legislative role.2 Others complain about centralization and advocate for broader local control.3 Still others argue against a system of general rules uniformly applied and in favor of rules more closely tailored to distinctions among case types.4 The level of discontent is unprecedented in the sixty-five year history of federal rulemaking in the field of civil procedure.5 The sharp criticism of court rulemaking reflects a profound shift in thinking about procedural law. Many critics today reject the idea that civil process is normatively independent of substance, an idea that justified court rulemaking at its inception and sustained it through much of this century.6 Moreover, the adversary model that guided rulemaking through the 1960s is deeply controversial today, with jurists calling for more use of alternative dispute resolution and greater reliance on case management and settlement.7 There is little agreement about how to measure the quality of procedural rules and considerable doubt about the practicality of making rules to control the complex strategic environment of litigation.

Under the pressure of these changing views, court rulemaking has moved toward a legislative model and away from the traditional model based on reasoned deliberation and expertise. Because procedure has substantive effects and involves controversial value choices, critics argue, rulemaking is “political” and therefore legitimate in a democracy only with broad public participation and accountability.8 At the same time, most of these critics reject the notion that Congress should make all the procedural rules. But no one has yet managed to articulate a coherent middle ground.9

This lack of direction is especially troubling in light of the strong interest in procedural reform today. Prompted by concerns about case backlog, litigation delay, and high litigation costs, the Advisory Committee on Civil Rules has been quite active in federal rule reform, and Congress has been keen to intervene in the rulemaking process.lo These developments have created conflict between the federal judiciary and Congress over the proper allocation of power to regulate procedure. Without a coherent view of the limits on court rulemaking, these conflicts are difficult to resolve, and the resulting tension has already strained relations between the two branches.11

What we need is a theory of procedural rulemaking that explains what the court and the legislature each have to contribute to the process. This article aims to provide such a theory. It argues that a centralized, court-based, and committeecentered process is well suited for making general constitutive rules that define the basic framework of a civil procedure system and more detailed rules that control particularly costly forms of strategic behavior. More generally, it defends a view of court rulemaking that sees its central function as developing and maintaining a system of rules that reflects the best principled account of procedural practice.

While in theory Congress has broad power over procedure, the view of court rulemaking that I develop in this article counsels in favor of congressional restraint.12 Court rulemaking is better suited than legislation to the task of inferring general principles from existing practice and designing an integrated system of rules based on those principles. Moreover, the legitimacy of the court rulemaking process does not derive from public participation or political accountability, but instead from a model of principled deliberation akin to common law reasoning. Thus, there is good reason for Congress to defer to court rulemaking. Congressional oversight is not necessary to legitimacy, and congressional intervention can easily distort the principled coherence of the rule system as a whole. As a result, Congress should intervene only if there is strong reason to believe that the current system is in need of radical overhaul or that tailoring a specific procedural rule to a particular statutory scheme promises substantial benefits. In either case, however, Congress should take account of the institutional advantages of court rulemaking and seriously consider the costs to the integrated system of rules court rulemakers have made.

Finally, although my main focus is on the relationship between court rulemaking and Congress, I also briefly discuss other rulemaking alternatives.’3 In particular, I argue that local districts should have only a limited role in rulemaking, and that trial judge discretion to adapt procedure to individual cases should be more constrained and more strongly guided by general principle than it is today.

This article is divided into six parts. Part I describes the current court rulemaking model for the federal courts. Part II reviews the history of federal court rulemaking. It describes the ideas and beliefs that fueled enthusiasm for court rulemaking at the beginning of this century and the romance with the court rulemaking model during the 1950s and 1960s. It then recounts the events that contributed to the rise of procedural skepticism during the 1970s and traces the impact of this skepticism on views about the legitimacy and efficacy of court rulemaking during the 1980s and 1990s.

This history sets the stage for the prescriptive project in Parts III through VI. Part III organizes the various critiques of court rulemaking into an account of the modern justificatory dilemma. This dilemma stems from the recognition that procedure has different substantive effects on different classes of litigants and that procedural rules can be justified only by reference to substantive value. This insight undermines the traditional rationale for court rulemaking based on technical expertise. Moreover, its troubling implications cannot be easily avoided through a simple analogy to administrative agencies-at least insofar as agency legitimacy depends on substantive delegation or political oversight. Congress has expressly withheld substantive lawmaking power from court rulemakers, and court rulemaking is not subject to executive control. In addition, the possibility of congressional oversight cannot justify the kind of rulemaking process that I defend in this article, one that operates largely independent of Congress and that derives its legitimacy outside the realm of politics.

Part III also critically examines current attempts to cope with the justificatory dilemma and to defend a legitimate place for court rulemaking distinct from legislation. It surveys arguments based on procedural expertise, empiricism, consensus, and trial judge discretion, and concludes that none is fully adequate to the task.

Part IV then turns to constructing a more convincing rationale. It analyzes the core elements of the court rulemaking model-nationally uniform rules made by the court through a centralized rulemaking process relying on expert committees-in light of three different theories of procedure: an efficiency theory, a rights-based theory, and a process-based theory. Part IV argues that court rulemaking is an attractive way to make good rules under each theory, and ultimately justifies court rulemaking as a vehicle for designing a system of rules consistent with the best principled account of existing practice.

Part V addresses an objection based on process-oriented legitimacy. According to this objection, court rulemaking is incompatible with democratic values even if it produces good rules, because it gives a politically insulated body, the United States Supreme Court, the power to make law that can significantly affect the implementation of substantive values and the distribution of social power. Part V argues in response that procedural rules can have substantive effects and their justifications implicate substantive values without requiring a politically accountable process. Finally, Part VI discusses the implications of this analysis for the limits of court rulemaking and proposes a test for allocating power between the Court and Congress.


Civil rulemaking for the federal courts illustrates the core features of the court rulemaking model-centralized rulemaking by courts relying on the expertise of committees and aimed at creating uniform, general, and systemically integrated rules. The Supreme Court has the power to “prescribe general rules of practice and procedure” for the federal courts,14 and the Judicial Conference of the United States has the authority to recommend rule changes to the Supreme Court.15 The Judicial Conference in turn oversees a committee structure that includes a Standing Committee on Rules of Practice and Procedure appointed by the Chief Justice of the Supreme Court’6 and various advisory committees accountable to the Standing Committee.’7 The Advisory Committee on Civil Rules, which consists of judges, lawyers, and legal academics, is responsible for the Federal Rules of Civil Procedure.

There are several stages in the rulemaking process. A proposed rule is first considered by the Advisory Committee. If the Advisory Committee approves the proposal,’8 it is then reviewed by the Standing Committee and finally by the Judicial Conference before being forwarded to the Supreme Court.19 If the Supreme Court concurs, the proposal is transmitted to Congress, which then has roughly seven months to exercise a veto.2 In the absence of a veto, the proposed rule goes into effect.

As even this brief summary indicates, the ideal of centralized and uniform court-made rules is qualified to some extent. For one thing, judicial power to make the Federal Rules of Civil Procedure is not exclusive; Congress also has some power through its veto option under the Rules Enabling Act. However, the assumption at the time of the Act’s adoption in 1934 was that Congress would use its veto power only sparingly, and from 1934 until 1973, Congress never vetoed a rule.2′

Furthermore, although centralization and national uniformity are key ingredients of the traditional model, federal districts have always retained some power to adopt local rules.22 This power, however, is limited. Local rules are supposed to be subordinate to and consistent with centrally adopted national rules.23 Although district courts have not always respected these limits24-and in recent decades have engaged in extensive local rulemaking25-there is no question that the traditional court rulemaking model contemplated a highly centralized process producing nationally uniform rules and that centralization and uniformity still have strong supporters today.26



The campaign for the court rulemaking model on the federal level began in 1906 with Roscoe Pound’s famous address to the American Bar Association, The Causes of Popular Dissatisfaction with the Administration of Justice.27 Pound’s address focused criticism on the technicality and formality of the code and common law systems and inspired the American Bar Association’s (ABA) lobbying effort to reform the rules.28 That effort culminated twenty-eight years later, in 1934, with congressional adoption of the Rules Enabling Act.29

The United States Supreme Court promptly exercised its power under the Rules Enabling Act and appointed an advisory committee to draft the Federal Rules of Civil Procedure (“Federal Rules”). The committee met for two years and completed its task in 1937.30 The new Federal Rules went into effect in 1938 with much fanfare and celebration.31

Between 1906 and 1938, many judges, lawyers, and legal academics wrote about the advantages of court rulemaking. The resulting literature is vast and the arguments varied.32 Yet advocates seem to have shared a common set of assumptions about procedure and substance, assumptions that supported their faith in court rulemaking as the ideal way to design procedural rules.

In brief, proponents of court rulemaking believed that procedure was normatively distinct from and subordinate to substantive law. Procedure was, in their view, merely a means to the end of finding the facts and applying the substantive law accurately.33 To the early twentieth-century reformers, this insight meant that the design of a procedural system was mainly a technical exercise in perfecting administrative machinery.34 The idea was to streamline the procedural machine so that it enforced the substantive law (whatever that law might be) with a minimum of friction and waste.35 Thus, the values relevant to procedural rulemaking were not substantive in nature.36 They were practical values of administrative design, such as efficiency (understood narrowly as minimizing administrative cost),37 simplicity, and flexibility.38

Edson Sunderland, a Professor at the University of Michigan Law School, an active member of the original advisory committee, and one of the leading procedure scholars of his day,39 summarized the prevailing view in the following succinct passage:

Now the law consists of two distinct and almost independent sets of rules and principles, one making up the field of so-called substantive law, the other the field of procedure. The first group is primary and constitutes an essential part of the structure of society; the second is secondary and derivative, and merely serves to make the first operative. The first group is relatively fixed, and only changes with the slow evolution of social relations; the second is relatively flexible, having no universal quality, but being the mere manifestation of opportunist ingenuity. To radically revise the first would mean a social revolution, but the second could be totally reorganized at a moment’s notice without causing a tremor in the social structure.40

From these premises, it followed that a democratic process was not necessary to the legitimacy of procedure, for procedure involved no substantive value choices. In fact, to some commentators, separation-of-powers principles actually dictated that procedure should be left to the control of courts.41 The general view was that courts were much better suited to the rulemaking task. Judges were experts in the science of procedure and had first-hand experience with rules in practice.42 Furthermore, they were immune from the pressures, special interests, and vagaries of the political process.43 As a result, courts were more likely than legislatures to approach the rulemaking task rationally and with due regard for the systemic integrity of the procedural system as a whole. Also, it made sense to use committees because of time and resource constraints and because a committee process was ideally suited to tapping the expertise of scholars and lawyers as well as judges.44

This view of procedural rulemaking as an exercise in instrumental rationality divorced from substantive value may seem alien to many proceduralists today, accustomed as we are to focusing on the substantive effects of procedural rules. The distinction made sense to early twentieth-century reformers partly because they never seriously questioned the basic features of adversarial adjudication, such as party control, oral presentation, and cross-examination of witnesses.45 Reform meant stripping away the technical rules of the code and common law systems and retrieving-and then perfecting-the core of the adversary system embedded in existing practice.46 Indeed, leading federal rule proponents assumed the integrity of adversarial process when they justified even the most controversial rule reforms, such as liberalized pleading, expanded joinder, broader discovery, and summary judgment.47


Enthusiasm for court rulemaking and the Federal Rules spread rapidly after World War II. Many states followed the federal lead in giving the rulemaking function to the courts.48 Additionally, first year civil procedure courses switched their focus from the practical details of local state practice to the general theory and content of the Federal Rules, and law schools added upper level courses on federal jurisdiction and federal courts to the curriculum.49 Procedure scholars debated the pros and cons of court rulemaking, the general philosophy behind the Federal Rules model, and the advantages of specific Federal Rules.50 The 1950s and 1960s were, in short, the golden age of court rulemaking.51

There are a number of reasons for these developments. Civil procedure professors, especially those at leading law schools, probably appreciated the opportunity to teach and write about their subject on a more theoretical level. Moreover, the fact that court rulemaking and the Federal Rules had the backing of prominent procedure scholars as well as the ABA elite no doubt added prestige to the reform movement. And as more students learned about the virtues of court rulemaking and the Federal Rules in law school, there would have been more lawyers willing to support the reforms in practice.

It is also important to note that the basic tenets of the court rulemaking model fit the most influential ideas about law and procedure during the 1950s and 1960s. These ideas, known collectively as Legal Process, stressed the importance of sound institutional design to good lawmaking.52 As I have described in some detail elsewhere,53 Legal Process theory assumed that each institution (for example, legislation, adjudication, administration) was organized around a core of general principles that defined its special competency and prescribed its ideal structure.54 In the case of adjudication, this core consisted of the basic elements of the adversary system: party presentation of facts and argument in an adversarial setting.55 Designing the more detailed features of an institution like adjudication was a pragmatic and incremental process of using reason to reflect critically on practice in light of the institution’s defining principles.56

Since procedural rulemaking involved the design of adjudication, it made sense on this view to construct a rulemaking process around reasoned deliberation by legal experts familiar with litigation practice.57 It also made sense to adopt rules, like the Federal Rules of Civil Procedure, that incorporated the basic elements of the adversary system. Moreover, the fact that the task involved reasoned deliberation rather than interest accommodation helped to support the legitimacy of the process despite the absence of political accountability.

Even so, legitimacy concerns occasionally arose in debates about the proper role of the legislature in the rulemaking process.58 Even supporters of court rulemaking at times conceded that procedure and substance were “inextricably intertwined,” that ostensibly procedural rules sometimes involved policy choices, and that such choices had important substantive effects.59 Nevertheless, the legitimacy critique took a relatively mild form during this period. Most commentators agreed that courts should have primary responsibility for rulemaking; they disagreed about whether the legislature should have any control over the process at all.6 Not until the 1970s did commentators begin to realize the full implications of the legitimacy critique.


1. The Rise of Procedural Skepticism

The professional romance with court rulemaking and the Federal Rules began to sour in the early 1970s. Critics attacked the notion that there was an ideal procedure embedded in existing practice and codified in the Federal Rules. As a result, the boundary between procedure and substance blurred, and the case for expert rulemaking weakened.

During the 1960s and early 1970s, new substantive rights were created in response to growing public concern about civil rights, consumer welfare, and environmental protection.6′ At the same time, public interest groups and lawyers inspired by the successes of the civil rights movement began to view litigation as a vehicle for social reform. The resulting changes in the character of federal litigation gave rise to concerns about the adequacy of the existing procedural system to promote substantive values.

Many of the new public interest cases, for example, pitted plaintiffs of poor or modest means against more powerful government and corporate defendants.62 These cases raised questions about court access and the fair distribution of litigating power in the context of procedural issues, such as class actions,63 joinder,64 and pleading specificity.65 Moreover, as complex injunctions became more common, judges experimented with procedural innovations at the remedial stage-including expanded intervention, use of court monitors, and greater reliance on negotiation and settlement-and justified those measures as necessary to enforce the substantive values at stake.66 These developments challenged the belief in a unitary procedural ideal and undermined the hegemony of the traditional Federal Rule model.67 Most important, they focused attention on the close relationship between procedure and substantive law.

Heightened sensitivity to the procedure-substance connection manifested itself in other ways as well. Some critics claimed that procedural rulings on issues such as standing were in fact disguised substantive decisions and that judges used value-neutral procedural rhetoric to conceal hostility to substantive rights.68 Moreover, outside of civil procedure, many scholars challenged processbased constitutional theories like that of John Hart Ely in Democracy and Distrust69 on the ground that decisions about process always implicated substantive value choices.70 And work by political theorists during the 1970s and 1980s showed that procedural rules could determine substantive outcomes.71

During this same period, lawyers and judges began to worry about the practical functioning of the federal system. Starting in the early 1970s, concerns about case backlog, lengthy delays, and high litigation costs grew more frequent and more intense.72 These concerns focused attention on alternatives to adversarial process and raised doubts about the desirability of the most basic elements of the Federal Rules, including notice pleading and broad discovery.73

In the midst of these developments, the Supreme Court weighed in with a procedural due process opinion, Mathews v. Eldridge,74 that rejected a unitary procedural ideal and embraced a pluralistic approach to procedure design. The Mathews Court held that the Due Process Clause did not require an oral evidentiary hearing before termination of social security disability benefits and that a procedure based on written submissions was constitutionally sufficient.75 The Court endorsed a cost-benefit balancing test that weighed the reduction in error risk made possible by the additional procedures (in this case, oral testimony and cross-examination), the nature of the private interest affected by an erroneous decision, and the cost to the government and the public of providing greater procedural protection.76 Although the case itself involved administrative process, the Court’s reasoning had much broader implications, for it suggested that no particular kind of procedure was necessarily ideal in all circumstances and that the adversarial process was not always optimal.77 Moreover, by including the recipient’s substantive interests in the balance, the Court signaled just how tenuous the distinction between procedure and substance really was.

Thus, assaults on the ideology of court rulemaking and the Federal Rules took place on three different fronts. On the theoretical front, challenges created by new substantive rights undermined belief in an ideal procedure and a clear normative divide between procedure and substance. On the practical front, complaints about the quantity and quality of federal litigation shook confidence in the adversary system. And on the doctrinal front, the Supreme Court’s procedural due process jurisprudence blurred the distinction between procedure and substance.

Another development, which came from outside the procedure and constitutional law fields, had a profound effect as well. From the early 1900s through the 1960s, courts and commentators had relied on arguments of technical expertise and scientific objectivity to support the efficacy of administrative agency decisionmaking and to reconcile agency discretion with democratic values.78 By the 1970s, however, concerns about regulatory capture, a loss of faith in expertise, and a growing awareness that administrative regulation involved social policy choices had eroded public confidence in administrative agencies.79 These same factors undermined confidence in the efficacy and legitimacy of the traditional court rulemaking model as well.

2. The Decline of the Court Rulemaking Model

One can see the impact of these developments in the checkered history of civil rulemaking for the federal courts over the past twenty-five years. Today, critics of court rulemaking are more numerous and much more influential. They question whether a rulemaking process dominated by federal judges can possibly produce good rules which treat all litigants fairly, and they challenge the legitimacy of a process that does not include representation by all affected constituencies. The result is a shift away from a model based on expertise and toward one based on representation and accommodation of competing interest groups.

The modern story begins in 1973. In that year, Congress for the first time exercised its veto power under the Rules Enabling Act by blocking the proposed Federal Rules of Evidence and enacting a revised version in statutory This event opened the door to greater political involvement in the rulemaking process.81

Since 1973, various interest groups-plaintiffs’ bar, defendants’ bar, civil rights groups, and corporate groups-have become more active (some would say aggressive) at all stages of rulemaking, from Advisory Committee hearings to congressional review.82 In 1983, for example, interest group lobbying blocked enactment of proposed changes to the federal service-of-process rule (Rule 4)83 and cut short Advisory Committee consideration of amendments to the offer-ofjudgment rule (Rule 68).84

In 1988, Congress made substantial revisions to the Rules Enabling Act for the first time in thirty These changes opened the rulemaking process to broad public participation by requiring public hearings, open meetings, publicly available minutes, and longer periods for public commentary.86 Moreover, at least since the mid-1980s, critics have advocated changing the membership of the Advisory Committee to include more lawyers as well as more representatives from groups that use the federal courts.87

This pressure to increase public involvement followed on the heels of sharp criticism during the 1970s and 1980s that the rulemaking process lacked public accountability and that more input from lawyers and other interested parties would produce better rules and bolster By the mid-1970s and continuing into the 1980s, public accountability through enhanced participation and political oversight rapidly overshadowed technical expertise as the primary model of legitimacy for administrative agencies.89 Court rulemaking responded to the same pressures: concerns about efficacy and legitimacy pushed the process toward an interest group model that assimilates rulemaking to legislation.90

Perhaps the most serious blow to the court rulemaking model took place with the adoption of the Civil Justice Reform Act of 1990 (CJRA).9′ The purpose of the CJRA was to address problems of case backlog, delay, and cost, and to improve the efficiency of the federal court system.92 The CJRA required federal district courts to develop case management plans tailored to the particular circumstances of their local districts.93 It also created a novel process for making rules: each district court appointed a local advisory group broadly representative of the different users of the court system;94 and each advisory group studied the local problems and recommended reforms to be enacted as local rules.95

The CJRA remained in effect for seven years, and during that period all federal districts adopted case management plans.96 The result is a hodgepodge of different procedural rules varying from district to district, with many of these rules representing major departures from the original Federal Rules.97 This may be only a temporary phenomenon, as the CJRA contemplates national rulemaking based on experience with case management plans,98 but many commentators still worry about the radical departure from the original ideal of uniform federal rules and the disturbing trend toward “balkanization” and “fragmentation” of federal civil procedure.99

The CJRA’s underlying philosophy poses a more serious challenge to the court rulemaking model. The CJRA is based on a “bottom up” approach to rulemaking. o The idea is that good procedural rules grow out of the particular circumstances of local practice communities and the needs of user groups, and that legitimacy can be achieved only if rulemaking is responsive to multiple points of view.’o’ Although the notion of anchoring procedural reform in actual practice would have been congenial to the original Federal Rules drafters, the idea of fitting rules to local circumstances and involving user groups centrally in the drafting process would have made little sense.

More disturbing still is Congress’s position that significant procedural reform should be controlled by the legislature.’02 In particular, CJRA proponents defended congressional involvement by arguing that CJRA reforms were so intimately tied to substantive policies and effects that public participation and congressional control were essential.l03 This argument employed a very broad definition of “substance,” as anything that increases public access to the federal courts or improves the efficiency of American business.’4 Although it may be unwise to read too much into the more extreme rhetoric,”5 it is certainly significant that CJRA supporters in Congress thought this argument sufficiently persuasive to make it publicly. Moreover, the logical extension of the argument is clear enough; it implies the virtual elimination of court rulemaking altogether.06

Congressional involvement and interest group lobbying did not stop with the CJRA. In 1993, for example, the Supreme Court transmitted proposed discovery rule amendments to Congress.107 The ABA and other interest groups opposed the reforms and launched a major political battle that almost succeeded in securing a congressional veto.los Two years later, Congress again entered the procedural arena. As part of the Republican Party’s “Contract With America,” the House considered several bills intended to make substantial changes in federal procedure, including the Attorney Accountability Act, the Common Sense Legal Reforms Act, and the Private Securities Litigation Reform Act.l09 These measures generated intense political controversy and heated public debate. As a result of these developments, the Advisory Committee has become keenly sensitive to the risk of congressional interference in the rulemaking process.110

Along with the increased politicization of rulemaking, the character of federal court adjudication has changed markedly over the past twenty-five years. These changes have undermined the traditional notion of court rulemaking as a process for implementing ideal norms of adversarialism. Federal judges today emphasize pretrial management and encourage settlement much more than ever before.l’l Also, disenchantment with adversarial process has spurred a strong movement for court-annexed arbitration and other forms of alternative dispute resolution.112


Thus, concerns about legitimacy and efficacy lie at the heart of modern discontent with court rulemaking. The traditional belief in a procedural ideal and a clear normative divide between procedure and substance is no longer persuasive. So long as lawyers and judges believed that rulemaking involved the perfection of an adversarial ideal, it was possible to conceive of the enterprise as divorced from substance. But if the adversarial system is not necessarily ideal, then procedural design is a matter of policy, and the relevant policies reflect substantive values-or so the argument goes. To be sure, few critics are willing to jettison the idea of judge-made rules altogether, but without a solid justification to anchor it, the court rulemaking model is vulnerable to congressional encroachment, interest group involvement, and strident calls for expanded participation.

The Supreme Court long ago recognized the constitutionality of court rulemaking under the Rules Enabling Act,113 and I do not mean to suggest that this constitutional holding is currently up for grabs. The challenge from legitimacy and efficacy operates at a lower policy level. Critics use these arguments to demand changes in rulemaking procedures, to object to particular rules, and to call for greater congressional involvement. The analysis in this article is targeted at a similar level.

As we have seen, the prevailing view of administrative agencies has also shifted from a focus on technical expertise to a more frank recognition of the contested policy stakes,114 and there is a rich literature analyzing legitimacy issues in the agency context. But court rulemaking differs from agency rulemaking in important respects. For one thing, Congress empowers administrative agencies to make substantive law, but it denies that power to court Moreover, most agencies are subject to political checks through the executive branch and oversight by Congress.116 By contrast, court rulemakers are not subject to executive control, and the traditional model (and the model I defend in this article) assumes that Congress will exercise its veto power under the Rules Enabling Act only rarely. Even judicial review does not offer the same kind of constraint as it does in the agency context, for successful challenges to federal rules are extremely rare and Supreme Court review relatively cursory.117

I do not mean to suggest that there are no helpful agency analogies. For example, the United States Sentencing Commission, an independent agency created by Congress to promulgate sentencing guidelines for federal criminal cases, is a useful analogy,”8 but it is hardly dispositive of all policy concerns in the court rulemaking context. Unlike the committees involved in court rulemaking, the Sentencing Commission operates pursuant to an explicit grant of substantive lawmaking power, and is also subject to some executive oversight (through the President’s power of appointment and removal for good cause). 19

Nor do I mean to suggest that no theories of agency legitimacy are relevant to court rulemaking. Those theories that privilege delegation or political and judicial oversight are not particularly helpful, nor are those that focus on public participation (since I defend a view of legitimacy that does not depend on participation). However, some commentators locate legitimacy in what they believe is the superior capacity of agencies to engage in the kind of deliberative process that advances the public interest.120 The argument I make for court rulemaking also focuses on reasoned deliberation, but does so in a way that fits the special nature of the court rulemaking task.

Thus, court rulemaking is sufficiently different from agency rulemaking to complicate the justificatory problem. The goal of my argument in this articleand the goal of the traditional model-is to define a domain for court rulemaking that is largely independent of Congress and that does not require political or public accountability to support its legitimacy.


To build a case for court rulemaking, however, we must first be clear about the precise nature of the connection between substance and procedure that gives rise to the modern justificatory dilemma. Critics focus on two different kinds of connection: those that rely on substantive effects, and those that rely on substantive justification.

1. Substantive Effects

Most critics focus on the substantive effects of procedural rules.lzl At first glance, this connection might seem obvious: if procedure affects outcome and outcome is substantive, then procedure affects substance. But if this syllogism were the entire critique, there would be no reason for concern. Even the early twentieth-century reformers understood that procedure affects outcome. Indeed, if outcome effects alone were sufficient to trigger concern, then all state decisions would have to be made by processes styled on legislation, even those decisions belonging to the common law.

The argument therefore must depend on more than mere outcome effect. In fact, most critics focus on the distribution of outcomes and, in particular, on distributions that systematically disadvantage particular classes of litigants.l22 Because compliance with a procedural requirement is always costly, the argument goes, the existence of wealth inequality means that some parties are less able to litigate than others and thus less able to vindicate their legitimate claims. If, given these wealth effects, a procedural rule is likely to create an outcome distribution that seriously and systematically disadvantages certain parties, such as the poor, women, or minorities, then rulemakers should at least hear from those adversely affected and perhaps even include them in the rulemaking body.

Unfortunately, proponents of this argument do not explain why the existence of wealth-contingent outcome effects should require public participation or mandate a rulemaking process that is politically accountable. One can concede the premise of outcome inequality without also accepting the conclusion that court rulemaking has questionable validity.l23 All social decisions are made against a background of wealth inequality, yet we do not believe that all such decisions must be made by a politically accountable process.124 I explore these issues and dispute the contention that wealth-contingent outcome effects impugn rulemaking legitimacy in Part V below.

2. Substantive Justification

The second type of substance-procedure connection focuses on justification rather than effects. On this view, it is not simply the fact that a rule has differential effects that matters, but rather the nature of the justification for those effects. The argument supposes that any such justification requires weighing substantive values and that this sort of weighing should be done by a more politically accountable process.

We should not confuse this argument with the much weaker one that supposes rulemakers are actually motivated by substantive bias.125 The argument from substantive justification is not that rulemakers are biased, but rather that they have no choice but to weigh substantive values. There is simply no other way to design a sound procedural system.

The basic premise of the argument is sound. Procedural justification does involve weighing substantive values. To illustrate, suppose that a rulemaker must choose between a notice pleading rule (Rule A) and a strict pleading rule (Rule B).126 Our hypothetical rulemaker should be concerned about outcome accuracy and thus should compare the risk of error under these two rules.l27 At this point, however, she faces a serious complication because error comes in two different forms: false positives and false negatives. For example, Rule A reduces the risk of an erroneous denial of relief-a false negative-by making it easier for meritorious cases to be brought.’28 But it also increases the risk of an erroneous grant of relief-a false positive-by making it easier for frivolous suits to be filed. Rule B has precisely the reverse effect. It reduces the risk of false positives by increasing the filing burden for frivolous suits, but it also increases the risk of false negatives by making filing harder in meritorious suits.

The only way our rulemaker can choose between these two rules is to compare the relative importance and relative frequency of the two kinds of error.129 But-and this is the crucial point-the relative importance of the two kinds of error depends on the seriousness of the harms each error creates, and those harms involve injuries to substantive interests. Thus, our hypothetical rulemaker must assess the value of the substantive interests at stake. To choose between notice pleading and strict pleading, for example, she must compare the seriousness of preventing a meritorious suit with the seriousness of allowing a frivolous suit. How serious is the harm when meritorious plaintiffs cannot recover (or even have an opportunity to recover)? The answer depends on the values underlying the substantive law, for an inability to recover is harmful only insofar as it frustrates the realization of those values. How serious is the harm caused by a frivolous suit? Frivolous suits consume litigation resources, cause reputation harms, and produce unjustified settlements.l30 Reputation harm is significant for its effect on an individual’s substantive welfare, and unjustified settlements create problems insofar as they affect substantive interests. Measuring these harms, therefore, requires consideration of substantive values. Even the seriousness of wasted litigation costs depends, explicitly or implicitly, on alternative uses to which the wasted resources might be put, and these alternatives implicate substantive interests.131

Our hypothetical rulemaker faces another complication, which does not depend on the distinction between false positives and false negatives. Rule A and Rule B will likely generate different process costs. A notice pleading rule, for example, invites more litigation than a strict pleading rule.l32 The same is true of a broad versus narrow discovery rule or a generous versus strict amendment rule. Our rulemaker must compare these process costs with the error costs created by the proposed rule. But to do this, she must have a way to measure the seriousness of error, and seriousness of error necessarily depends on substantive values.

To illustrate, suppose that most suits are meritorious and that plaintiffs frequently have difficulty obtaining information about liability before bringing suit. Our rulemaker might reasonably estimate that under these circumstances a notice pleading rule would generate fewer total errors than a strict pleading rule, but that it would also generate higher litigation costs. To determine whether the reduction in error justifies the increase in cost, the rulemaker must have some metric to measure error cost, and any such metric must refer to substantive values. The same reasoning applies to choice of discovery rules, amendment rules, summary judgment rules, and any other rule that significantly affects the likelihood of success or the cost of litigating a suit.

One would have little cause for concern if the values to be assigned to the various interests were obvious from inspecting the substantive law. But this is clearly not the case. Assigning values to substantive interests is both difficult and controversial. Reasonable people disagree, for example, about the relative importance of the different interests protected by the Constitution.

This point about substantive values is perfectly general. The reasoning depends only on two relatively obvious propositions: first, that choice of procedural rules should turn, at least in part, on outcome accuracy; and second, that accuracy is valuable not as an end in itself but as a means to protect substantive interests. In particular, the argument does not depend on any specific standard for measuring the accuracy of adjudicative outcomes. Whether adjudication aims to maximize efficiency or protect rights, it is still necessary to make substantive judgments about error cost, although in the case of rights the cost would be measured in moral terms.133

One cannot avoid the implications of this argument by positing general principles of procedure, such as that decisions should be on the merits insofar as possible or that a procedural system should fairly apportion burdens between parties.’34 A principle that decisions should be on the merits is just another way of saying that procedure should yield accurate outcomes, and as we have seen, application of that principle requires attention to substance. Similarly, a principle of fair apportionment begs the question of what is “fair” and answering that question requires reference to substantive values.l35

Nor can one carve out a substance-free domain through the idea of balancing “process values.”’36 Depending on who makes the list, process values can include such things as legitimacy, efficiency, fairness, accuracy, finality, and participation.’37 The process value approach lacks sufficient coherence to be analytically useful. For example, accuracy, finality, and participation are not values in and of themselves. They are just factual states of affairs-a certain probability of error, the termination of a process, or the opportunity to be involved in a case. They have value only insofar as they are given value by some normative view of procedure, and as we have seen, any such view, to the extent it focuses on outcomes, requires reference to substantive value.’38 To be sure, some procedures, such as participation, may also have dignitary value unrelated to outcome quality. I discuss this “process-based” approach in Part IV. But any litigation system must place primary weight on achieving good outcomes and therefore must take account of substantive values.139

There is no denying therefore that procedure is connected to substance through effects and through justification. To what extent this poses a challenge to court rulemaking remains to be seen in Parts IV and V below. One thing is clear, however. This close relationship means that concerns about efficacy and legitimacy must be taken seriously. For those who would support court rulemaking, this presents a difficult justificatory dilemma.


In addition to sometimes urging broader public participation and accountability, proponents of court rulemaking employ four different strategies to resolve the modern dilemma. The first relies on procedural expertise, the second on empiricism, the third on consensus, and the fourth on trial judge discretion. None is fully equal to the task.

1. Procedural Expertise

The first strategy employs a version of the traditional argument from expertise.l40 Proponents of this approach suppose that a domain of pure procedure exists that is distinct from substance and that involves complex technical issues suitable for procedural experts.’41 They argue that the problems plaguing court rulemaking today are largely the fault of overzealous rulemakers who take on issues, such as the settlement class action, that exceed the bounds of rulemaking competence and invite political controversy.l42 The solution is for rulemaking to return to the purely procedural matters that define its proper domain.

The justificatory dilemma is not so easily resolved, however. It is simply not possible to separate out a domain of purely technical procedure divorced from substance in the way this argument requires. While some procedural issues are relatively uncontroversial and yield mainly to a technical analysis, this is certainly not the case for most of the important features of a procedural system. As we have seen, designing even the basic elements requires attention to substantive value and in a way that can easily invite interest group controversy.l43 Indeed, given that most significant procedural reforms today are likely to be controversial, any approach that seeks to avoid controversy risks eviscerating rulemaking in the name of saving it from politics.

2. Empiricism

The second strategy focuses on improving the empirical basis of rulemaking.l44 On this view, the rulemaking crisis today is largely a result of empirical ignorance rather than committee overreaching. If rulemakers had a better understanding of the effect of procedural rules, the argument goes, disputing factions would find it more difficult to make speculative claims and rulemakers could more easily go about the business of assessing costs and benefits in a reasonably objective way.l45

Certainly, better empirical work can produce better rules, and hard data is sometimes an effective antidote to extreme factual claims.l46 Yet there are serious methodological and funding obstacles to conducting rigorous empirical studies of the litigation system,l47 and the results of those studies are often subject to interpretation depending on one’s point of view.l48 Furthermore, the empirical approach does little to defend rulemaking against legislative intrusion. Congress, after all, is better equipped to fund, gather, and process empirical information.

More important, the empirical approach understates the significance of normative conflict. Evaluating procedure’s costs and benefits is not simply an empirical matter. It involves applying values-frequently substantive values-to the facts. And disagreement about values is responsible for much of the rulemaking controversy today.149

3. Consensus

The third strategy, which seems to be the most popular today with commentators and the Advisory Committee itself, relies on consensus to resolve normative conflict.’50 Because no objective way exists to choose among competing values, consensus proponents argue, rulemaking must act with the input and the consent of all competing interests if the process is to produce good rules, operate legitimately, and avoid political controversy.

Consensus, however, is a problematic basis for rulemaking. In fact, consensus rulemaking in the administrative agency context is quite controversial and so far has received mixed reviews.’5′ Consensus in the civil procedure field may be feasible for some procedural rules (although even then it might be difficult to ensure that all interested groups consent), but these rules are likely to implicate only technical issues with slight substantive impact. Moreover, insofar as consensus involves accommodating conflicting interests, it aligns court rulemaking so closely with legislation that it is unclear why the rulemaking task should be given to the courts at all.’52 The legislature is the institution in our democracy that is designed to accommodate interests, and the legislature may be more effective in doing so since it is not limited by a consensus requirement. Indeed, it is even doubtful that satisfying user preferences should replace values such as accuracy when rules are made outside of the legislature.

In addition, a search for consensus can paralyze the rulemaking process.’53 The high level of contentiousness over the past twenty years, even with respect to such core procedural matters as pleading and discovery, bodes poorly for the likelihood of reaching consensus on important procedural issues.154 Furthermore, if consensus is to confer legitimacy, the rulemaking process must represent all constituencies. But then negotiations can become unwieldly.’55

Finally, even when compromise is possible, it is likely to coalesce around highly general rules that leave the difficult normative issues unresolved.’56 Rules of this sort can be attractive to all sides of a controversy because they give everyone a chance to wage the battles later in the context of individual suits.157 This result, however, strips rulemaking of any real purpose. Moreover, for reasons discussed in Part IV below, trial judges are not likely to make good procedure in individual cases without the constraint and guidance of centrally promulgated rules.158

4. Trial Judge Discretion

The fourth approach to the problems of efficacy and legitimacy relies on the case management model and its core feature, trial judge discretion.’59 According to this model, procedural rules should serve primarily to empower casespecific discretion by furnishing a wide range of case management options. The appeal of this approach is that adding more options seems neither to limit choice nor to require resolution of controversial normative issues at the rulemaking stage. 160

The management model, however, is plagued by many of the same problems that beset the consensus approach. It provides no reason to prefer courts over the legislature, especially given that the legislature may be better able to evaluate the different options. Moreover, the model lacks a convincing explanation of why trial judges are likely to make better procedure than a centralized rulemaking committee. To be sure, trial judges are better informed about individual cases, but vesting broad (and essentially unreviewable) discretion in the trial judge undermines predictability and consistency and raises its own legitimacy concerns.

To defend case management and judicial discretion as the answer, one must explain why trial judges are able to use contextual information to make socially optimal procedures in an ad hoc, case-specific way. Perhaps they have a special talent for harmonizing competing values through a pragmatic and largely intuitive process of figuring out what works in the concrete setting of individual cases. This idea is superficially appealing, 1 but ultimately unpersuasive. Deciding on the best procedure cannot be simply an intuitive process of figuring out what “works,” because resolution of value questions is necessary to define what it means for a procedure to “work.” 162 Moreover, to manage the costs of strategic interaction effectively, trial judges need the constraint and guidance of general rules, as the following discussion demonstrates.


As we have seen, there are two distinct complaints about court rulemaking today. One focuses on the efficacy of the rules themselves and the other on the legitimacy of the process which creates them. Moreover, although the distinction is seldom made clear, there are two different aspects to efficacy. Efficacy can refer to the quality of the rules produced or to the likelihood of their implementation in the face of political opposition. Even the best rules will be ineffective if political opponents can block them in Congress. By the same token, however, in a world of public choice distortion, the likelihood of congressional disapproval says nothing necessarily about the quality of the rules themselves.

This Part addresses concerns about efficacy, and Part V below addresses concerns about legitimacy. A process can lack legitimacy in two different ways: either because it produces bad outcomes, or because it treats individuals unfairly by denying them voice, participation, power, or the like. The first kind of legitimacy-“outcome-oriented” legitimacy-equates legitimacy with efficacy. Thus, the discussion in this Part, insofar as it demonstrates efficacy, also establishes outcome-oriented legitimacy.

However, the second kind of legitimacy-“process-oriented” legitimacy-is distinct from efficacy. It is not enough on this view to show that court rulemaking produces good rules. Rather one must show that rulemaking is not the sort of activity that requires participation and political accountability. Part V takes on this task.

In the discussion of efficacy that follows, I focus more on the quality of rules than on the politics of implementation. I do so for two reasons. First, a coherent and attractive justification of court rulemaking can help with implementation problems. Ideas have power in the political process notwithstanding the force of raw interest. Armed with a persuasive justification of their role, court rulemakers can make it more difficult for Congress to justify intervention.

Second, we must understand what court rulemaking does well, and why, in order to choose an optimal response to pragmatic problems of implementation. Even if we succeeded in quelling political controversy-say, by drastically limiting the scope of rulemaking, acting only on the basis of consensus, or leaving controversial issues to case-specific discretion-it would hardly be worth the effort if we ended up with bad rules. To be sure, rulemaking is a practical enterprise, but good practical tradeoffs are possible only with a clear theoretical understanding of what rulemaking is supposed to do.

To evaluate the efficacy of court rulemaking, we need a metric for judging the quality of procedural rules. There are three such metrics, which differ according to what they assume is valuable about procedure: an efficiency metric that values procedure in terms of its effect on aggregate social costs; a rights-based metric that values procedure in terms of its contribution to enforcement of individual rights, and a process-based metric that values procedure in terms of its intrinsic regard for the dignity and autonomy of litigating parties.

The first three sections in this Part discuss court rulemaking under each of the three metrics. The fourth section then addresses the problem of how to make rules when there is reasonable disagreement about the proper metric to apply.


Efficiency requires a tradeoff between the costs of error and the costs of process to safeguard against error.’63 Adding procedure reduces error costs by reducing the risk of mistakes, but it also increases process costs. An optimal procedural system minimizes the sum of error costs and process costs.

From an efficiency point of view, there is no fundamental difference between substance and procedure. Both types of rules share a common goalminimization of social costs-and both operate in tandem to further that goal. Substantive law creates incentives for efficient behavior by adjusting the payoffs from different courses of action, and procedural law facilitates the substantive goal by increasing the likelihood of accurate enforcement. The rational actor decides what to do by comparing the expected costs of alternative actions, and those costs are a product of the sanction imposed by the substantive law and the probability of enforcement created by the procedural law.164

Thus, a lawmaker can induce efficient behavior by tinkering with substance, or by holding substance constant and tinkering with procedure. As a result, what we call “substance” and “procedure” are from an efficiency perspective simply different tools for achieving the same goal. The tools are not equally effective for all purposes, however, so the objective is to choose the mix of tools that minimizes total social cost.165

The question for court rulemaking is why someone with this view of procedure would choose a court-based, committee-centered, and centralized process aimed at producing nationally uniform rules.

1. Court-based and Committee-centered

At first glance, it might seem optimal to vest power over procedure in Congress so as to facilitate efficient coordination between substance and procedure in statutory law. Once the benefits and costs are considered, however, a strong case can be made for giving the courts power to make a default procedural system subject to legislative override in particular statutory contexts.

a. A Default Approach. Assuming that procedural rules set in advance should fit substance to some extent,66 there are two different ways to accomplish the fit. One possibility is to leave it to the legislature to make the entire procedural system anew for each substantive area it regulates. The other possibility is to design a default system subject to legislative exceptions for particular substantive areas.

The first approach would be extremely costly and thus justified only if procedure varied extensively with substance.167 An optimal system of procedure, however, is not likely to be so substance-specific. For one thing, procedure is a cruder instrument than substantive law for affecting real world incentives. As a result, one would expect more reliance on substance to shape incentives and less fine-tuning of procedure.168 Substance-specific rules are likely to have more impact on process costs, but even here substantial benefits are likely only in those substantive areas noted for large and complex suits, such as antitrust, securities fraud, and product liability,169 and for those procedures, such as pleading and discovery, that are highly sensitive to process costs.170 Thus, a default system should apply to a wide range of cases, saving the cost of making the same rules repeatedly.171

This brief analysis, though hardly rigorous,172 supports the intuitively obvious proposition that redesigning a procedural system for each substantive area would be wasteful. In fact, Congress currently legislates on the basis of a default approach. Only rarely does it create an entire system of procedure for a particular substantive area.173 Instead, Congress tinkers with elements of the default system created by the Federal Rules of Civil Procedure.174 This does not mean that the default approach is necessarily efficient, of course, but it does mean that Congress would likely create a default system itself if given exclusive power over the rulemaking function.

b. Constructing Default Rules. Thus, a default approach seems optimal. The next question is who should design the default rules-the legislature or the courts. Since default rules are not tightly fitted to substance, the coordination benefits of assigning the task to Congress are limited. Moreover, congressional rulemaking is likely to generate serious public choice inefficiencies, which could be reduced significantly by using a court-based committee-centered process.

The public choice problems with legislation are familiar.l75 In public choice theory, legislators further their own self-interest and respond to interest groups insofar as those groups make credible promises or threats to reward or punish, such as promising campaign support or threatening to support opponents.l76 Interest groups have conflicting interests, however, and one common way legislators manage the conflict is through logrolling. The typical logrolling scenario involves a deal between two legislators, each eager for the other to support a project that benefits politically powerful constituents.

Logrolling is not necessarily inefficient,l77 but it becomes so when collective action problems prevent some interested groups from organizing and thus participating in the legislative bargain. Under those circumstances, externalities often plague policy decisions, because organized groups with political power have incentives to secure legislation that benefits themselves at the expense of unorganized and less powerful groups.178

The experience with rulemaking over the past twenty years confirms that powerful groups have strong and conflicting interests in procedure.l79 In addition, collective action problems pose serious impediments to the ability of some affected groups to organize in the political process. For example, plaintiffs have strong interests in procedure, but they are extremely difficult to organize. Transaction costs and free-rider problems create substantial obstacles for a group so widely dispersed.lso Moreover, prospective plaintiffs who have not yet been injured have little incentive to invest in securing procedural rules that would benefit them only in the event of serious injury. And because most plaintiffs are one-shot litigants, even those with lawsuits pending are not likely to care much about rule reform, as their suits will probably terminate before any rule change takes effect.

It is true that there are organized groups that represent plaintiffs. But those groups tend to focus selectively on particular kinds of litigation, such as civil rights, and their interests do not necessarily match the interests of all their constituents. Bar groups, such as the American Trial Lawyers’ Association, represent the interests of plaintiffs to some extent, but lawyer incentives are sufficiently different from client incentives that bar groups cannot be trusted always to represent plaintiffs faithfully.181

Because of these collective action problems, therefore, legislative rulemaking is likely to be plagued by inefficient logrolling. But this is a persuasive argument for court-based rulemaking only if court rulemaking involves less severe public choice distortions. Professor Jonathan Macey suggests otherwise in what is the most extensive public choice analysis of court rulemaking to date.182

Professor Macey argues that federal judges will adopt rules that maximize their private utility, understood as a function of preferences for leisure time, prestige, and power to make decisions that advance private conceptions of the public good.l83 Macey predicts that judge-rulemakers with these incentives will support procedural rules that maximize trial judge control over litigation, because rules of this sort provide numerous opportunities for advancing judicial self-interest in individual cases.l84 The resulting procedure, he concludes, is likely to be inefficient because of externalities.’85

Macey’s account, even on its own terms, is incomplete in an important respect.’86 It ignores Congress.187 As we have seen, Congress plays a significant role in procedural rulemaking today, and the Advisory Committee and federal judiciary are concerned about political end runs around the Enabling Act process.’88 Thus, a public choice analysis should treat court rulemaking as a strategic game among the Advisory Committee, Congress, and the various interest groups.189

Self-interested rulemakers faced with the threat of congressional intervention should be willing to make concessions to powerful interest groups in order to maintain some control over the rulemaking process.’90 Thus, interest groups should have influence over rulemaking, though their influence will be limited by the costs of successful lobbying in Congress.191

Nevertheless, inefficient logrolling is less likely in a court-based rulemaking process which relies on a committee system. 192 Because the Advisory Committee has control over a limited range of subject matter, court rulemaking involves a much narrower policy space and thus a smaller region over which trades can be made. Moreover, vigorous logrolling is likely only if rulemakers are strongly allied with distinct constituencies-an unlikely scenario for a committee composed mainly of federal judges. To be sure, interest groups can trade with one another directly. A corporate group, for example, might agree not to oppose a class action amendment favored by a civil rights group in return for the civil rights group agreeing not to oppose a strict sanctions rule. But here too, opportunities for compromise are limited by the confined policy space as well as by the more limited opportunities for agenda manipulation.’93

Even so, this analysis hardly supports a ringing endorsement of court-based rulemaking. The conditions that frustrate logrolling in the committee settingrulemaker incentives and a limited policy space that together make compromise difficult-are also a recipe for stalemate. In addition, rulemakers who seek to satisfy interest groups will tend to focus on rules capable of garnering consensus. As we have seen, these rules often include highly general discretionary standards that give competing interest groups a chance to wage their battles in individual suits.l94

Public choice therefore offers two rather dismal options for rulemaking: inefficient logrolling in the legislature or toothless rules made by the courts. There is a way out of this dilemma, however. With a reasonably coherent idea of what they are supposed to do-something that has been missing since the demise of the traditional justification-rulemakers should be able to transcend narrow self-interest, engage in reasoned deliberation, and avoid the public choice trap. Indeed, individuals are capable of choosing an action simply because they believe it is the morally right thing to do. If one insists on an explanation in terms of utility maximization, then the desire for reputation and prestige should suffice, at least so long as there is a robust reputation market for public-regarding behavior.’95 Accordingly, clear rulemaking norms can support development of professional standards, which in turn can constrain the rulemaking process and guide it in the direction of the public interest.

However, this happy result requires that rulemakers not be strongly dependent on outside interests and that their numbers be small enough to facilitate reasoned deliberation. This argues for a court-based rulemaking process with committees staffed mainly by judges, who are not wedded to outside interests or committed to compromise in the way legislators are. My point is not that judges have no private stake in rules. Certainly they do, and so do a lot of other groups. But insofar as outcome quality is concerned, the better response to public choice problems is to improve the deliberative process by controlling rulemaker incentives rather than accommodate competing interests. And this response is likely to be more successful with a court-based, committee-centered approach.

It is true that even an advisory committee with the right incentives still must deal with the risk that interest groups will lobby Congress and seek congressional intervention. But this is inevitable in any rulemaking system. Moreover, a strong justification for court rulemaking, especially one that focuses on efficacy and legitimacy concerns, may be the best possible defense against legislative intrusion. Indeed, a justification of this sort is essential to any effort to negotiate with Congress for new rulemaking limits.l96

To recap, the argument from efficiency starts with the proposition that the best way to achieve an efficient coordination of substantive and procedural law is to adopt a default system of procedural rules that allows for legislative exceptions. It then shows that the best way to construct the default system is through a court-based and committee-centered rulemaking process.’97

2. Centralized

Assuming that courts aided by committees should make the system of default rules, is it better to do so in a centralized or a decentralized manner? Rulemaking can be decentralized in two ways: first, individual judges can make rules for individual cases, or second, a committee-based process in each local judicial district can make general rules applicable to all the cases filed in that district. However, neither approach can substitute for a centralized process aimed at national uniformity.

a. Case-specific Rulemaking. The first alternative, case specific rulemaking, is an extreme version of what many commentators think happens today under rules that delegate broad power to trial judges.l98 This approach, however, has serious limitations from an efficiency perspective. Some degree of judicial discretion is valuable (indeed, essential) to tailor procedure to the special circumstances of individual cases, but discretion should not replace a centralized rule system.

Obviously, case variation is not extensive enough to justify inventing all procedures anew for each separate case, given the high cost of doing so. To reap the benefits of case-specific tailoring on this level, the judge would have to gather a great deal of information through adversary hearings, thereby multiplying litigation costs. Moreover, even experienced judges are likely to have considerable difficulty selecting efficient rules to control the complex strategic environment of litigation.’99

These factors are also relevant to deciding how much discretion rulemakers should give trial judges under a default system. Although discretion allows for case-specific tailoring, it also breeds litigation costs by opening up possibilities for adversarial contention. Moreover, discretion has questionable efficiency benefits compared to clear ex ante rules given the difficulties of anticipating strategic effects on a case-by-case basis.2 Indeed, even without strategic complexity, trial judges should have a hard time predicting the global costs and benefits of procedural alternatives in the setting of a particular case.

There is another important efficiency reason to limit case-specific discretion: discretion will likely exacerbate collective action problems and undermine the judge’s ability to manage the litigation through credible threats. This conclusion follows from the strategic dynamics of litigation.

One of the most well known strategic problems is the Prisoners’ Dilemma.201 To illustrate, consider discovery abuse. Although it is quite likely that discovery abuse has more than one cause,202 many commentators point to the Prisoners’ Dilemma as a particularly salient factor.203 Each party to a lawsuit has an incentive to engage in abusive discovery for fear that her opponent will do so and take her for a “sucker.” This is so even though she knows she would be better off saving costs by avoiding the abusive strategy. The problem is that she cannot trust her opponent to do the same. Thus, parties arrive at an inferior outcome-costly abusive discovery on both sides-because they cannot trust one another to stick to the Pareto-superior strategy.

The way out of a Prisoners’ Dilemma is to enable all players to make credible commitments in advance to follow the less costly strategy-in our example, to commit to limited discovery. The adversarial nature of litigation, however, makes it difficult for parties to commit on their own. Procedural rules provide a basis for credible commitment by imposing severe sanctions on any party who defects.

Such rules work optimally, however, only if they are clear enough to be understood in advance and only if they are enforced strictly whenever an infraction occurs. In theory, there is nothing to stop a trial judge from fashioning a clear rule, announcing it at the outset of the litigation, and strictly enforcing it against all violators. But there are two problems. First, to have a meaningful impact on party incentives, the judge must announce the rule at the beginning of the case; otherwise parties will gamble on the chance of a lax approach. But the trial judge has only limited knowledge at the outset of the case and only limited ability to adapt the procedure to the case.

Second, trial judges, in the federal courts at least, are not likely to announce a clear rule at the outset and even less likely to stick to it ex post. Federal judges covet their discretion to manage individual cases and are accustomed to dealing with procedural issues on an individualized and ex post basis.2 Moreover, because many of these judges were once trial lawyers themselves, they tend to empathize with the violator and exercise leniency at the sanctions stage.25 As a result, parties will not be able to rely on rules for a guarantee of commitment and will thus be thrown back into the Prisoners’ Dilemma with its costly consequences.

Of course, federal judges bent on exercising case-specific discretion might simply ignore centrally promulgated rules. But clear rules provide sharper standards for evaluating judicial performance and a firmer basis for appellate review. Moreover, I have enough faith in federal judges to think it highly implausible that they would systematically flout a clear rule obviously meant to constrain them.206

The efficiency argument for clear and strict rules is not limited to the Prisoners’ Dilemma. Clear rules also help a judge make credible threats in advance, which enlarges the range of regulatory options available to the judge.27 To illustrate, suppose we know that most frivolous suits are filed by plaintiffs who do not investigate, and that the efficient rule would impose sanctions whenever a plaintiff fails to do a reasonable pre-filing investigation even if suit turns out to be meritorious.28 Suppose the trial judge strongly believes that only meritless suits should be sanctioned. Without a binding rule set in advance, the judge, acting in a perfectly rational way, would not impose sanctions whenever a suit proved to be meritorious. Plaintiffs, knowing this, would investigate less often, gambling on the chance that their suit is meritorious.

The obstacle to the efficient result in this example is that the judge with her publicly known beliefs cannot credibly threaten in advance to impose sanctions on meritorious suits. Even if she threatens to do so, plaintiffs will call her bluff because they know she will not follow through on her threat. A clear procedural rule exogenously adopted and strictly enforced, with strict enforcement monitored on appeal, reduces the inefficiency by making threats credible.209 And the rule is efficient not because it makes everyone better off, as in a Prisoners’ Dilemma, but because it creates social benefits that exceed the additional costs to plaintiffs.

To recap, a rulemaking system designed to achieve efficiency should place limited reliance on case-specific discretion, more limited in fact than under the current Federal Rules of Civil Procedure.2’o Broad discretion generates high administrative and litigation costs with questionable efficiency benefits, impedes the solution of collective action problems, and frustrates the use of credible threats to achieve regulatory objectives.21 As I argue below, even though case-specific discretion can sometimes be useful as part of a commonlaw method of learning what works in practice, discretion must be meaningfully guided and subject to the periodic oversight of a centralized rulemaking process.212

b. Local District Rulemaking. Ruling out broad case-specific discretion, however, does not make the case for centralization at the national level. Rulemaking power could be given to local districts to exercise on a district-wide basis. This kind of rulemaking is quite common in the federal system despite the Federal Rules of Civil Procedure,213 and it has been much discussed in recent years.214

Local rulemaking is unproblematic insofar as it deals with routine administrative matters, such as where to file and how to get access to court records.215 Moreover, local rulemaking offers potential benefits beyond this limited area. Some commentators argue, for example, that local rulemaking facilitates the matching of procedures to local conditions,216 makes room for experimentation, and under proper conditions even supports competition for efficient rules.217

However, these benefits are not as obvious as supporters of local rulemaking assume. For instance, it is unclear that local conditions vary all that much in ways relevant to the design of efficient procedural rules.218 It is true that the character of the caseload varies somewhat from district to district. A district inundated with large and complex cases, for example, might have a particularly strong need for discovery limits or ADR. But this does not mean that the procedures must be set at the local level. Instead, a central rulemaking body could authorize a menu of procedures by general rule and set national standards to control their use.

The argument from local experimentation also has flaws. Local rulemaking can provide some useful information, but lack of rigorous experimental design limits its utility.219 If there were an autonomous ordering device, such as the market, that rewarded efficient rules, then it might be possible to rely on an evolutionary process to select for efficiency.220 However, conditions for efficient competition are unlikely especially given the incentive to externalize burdensome litigation costs.221

In addition, local rulemaking increases the risk that interest groups will dominate the rulemaking process. Although it would be costly for national groups to organize in each locality, those powerful bar and client groups with local chapters should have a much easier time wielding influence over a local than a nationally-centralized process. Lobbying is probably less costly in a local district, and close professional ties between local lawyers and judges increase the chance of success. To be sure, rulemakers can transcend narrow self-interest in a committee format, but doing so is likely to be more difficult on the local than on the national level.

At the same time, if lobbying is less costly on the local level, one might expect a broader range of groups to have a voice in local rulemaking. But collective action problems still raise substantial obstacles to participation, especially in geographically large districts, and the more powerful groups are likely to have greater influence on the local level.

Locally divergent rules also raise the costs to lawyers of practicing in multiple districts.222 It is true that an outside lawyer could just retain local counsel, but the need to do so creates barriers to entry. Specialized local rules add to entry costs and reinforce the market power of local lawyers, at least if competition in the district is already weak. Indeed, one might predict the emergence of rather detailed local rules, since complex rules raise the costs of entry.223

Decentralized rulemaking presents yet another problem. Delegating exclusive rulemaking power to local districts can create free riding and externalities, and even trigger an inefficient “race to the bottom.”224 For example, a local district bent on limiting its caseload might enact strict pleading or narrow joinder rules that encourage plaintiffs to sue elsewhere. Faced with this prospect, every district would have an incentive to enact similar rules in order to avoid being the only plaintiff-friendly jurisdiction left and thus a magnet for litigation. The likelihood of this scenario depends on how different districts weigh the costs and benefits of litigation and on the latitude for forum shopping left after venue and personal jurisdiction constraints. Nevertheless, the possibility cannot be ignored. In cases involving significant contacts with multiple states or implicating federal causes of action with nationwide jurisdiction, venue and personal jurisdiction requirements will often be satisfied in several federal districts.225 Moreover, cases with multistate contacts tend to be large and complex, as do cases involving some of the more popular federal causes of action for which Congress has authorized nationwide jurisdiction.226 Thus, a local district wishing to limit its caseload might expect to gain from creating a procedural environment hostile to plaintiffs.227

To be sure, more cases mean more litigation business for local lawyers. However, rational lawyers will balance this benefit against the potential loss of business clients attracted to more defendant-friendly jurisdictions. Furthermore, not many districts need be willing to opt for plaintiff-hostile rules in order to induce an inefficient race to the bottom. All one needs is a large enough number so that the expected judicial and litigation burden on the remaining districts overwhelms the expected benefit from additional litigation business. Finally, judges are likely to be more concerned than lawyers about additions to the district’s caseload, and judges usually dominate the local rulemaking process.

In sum, a case can be made on efficiency grounds for a centralized, courtbased, and committee-centered rulemaking process aimed at designing a nationally uniform system of default rules for the federal courts subject to legislative override. Room would still exist for some case-specific trial judge discretion and for local rulemaking, but these would be more limited than they are today and guided by meaningful constraints.

At the same time, it is important to note that the efficiency case against local rulemaking is not clear cut. The analysis is quite sensitive to empirical factors that are poorly understood, such as the magnitude of public choice problems on the local level, the competitiveness of local legal markets, and the costs and benefits to localities of different procedural packages. Even so, the risks are serious enough to counsel against leaving the basic elements of a procedural system to local control.


A rights-based metric of outcome quality, though different from the efficiency metric, supports similar conclusions about court rulemaking. In contrast to efficiency, a rights-based theory defines the value of accuracy in terms of the moral value of protecting individual rights. This focus on morality makes concerns of distributive fairness relevant to rulemaking, and a court-based, committee-centered, and centralized rulemaking process is well suited to the kind of fairness analysis that a rights-based theory entails.

The goal of adjudication in a rights-based theory is to enforce the substantive rights of the parties rather than to minimize social costs.228 A substantive right gives the individual rightholder a claim that society must honor, even if doing so increases social costs and reduces aggregate welfare.229 Thus, rights trump or act as side-constraints on the pursuit of utilitarian goals; a right cannot be limited merely because doing so would create greater net benefits for all.230

The existence of substantive rights implies procedural rights.23′ Because effective enforcement is critical to respecting a substantive right, limiting procedure for reasons of utility would be tantamount to limiting the substantive right on utilitarian grounds. Put differently, without the constraint of procedural rights, courts could undermine substantive rights in order to serve utilitarian goals by denying the costly procedures necessary to their enforcement.

It is no easy matter, however, to specify the content of a party’s procedural rights. At first glance, it might seem that a party should have a procedural right to a perfectly accurate outcome, since anything less would violate her substantive rights. The problem with this approach is that it makes impossible demands. On this view, no system could ever respect rights, since error is an inevitable feature of any procedural system no matter how well designed. Moreover, any effort to achieve perfect accuracy would commit most of the public treasury to financing procedure at the expense of other public goods.

There are two ways to deal with this problem under a rights-based approach. One way supposes that procedural rights are rights to perfectly accurate outcomes but that violations due to outcome error can somehow be excused (but not justified) if society does the best it can with its limited resources.232 The second way conceives of procedural rights as rights to fair treatment, not to accurate outcomes.TM On this alternative view, excusing outcome error is unnecessary because error does not violate procedural rights so long as the error is not due to skimpy procedures justified on utilitarian grounds and so long as the procedural system as a whole treats litigants fairly.

Each of these two approaches has shortcomings, and we need not choose between them here.TM In both, procedural rights place constraints on the risk of error a party must bear. These constraints, while they differ in other respects, overlap in recognizing the central importance of distributive fairness: error risk must be distributed fairly across cases and litigants.235 That each litigant holds a substantive right implies that each stands as an equal alongside others and has a claim to be treated as an equal in the allocation of scarce process resources. This does not necessarily mean that resources must be distributed equally, but it does mean that the error distribution must be justified by general principles that treat all litigants as equal rightholders with entitlements to the process resources that make accuracy possible. For example, if all that mattered for procedure was the formal identity of substantive rights qua rights, then fairness would require an equal distribution of error risk, since the principle of formal equality gives no reason to draw distinctions.

The central questions for our purposes are: first, what sort of institutional structure is best suited to implement distributional fairness, and second, whether a court-based, committee-centered, and centralized process has any advantages.

1. Court-based and Committee-centered

First, consider the choice between court and legislature. Since Congress makes the resource allocation decisions, perhaps Congress should also make the distributional decisions by choosing the procedures that distribute scarce judicial resources, and thus error risk, across different cases and litigants. The problem, however, is that Congress is not particularly well suited institutionally to the task. Judges are better equipped to engage in the kind of inquiry that a fair distribution of error risk demands.

The first problem with a congressional process, even one free from the usual public choice problems, is its sensitivity to the systematic bias of public opinion. For example, false positive errors are likely to be more salient to the public than false negatives.236 Because a false positive alters the status quo, it is frequently more dramatic and more visible.237 Therefore, because legislation is highly sensitive to public opinion, a legislative process is prone to distort the error analysis.

The second problem with a legislative approach runs deeper. To understand it, we must examine more closely what is involved in designing a procedural system that distributes error risk fairly. Most significantly, a fair distribution must take account of the relative importance of the substantive interests at stake in different cases.238 There would be no need to do this under a purely formal view of substantive rights, for then all rights would be identical and error risk would be distributed equally. But such a view is an implausible account of American procedural law.239 If it were not, then serious class-wide racial discrimination would have to be treated the same as a minor landlord-tenant dispute.

In fact, there are many examples of situations in which American procedure takes account of the harm from error, not just the risk. Ever since New York Times v. Sullivan,240 for instance, courts protect the media’s First Amendment interests by forcing plaintiffs to bear the burden of pleading and proof on actual malice in a public figure defamation case-despite the obvious informational disadvantage-and by imposing a clear-and-convincing standard of proof. This result reflects a greater concern for false positives (successful frivolous suits) than for false negatives (unsuccessful meritorious suits), and this difference in turn reflects a higher value placed on First Amendment interests.241

Thus, a fair distribution of error risk should take account of the relative value of different substantive interests. The question is how to make the value judgments. It is not necessary to explore this difficult question in detail. It is enough to note that, for rulemakers, determining relative value is not a matter of legislative-type policy judgments. If it were, then procedural rights would be held hostage to social welfare, contrary to their status as rights, and we should have serious doubts about whether courts should be making rules at all. Instead, rulemakers should apply relative value judgments already embedded in the law. Procedural rights, after all, guarantee that parties be treated as equals in the way the legal system actually distributes error. What matters, therefore, is the internal consistency of the legal system in light of the values that the system actually assigns to the interests at stake.

More specifically, rulemakers must construct from existing practice a coherent general theory of value that fits and justifies the pattern of protection given to interests by the legal system as a whole.242 For example, if the law tends to be more solicitous of constitutional interests or of interests in avoiding bodily harm (as opposed, for example, to property damage) insofar as liability, remedy, or procedure is concerned, and if the more generous treatment fits a coherent and principled account of why values are assigned as they are, then rulemakers should give more weight to an error that burdens the favored interest and adjust procedureal benefits and burdens accordingly. The same is true if Congress makes clear that the substantive interest protected by a particular statute is especially important.243

The analytical process I describe (admittedly somewhat loosely here) is an exercise in legal interpretation similar to what many believe judges do when they decide individual cases, construe the constitution, or make common law.24 The close affinity between the rulemaker’s and the judge’s task suggests that rulemaking should be assigned to the courts rather than the legislature.245 Courts are experienced in legal interpretation and are more likely than legislators to take the job seriously. For the same reasons, rulemaking should also be handled in the first instance by a relatively small committee to facilitate the kind of reasoned deliberation that the task requires.

2. Centralized

The rights-based requirement of principled consistency also means that a system of rules must be made on a uniform basis for all courts in the relevant jurisdiction. This conclusion follows from a proper understanding of the domain over which distributional fairness claims operate. By virtue of her procedural right, litigant A can demand that the error risk in her case be fair relative to the error risk facing litigant B in another case. To define the scope of A’s right, however, we need to determine the class of Bs to which A has a right to compare herself. If that class is limited to litigants with cases before the same judge or cases filed in the same federal district, then it would be sufficient for rulemakers to do fairness comparisons for each judge’s or each district’s caseload taken separately. On the other hand, if the class of Bs encompasses all litigants in all cases filed in federal district courts nationwide, then a nationally centralized process would be necessary to coordinate the demands of fairness on a global scale.

I can think of no principled reason to define the class of Bs other than as all litigants in all federal cases nationwide. A’s procedural right gives her a claim to fair treatment available against the sovereignty responsible for establishing the court system, and by virtue of that claim, A can insist that she be treated fairly relative to every other rightholder within the same system. The federal courts comprise a unitary system recognized as such by Article III of the Constitution and established as such by Congress. Accordingly, the domain over which distributional fairness claims operate is the federal court system as a whole, and not individual judges or particular districts. The rulemaking power therefore should be nationally centralized to ensure global justice.

In sum, the core task of rulemakers committed to a rights-based theory is to design a procedural system that fairly distributes error risk. To do this task well, rulemakers must construct a general account of how and why the law assigns value to the interests underlying substantive rights, and this account must be based on general principles that accord each litigant equal concern and respect as a rightholder. Because the process requires rulemakers to infer general principles from practice, it is much better suited to the courts than to the legislature. And because rights-claims operate over the entire federal court system, the rulemaking task should be centralized.

At the same time, estimating error risk is a rough process, even with access to empirical data. What is important is that rulemakers work from a well justified account of what constitutes a fair distribution, that they do their best to realize that account in practice, and that they publicly explain their predictions and justify their rule choices by reference to distributive effects. I offer one final point. The rights-based account I give here leaves room for legislation and local rulemaking. On this view, court rulemakers monitor the basic elements of the procedural system to ensure that the system as a whole distributes error risk fairly given the value choices implicit in the substantive law and the way procedural rules actually work in practice.246 For example, if pleading or discovery rules come to have an unfair impact on a class of litigants, say, because of more intense strategic behavior or an increase in the number of frivolous suits, court rulemakers should change the rules or adjust other features of the system insofar as-and only insofar as-necessary to eliminate the adverse effects. In a rights-based system, it is the legislature that has the power to make major changes in the error distribution to serve policy goals, so long as it seriously considers the harm to litigants and conforms to constitutional requirements.247

Furthermore, only major features of a procedural system are likely to make a noticeable impact on relative error risk. Consequently, centralized rulemaking should focus primarily on the fundamental elements of the system, such as pleading, discovery, and summary judgment. Relatively minor details can be left to local rulemaking without risking a serious violation of distributive norms.


Although the matter is controversial, many proceduralists believe that litigants enjoy process-based procedural rights unrelated to outcome quality.248 Elsewhere I have expressed doubts about the process-based right to participate,249 and I will not repeat that discussion here. The point I wish to make (briefly) is that court rulemaking is an effective way to determine the content of process-based rights if such rights exist.

To illustrate, consider the participation right.250 In its process-based form, this right derives from the state’s obligation to respect the dignity and autonomy of persons affected by adjudication. For one committed to this view, the challenge is to define the specific participation opportunities the right guarantees. One might reason from a general theory of democratic participation, but it is hard to imagine a general theory specific enough to prescribe the working details of standing, joinder, intervention, class actions and the like. These institutional details are developed through institutional practice.

Thus, the process-based participation right is an institution-specific right, in contrast to the general background right of participation justified as a requirement of political morality. The institutional right of participation takes shape through an incremental process of determining what the background right means and how best to realize that meaning in practical litigation settings. This incremental process produces a set of institutional principles which are manifest in litigation practice and which define the scope of the process-based right.

For example, parties must often give up some of their litigating autonomy so that others can be treated fairly. Absentees facing potential unfairness, for instance, are allowed to intervene in a pending suit to protect their interests even over the objection of existing parties.TM So too litigants are sometimes forced to join class actions when separate suits would prejudice other class members.252 These rules reflect principles of fairness that limit the freedom of parties to control their own suits, and thus help define the scope of the institution-specific participation right. Court rulemakers must infer these principles from the rules and practices that express them.

This account of the participation right has important implications for rulemaking. Since the right is defined by adjudicative practice, rulemakers must look to practice to infer institution-specific principles defining the right. At the same time, rulemakers must also check whether those principles optimally further the background right they were meant to express. This process is similar to the interpretive method used to value substantive interests under a rights-based metric, and for the reasons discussed above, a court-based, committee-centered, and centralized rulemaking process is the best way to handle the task.


The previous discussion provides a justification for court rulemaking on outcome-quality grounds once a metric is chosen and even if (as is almost certain to be the case) that metric is some complex hybrid of the three already discussed. However, reasonable people disagree about the metric, and this creates an analytic and practical problem: how should we choose a proper metric if the right choice is not obvious?

This question is complex and difficult. If choosing a metric is a matter of social policy, it is not clear that it should be done by a court rulemaking process rather than by Congress (or a congressionally-created agency with political and public accountability). On the other hand, controversy alone cannot require a politically accountable social choice mechanism. For example, courts make decisions about competing values and sometimes trigger intense controversy when they interpret the constitution or make common law, and these decisions are not subject to a political process. Thus, what matters is how the choice of metric is made.

Constructing a Hybrid Metric

I do not have a complete answer to the question of how to choose a metric, but I can outline what I believe is a promising approach. The idea is to use the hybrid metric already embedded in current practice. Our procedural system, as it currently exists, embodies a complex mix of different metrics. The rules of preclusion, for example, give considerable weight to process-based participation values by recognizing a broad right to a “day in court,” but they also impose limits (in class actions and compulsory joinder, for example) justified on rights-based and efficiency grounds.3 This mix of values has developed through long experience with procedural rules in practice and reflects an ongoing commitment to working out the best features of a litigation system.

To employ such an approach, however, rulemakers must tease a metric out of existing practice. This task is complex for two reasons. First, it requires interpretation. Rulemakers must aim to construct a reasonably coherent set of principles254 that both fits current practice and also justifies that practice as the best it can be given what it is. This analytical process is similar to Ronald Dworkin’s interpretive approach to law255 and also shares much in common with John Rawls’s method of “reflective equilibrium.”256 The interpreter moves back and forth between concrete features of practice and well considered judgments of morality. Sometimes she fits the general principles to the practice, and sometimes she modifies the principles in light of moral demands. Throughout this process, the interpreter’s goal is to construct a set of legal principles that justify the practice as a reasonably coherent and morally attractive whole. At the same time, the requirement of fit means that the interpreter cannot deviate too much from existing practice.257 The result is a set of principles that are morally attractive (the justification requirement) but not morally ideal (because of the fit requirement); in other words, legal principles that strike a rough accommodation between the realities of practice and the ideals of morality.258

Second, the task is complex in terms of its scope. It would take the superhuman intelligence of a Dworkinian Hercules259 to construct a complete set of principles all at once for the entire procedural system. Thus, the job of constructing principles must be done incrementally in the context of particular rule choices. Rulemakers must build a set of general principles rule by rule, just as judges build the principles of the common law case by case.

As we have seen, this kind of interpretive method is well suited to court rulemaking by committee.260 The selection of a metric is not a social policy choice for the legislature to make; instead, it involves the sort of reasoned deliberation that courts do well. The legislature would still have broad power over procedure, but it would also have strong reason to exercise restraint given the advantages of court rulemaking and the difficulties inherent in predicting the effects of major reform. In fact, the legislature should probably act only when the regulatory benefits of a procedural rule are especially clear and compelling.

This account still leaves an important question: Why defer to existing practice at all, especially when some aspects of that practice are flawed?261 In answering this question, it is important to realize that the interpretive method I describe here does not simply accept uncritically the procedural status quo with all its flaws. Practice falls short of its own ideals in many areas, and rules that operate well can easily misfire when external conditions change. Rulemakers have the responsibility to revise the rules to address these deficiencies and to make the current system better realize the mix of values that it means to serve. Indeed, I have argued elsewhere that highly novel statistical methods of adjudication might meet standards of fit and justification-and so might be appropriate for court rulemaking-at least when delay costs are severe and process resources scarce.262

Furthermore, there are pragmatic reasons to defer to existing practice to some extent. Given the complex strategic dynamics of litigation, it is terribly difficult to know with certainty whether major changes will actually realize their intended goals. Although the test of time is by no means perfect, the fact that a procedural system has endured is a reason to accord it respect. From this perspective, court rulemaking is justified as an institutional device for harvesting the insights of experience.

Indeed, it is difficult to imagine how a legislature would go about choosing in the abstract a hybrid metric sufficiently comprehensive and complete to provide a normative framework for analyzing all procedural questions. Any acceptable metric is likely to consist of a complex array of competing values. Working out such an array in advance of concrete procedural issues would be an enormously difficult-perhaps even an impossible-task. Of course, one might give the entire rulemaking job to Congress, but doing so would sacrifice the benefits of a court rulemaking approach.

Finally, the interpretive approach I describe here holds out hope of reconciling disagreement on grounds that most people could accept.263 Many people who disagree about procedural specifics nonetheless accept the fundamental principles of our adjudicative system. Thus, an interpretive process aimed at those principles and their practical implications has the potential of garnering broad support.

Those who believe that the current procedural system is in need of radical reform are likely to be disappointed by my account of court rulemaking. Although rulemakers do have power to implement some major reforms, there are limits to what they can do in the name of perfecting current practice.264 This is not particularly troubling to me. I am rather skeptical of our ability to improve the procedural system through radical overhaul, especially given the epistemic difficulties inherent in regulating a complex strategic environment. Accordingly, I tend to favor a more gradual approach and thus see much to commend the process of court rulemaking described here (at least when supplemented by legislation for more substantial reforms). Yet there is also a lesson for those who favor radical change implemented quickly: legislation, not court rulemaking, is the appropriate vehicle for this kind of reform.

2. An Example: The Small Claimant Class Action

To illustrate, consider briefly the problem of whether and when to allow small claimant class actions. Rule 23(b)(3) of the Federal Rules of Civil Procedure currently authorizes class actions on behalf of persons with claims too small to justify individual suits but large enough in the aggregate to attract a contingency fee lawyer.265 Whether to allow such actions has always been controversial, because they provide little benefit to injured parties and create litigation where none would otherwise exist.266

Suppose this issue was one of first impression. In considering the processbased metric, rulemakers would first have to decide whether our procedural system recognizes a process-based participation right. If it does, rulemakers would then have to deal with the fact that the class action limits the participation opportunities of absentees, and decide whether absentees must be given notice and a right to opt out when their claims are small.267 In this regard, rulemakers might reason that the small claimant class action furthers processbased participation values even without notice to absentees and an opportunity to opt out, because it provides at least vicarious participation where none would otherwise exist. Alternatively, they might argue that the principle of processbased participation has little weight in the small claimant context, since class members are not likely to participate anyway and dignity is not significantly implicated for trivial losses.268 Whatever they decide, however, rulemakers should be guided in their reasoning by how small claims are treated in other areas of legal practice.269

Next, consider the distributive fairness concerns associated with a rightsbased metric. Rulemakers might well conclude, based again on an interpretation of current practice, that cases with very small stakes should receive little moral weight for purposes of distributing error risk, whatever the substantive interest involved. This conclusion might suggest that directing scarce process resources to costly class actions for small claims would be unfair if doing so meant a significantly higher error risk for more important claims. On the other hand, the best understanding of fairness, as embodied in practice norms, might guarantee each rightholder at least some practical opportunity to obtain a remedy, and the class action might be necessary to satisfy this guarantee for small claims.

The case is clearer from an efficiency perspective. The small claimant class action in effect empowers the attorney for the class to act as a private attorney general to enforce the substantive law.27 Some substantive areas, such as antitrust and securities fraud, are characterized by small, widely dispersed injuries, and the costs of those injuries might not be internalized without the small claimant class action.271

As this brief analysis shows, the case for the small claimant class action could depend on the choice of metric. Suppose that rulemakers conclude that authorizing a class action is efficient but that it would skew the error distribution in a way that cannot be justified on rights-based grounds. Rulemakers must then address the more general question of which metric to apply. Stated more concretely, rulemakers must decide whether it is ever proper for a court to grant a remedy solely for deterrence purposes when no substantial compensatory interest is at stake. In making this decision, they should look to how the law deals with private attorneys general in other contexts and develop a principled account of that practice, an account that defines the conditions under which the private attorney general device is appropriate.272

The example of the small claimant class action highlights the main features of the court rulemaking method. First, rulemakers should not implement their own policy preferences, harvest the preferences of some constituency, or seek consensus among competing interests. Instead, they should act on the basis of a general account of procedure that coheres with existing practice, and they should justify rules in terms of a best understanding of the metric selected by that account.

Second, the rules should leave some room for trial judges to adapt procedure to individual cases, subject to whatever strict rules are necessary to solve collective action problems and support credible judicial threats.273 To some extent, this is inevitable, since rules cannot cover all exigencies in advance. But additional discretion is also desirable despite its costs, because it provides valuable information to rulemakers about how rules and their underlying principles work in practice. For this to be possible, however, rulemakers must provide transparent justifications and clearly identify the guiding principles, so trial judges know what they should aim to do. By the same token, trial judges must report their reasoning to rulemakers, perhaps by publishing opinions on the more difficult procedural issues.274

The obvious connection between court rulemaking, as I describe it here, and a common law process raises the question of whether rulemaking should be left entirely to case-by-case common law adjudication. This is an important and difficult question;275 after all, we entrust large portions of our substantive law to the common law process. The reason to prefer court rulemaking for procedure, I believe, has to do with the special importance of systemic integrity to procedural design.276

Procedural rules are more interdependent than substantive rules. Parties view a lawsuit as a unitary event with a single objective, and they pick their strategies at each stage with an eye to the possible effects at every other stage. To regulate party conduct effectively, therefore, procedural rules must also be tightly coordinated across stages, and centralization is necessary to maintain coordination.277 Relying exclusively on decisions by trial judges in individual cases would place rulemaking at the mercy of case selection effects and random influences on the ordering of issues, factors that could frustrate any effort to achieve a tightly integrated system.278 But so long as rulemakers intervene periodically to correct serious threats to systemic integrity, case-specific decisionmaking along the lines of a common law process can be a useful mechanism for providing information about rules in practice.

I offer one final thought. The interpretive method I describe here aims at constructing a coherent account of the mix of values embedded in current practice. There are limits, however, to the degree of coherence possible when values conflict. To return to the small claimant class action example, a solution to the inherent conflict between efficiency and rights-based metrics might not be possible. Under these circumstances, the best we can hope for is a reasonable accommodation, perhaps along the lines of a rough lexical ordering of competing norms.279 If this is not possible, however, then the matter is beyond the power of rulemakers to address.


Part IV demonstrated the efficacy and outcome-oriented legitimacy of court rulemaking. This leaves the argument from process-oriented legitimacy. Simply put, the argument is that procedure is “political” because it has substantive effects and because its justification involves controversial substantive value choices, and in a democracy, politics is properly left to the legislature or to legislatively created agencies with greater political and public accountability.


That procedure has substantive effects does not by itself impugn legitimacy. As we saw in Part III, the argument gets off the ground only if it focuses on distributive effects, and then only if it claims that procedural rules systematically burden the poor and less powerful.2so But the argument has a serious problem even in this form. Inequality in the background wealth distribution affects many decisions that are not committed to a political process, such as decisions about the common law. Why then is procedure special in this regard? Perhaps procedure is special because wealth inequality has stronger and more systematic effects in the costly and risky environment of litigation.281 But even if this is true, the argument still needs a link to legitimacy. The only plausible link, I believe, is to suppose that a procedural system that produces wealthcontingent outcome inequality is unjust, unless those burdened by the system (or their political representatives) have some control over its design.

But what is the source of injustice in this argument? It cannot simply be that the background distribution of wealth is unjust (assuming that it is). If that were enough, then one would have to condemn even a perfectly accurate procedural system, since accurate outcomes will inevitably reproduce the injustice embedded in the substantive law.282

Procedure is not perfectly accurate, however, and as we saw in Part IV, distributive justice is relevant to the way a procedural system allocates the risk of error.2a3 But whether an unequal error distribution is unjust depends on the reasons for the inequality. What is significant about inequality of error risk due to wealth is that it is an inevitable byproduct of even the very best efforts to design an optimal procedural system.

To criticize court rulemaking for creating an unjust procedural system makes no sense when the injustice is inherent in any system no matter who makes it or how it is made.TM Furthermore, the most effective methods for redressing inequality are beyond the power of court rulemakers to adopt. For example, only Congress has power to fund legal aid and similar programs for the poor, and Congress has much more power than the courts to adopt fee-shifting rules.285

To be sure, rulemakers have some ability to redress wealth effects. For example, rules facilitating aggregation help plaintiffs take advantage of economies of scale, spread litigation costs, and attract attorneys. But each of these measures involves a difficult cost-benefit tradeoff that makes implementation less than obviously desirable.286

Thus, there is no obvious rulemaking response to wealth disparity. So long as rulemakers do what they can to correct the problem consistent with designing a sound procedural system, they act legitimately even in the absence of political accountability.287 It is for other institutions, such as Congress and state legislatures, to address the residual problems, and these institutions are already designed with participation and political accountability in mind.


The response to the argument from substantive justification is already clear from the discussion in Part IV above. Although rulemakers must value substantive interests and although assigning values can be a controversial matter, court rulemaking is not politics. Rulemakers do not simply make their own value choices. Instead, they implement their best understanding of the choices already embedded in the law.

If the legislature, for example, were to make clear that substantive interests protected by civil rights statutes should be valued more highly than other interests, court rulemakers would be obligated to abide by that choice. Of course, things are seldom so clear. Yet the rulemakers’ task is the same, just made more difficult because it requires interpretation. Personal values are bound to enter into the analysis to some extent, but those values must also fit the best interpretation of practice. Reasonable people may disagree, but this does not make the rulemaking process “political” anymore than reasonable disagreement makes constitutional interpretation or judge-made common law political.

To be sure, common lawmaking is not the same as procedural rulemaking. For one thing, the common law has a retrospective quality, developing gradually out of judicial decisions in particular cases and controversies, while procedural rulemaking is significantly prospective in nature. Yet the difference is not as sharp as it may seem. In a system of precedent, the common law is not simply retrospective; it also operates prospectively when a common law judge articulates rules and principles for a broad class of similar cases. Moreover, the procedural rulemaking process that I describe here is not simply prospective; it is also significantly retrospective insofar as it privileges past practice norms. In addition, rulemaking involves a mode of reasoning that is similar in important respects to the reasoning of a common law judge.288

None of this means, however, that we should exclude the public from rulemaking. It simply means that public participation is not essential to the legitimacy of the process. Public input can be useful, and public participation can sometimes ease acceptance of especially controversial rules. On the other hand, broad-based participation increases the risk of public choice problems and can distract from the message that legitimacy inheres in the efficacy of the institutional design rather than in the degree of public accountability.

Given these costs and benefits, the Advisory Committee should have broad discretion to weigh the advantages and disadvantages of participation and determine the optimal level of public input. One might worry that a Committee with this much discretion would closet itself away from public scrutiny. But Committee members should understand the benefits of public input and are unlikely to ignore those benefits, especially if they know that the rules they adopt will be judged by the quality of the justifications they provide.


Over the years, courts and commentators have struggled to define the proper limits of court rulemaking and the boundary between the Court and Congress, but with little success.289 The Rules Enabling Act describes those limits in terms of the slippery distinction between substance and procedure.29 The history of the 1934 Act indicates a congressional intent-not significantly changed by the 1988 amendments-to foreclose Supreme Court rulemaking where the choice among rules would have a “predictable and identifiable effect” on rights recognized by the substantive law or create rights that “approximate the substantive law in their effect on persons or property.”291 As to what constitutes a “predictable and identifiable” or other problematic effect, however, the Act furnishes little guidance.

The most helpful Supreme Court precedent to date, Burlington Northern Railroad Co. v. Woods,292 adopts a test of validity that turns on whether a rule’s substantive effects are incidental and whether the rule is “reasonably necessary to maintain the integrity of [the] system of rules.”293 But Burlington Northern says nothing about how to distinguish “incidental” from “direct” or how to tell whether a rule is “reasonably necessary” to maintain the integrity of the rule system. Nor does it say whether the drafter’s purpose, the rule’s apparent aim, or the rule’s actual effects ought to control the analysis.294

Courts and commentators have been unable to provide clearer guidance because they lack a coherent conception of the rulemaking task.295 The account I have given in this article helps fill this gap. In particular, we should measure the validity of a Federal Rule by reference to its proffered justification. In justifying a rule, rulemakers should meet three requirements. First, they should explain the rule explicitly in terms of its impact on the distribution of litigation error, its effect in reducing process and error costs, or its connection to processbased values.296 Second, they should not assert their own values, but instead implement their best understanding of the values embedded in legal practice.297 Third, they should show how the proposed rule fits current practice and furthers the principles embedded in a reasonably coherent and attractive normative account of that practice.298

This three-part test does not give precise answers, but it does define what is at stake and provides a helpful framework for analysis. On this approach, for example, a rule is not valid simply because it aims to regulate litigation-related conduct. It matters why rulemakers consider the rule an attractive mode of regulation. For instance, constitutional considerations aside, a rule substituting a truncated alternative dispute resolution mechanism, such as the mini-trial, for adversarial adjudication would probably be inappropriate for court rulemaking, though it aims at litigation-related conduct. Even if the first two requirements were satisfied, the third would be difficult to meet. The rule cuts back on participation opportunities, thereby triggering process-based objections. Moreover, the rule, if applied generally, might increase and distort the error risk in a way that raises fairness and efficiency concerns. Rulemakers would have the extremely difficult task of explaining how these various effects are consistent with the best principled account of current procedural law and practice.

At the same time, my test would not invalidate a rule simply because its justification depends on a judgment about the relative value of substantive interests or because the rule has substantive effects.299 In particular, a rule can still be valid even if it creates filing obstacles for some litigants or adversely affects the ability of some groups to litigate.3 Nor would my test condemn a rule just because the rule makes its procedural prescriptions vary with the substantive nature of the case. So long as rulemakers do not rely on their own independent judgments of substantive value and so long as the rule’s substancespecific distinctions are justified as consistent with a principled account of current practice, I can see no reason to bar the rule just because it is substancespecific.301 However, there are limits to how much rules should vary with substance. Too large a variance can add significantly to the error risk by increasing the likelihood of mistakes and also can increase litigation costs by inviting disputes over the classification of a case. Moreover, there are epistemic limits to the ability of rulemakers to distinguish the strategic effects of rules by substantive area.

The focus on justification underwrites a requirement that rulemakers explicitly and clearly justify their rule choices. The most obvious place to do this is in Advisory Committee Notes. This means, however, that Committee Notes must provide much more complete reasoning than they currently do.32 To be sure, there are practical limits to what one can reasonably expect from rulemakers with limited time and resources, but there is also considerable room for improvement over the current approach.303

In addition to normative issues of legitimacy, rulemaking also has practical constraints. In particular, rulemakers need not-and perhaps should not in the exercise of wise discretion-use their power when the potential costs of doing so are extremely high. For example, if a rule evokes especially strong political rancor, rulemakers might be justified in exercising restraint. However, these occasions should be limited to situations where the likely consequences are unusually serious. That a rule is politically controversial is not by itself a good reason to avoid the subject matter, anymore than political controversy provides a good reason to come out one way rather than the other on a common law issue or a matter of statutory interpretation. Indeed, it is through the responsible exercise of rulemaking power that the public will come to understand the role of court rulemaking in a democracy and appreciate the true basis for its legitimacy.


The court rulemaking model has lost its moorings. The original justification, based on a sharp normative divide between procedure and substance and the objectivity of technical expertise, no longer persuades. And no one has yet found a convincing alternative. Without a clear rationale, proponents have trouble defending the model against challenges to its legitimacy and efficacy. As a result, political interest groups lobby aggressively and Congress intervenes frequently. Indeed, rulemaking today more closely resembles a legislative process with broad public participation and interest group compromise than the process of principled deliberation it was originally conceived to be.

This article proposes a new justification for court rulemaking. The critics are correct in one respect: there is no clear normative divide between procedure and substance. Procedural rules do have substantive effects and require a potentially controversial weighing of substantive values. However, this does not mean, as the critics claim, that a court rulemaking process divorced from political accountability and allowing only limited public participation is necessarily illegitimate. The legitimacy of the process depends on what rulemakers are supposed to do when they make procedural rules, and that in turn depends on what a system of procedural rules is supposed to accomplish.

Thus, any defense of court rulemaking must be anchored to a theory of procedural law. This article considered three such theories and showed that the main features of the court rulemaking model fit the kind of deliberative process each theory requires. So long as court rulemakers stick to their area of special competence-inferring general principles from existing practice and choosing rules that implement those principles well in light of practice realities-they act legitimately, whether or not user groups are represented in the process.

Of course, the approach I have outlined cannot guarantee a rulemaking process free of interest group politics or congressional interference. Whenever a procedural rule has adverse effects on a politically organized group, there is always a risk that the group will apply pressure directly or indirectly through Congress. Pragmatics counsel against ignoring this risk, but its existence does not impugn the legitimacy of court rulemaking. Indeed, the integrity of the institution depends not on dodging hard issues, but instead on developing a clear and persuasive account of the institution’s legitimacy. The future of court rulemaking-indeed, the future of procedural law-hangs on the success of this venture.

1. The current version of the Rules Enabling Act is codified at 28 U.S.C. 2072 (1994). The term “court rulemaking” has been used at least since the early part of this century to distinguish rules made by the legislature from rules made by courts through an administrative-type process. It is important not to confuse court rulemaking with case-by-case common law adjudication. Court rulemaking involves centralized promulgation of general rules, but these rules are made by the court not the legislature.

2. See, e.g., Jeffrey W. Stempel, New Paradigm, Normal Science, or Crumbling Construct? Trends in Adjudicatory Procedure and Litigation Reform, 59 BROOK. L. REv. 659, 754-59 (1993).

3. See, e.g., Steven Flanders, Local Rules in Federal District Courts: Usurpation, Legislation, or Information?, 14 LoY. L.A. L. REv. 213 (1981); Bruce H. Kobayashi, et al., The Process of Procedural Reform: Centralized Uniformity versus Local Experimentation 19-29 (1994) (unpublished manuscript) (on file with author).

4. See, e.g., Stephen N. Subrin, Fudge Points and Thin Ice in Discovery Rejorm and the Case Jor Selective Substance-Specific Procedure, 46 FLA. L. REv. 27, 48-49, 54-55 (1994).

5. Indeed, criticism is so intense that the Standing Committee on Rules of Practice and Procedure commissioned a self-study of the rulemaking process in 1993. See A SELF-STUDY OF FEDERAL JUDICIAL RULEMAKING: A REPORT FROM THE SUBCOMMITTEE ON LONG RANGE PLANNING TO THE COMMITTEE ON RULES OF PRACTICE, PROCEDURE AND EVIDENCE OF THE JUDICIAL CONFERENCE OF THE UNITED STATES, 168 F.RD. 679, 683 (1995) [hereinafter SELF STUDY]. The reader should bear in mind that this article focuses on the rules of civil procedure, although some of its analysis might be relevant to other rules subject to a court rulemaking process, such as the rules of appellate procedure, the rules of criminal procedure, and the rules of evidence.

6. See, e.g., Mistretta v. United States, 488 U.S. 361, 392 (1989) (noting that rulemaking is “substantive and political”); Stephen B. Burbank, The Costs of Complexity, 85 MICH. L. REv. 1463, 1472-73 (1987) (challenging the neutrality thesis).

7. See, e.g., Judith A. Resnik, Failing Faith: Adjudicatory Procedure in Decline, 53 U. CHI. L. REv. 494, 535-37 (1986).

8. See Erwin Chemerinsky & Barry Friedman, The Fragmentation of Federal Rules, 46 MERCER L. REv. 757, 775-76, 792 (1995) (noting deep problems with the rulemaking process and the tendency today to advocate a pluralistic political approach).

9. See Stephen B. Burbank, Ignorance and Procedural Law Reform: A Call for a Moratorium, 59 BROOK. L. REv. 841, 846-55 (1993) (calling for a moratorium on rulemaking until we figure out what court rulemaking is supposed to do).

10. See id. at 853-55.

II. See Charles Gardner Geyh, Paradise Lost, Paradigm Found: Redefining the Judiciary’s Imperiled Role in Congress, 71 N.Y.U. L. REv. 1165, 1207-10 (1996).

12. Thus, this article has something to say about two issues: the scope of power that should be granted to court rulemaking as a policy matter, and the scope of power that Congress has in fact granted by the terms of the Rules Enabling Act. The article takes no position on the difficult question whether courts have the constitutional power to make a system of procedural rules in the absence of congressional authorization (or perhaps even in the face of a contrary congressional intent).

13. Accordingly, my purpose here is not to develop a comprehensive theory of lawmaking in the civil procedure field. Our current system relies on a mix of legislation, centralized court rulemaking,

local district rulemaking, standing orders adopted by individual judges, case-specific trial judge discretion, and common law-all subject to constitutional constraint. Many of these choices are difficult to parse, and it would take much more space than I have available in this article to analyze each thoroughly. My main purpose is to demonstrate that court rulemaking is an important part of this scheme and that it can be justified on grounds distinct from those that justify legislation.

14. 28 U.S.C. sec 2072 (1994)

15 See id secs 331, 2073. The Judicial Conference is the policy-setting body for the federal courts and its membership includes a mix of federal judges. See id sec 31.

16 See id

17 See id sec 2073.

18. In 1988, Congress mandated a period of notice and public comment, including oral hearings, at the Advisory Committee stage. See Judicial Improvements and Access to Justice Act, Pub. L. No. 100-702, 102 Stat. 4642 (1988) (codified at 28 U.S.C. 2073(c) (1994)). In Part IIc2 below, I discuss this development as one of several that reflect modem discontent with the traditional court rulemaking model.

19. See 28 U.S.C. 2073(b) (1994). 20. See id. 2074(a).

21. See Geyh, supra note 11, at 1187. I discuss the 1973 events and subsequent developments in Part II.C.2 below. Furthermore, Congress has power to act on its own and has from time to time created special procedures for particular substantive rights, see, e.g., Private Securities Litigation Reform Act of 1995, Pub. L. No. 104-67, 109 Stat. 737 (1995) (codified in various provisions of IS U.S.C. 77 (Supp. 1997)). Nevertheless, the Federal Rules of Civil Procedure still dominate federal court adjudication.

22. See FED. R. Civ. P. 83. Individual judges can even issue standing orders that spell out procedures specific to their courtrooms. See generally Stephen N. Subrin, Federal Rules, Local Rules, and State Rules: Uniformity; Divergence, and Emerging Procedural Patterns, 137 U. PA. L. REV. 1999 (1989) (discussing the proliferation of local rules).

23. See 28 U.S.C. 2071(a) (1994); FED. R. Civ. P. 83(a). However, there was some disagreement historically about exactly how limited local rules should be. See Subrin, supra note 22, at 2013-16. 24. In 1940, two years after adoption of the Federal Rules, the Knox Committee Report found a disturbingly large number of local rules. See Subrin, supra note 22, at 2016-18; see also Judith Resnik, Changing Practices, Changing Rules: Judicial and Congressional Rulemaking on Civil Juries, Civil Justice, and Civil Judging, 49 ALA. L. REv. 133, 201 (1997) (noting a 1966 survey that reported on local rulemaking excess).

25. See COMMITTEE ON RULES OF PRACTICE AND PROCEDURE OF THE JUDICIAL CONFERENCE OF THE U.S., REPORT OF THE LOCAL RULES PROJECT pt.l, at 1-5 (Daniel R. Coquillette rep., 1988) [hereinafter LOCAL RULES PROJECT]. The Rules Enabling Act was amended in 1988 to limit the rulemaking power of local district courts, see 28 U.S.C. 2071, 2077(b) (1994), and Rule 83 was revised in 1995 to the same effect. See Paul D. Carrington, Renovating Discovery, 49 ALA. L. REV. 51, 57-58 (1997).

26. See, e.g., Chemerinsky & Friedman, supra note 8, at 757-59; Lauren Robel, Fractured Procedure: The Civil Justice Reform Act of 1990, 46 STAN. L. REV. 1447 (1994); see also THOMAS E. WILLGING ET AL., DISCOVERY AND DISCLOSURE PRACTICE, PROBLEMS, AND PROPOSALS FOR CHANGE: A CASE-BASED NATIONAL SURVEY OF COUNSEL IN CLOSED FEDERAL CIVIL CASES 9-10 (1997) (reporting widespread lawyer discontent with nonuniformity in disclosure rules and support for a national rule). 27. 29 A.B.A. Rep. 395 (1906).

28. See id. at 408-14.

29. For the definitive history of the Rules Enabling Act campaign, see Stephen B. Burbank, The Rules Enabling Act of 1934, 130 U. PA. L. REv. 1015, 1043-98 (1982). The idea of court rulemaking was not new with the Enabling Act. The English Judicature Act of 1875 empowered English courts to make their own practice rules. Moreover, during the nineteenth and early twentieth century, the United States Supreme Court adopted rules for cases in equity and admiralty.


31. With the Federal Rules as an example of what could be accomplished, the ABA in 1938 launched a campaign to get court rulemaking adopted on the state level. See Report of the Section of Judicial Administration, 63 A.B.A. Rep. 522 (1938). However, the ABA’s campaign met with unexpected resistance from local lawyers and state bar associations bent on preserving the advantages of local practice, and it made little headway before the 1950s. See 64 A.B.A. Rep. 117, 404-06 (1939); see also John B. Oakley & Arthur F. Coon, The Federal Rules in State Courts: A Survey of State Court Systems of Civil Procedure, 61 WASH. L. REV. 1367, 1428 (1986) (listing only Arizona, Colorado, New Mexico, and the District of Columbia as adopting rules modeled on the Federal Rules before World War II).

32. For a summary of the major arguments and a partial bibliography of the literature up to 1926, see Judicial Versus Legislative Determination of Rules of Practice and Procedure-A Symposium, 6 OR. L. REv. 36 (1926).

33. See, e.g., Charles E. Clark, The Handmaid of Justice, 23 WAsH. U. L.Q. 297 (1938); Roscoe Pound, The Rule-Making Power of the Courts, 12 A.B.A. J. 599, 602 (1926); Edson R. Sunderland, An Inquiry Concerning the Functions of Procedure in Legal Education, 21 MicH. L. REv. 372, 382 (1922). This insight was at the heart of the reform program. It fit the popular explanation for the breakdown of the code and common law systems, which focused on the tendency of judges and lawyers to value formalistic and technical compliance with procedural rules. Pound stressed procedure’s instrumental function in his 1906 address, and the ABA made it one of the major planks of its procedural reform program. See, e.g., Report of the Special Committee to Suggest Remedies and Formulate Proposed Laws to Prevent Delay and Unnecessary Cost in Litigation, 34 A.B.A. REP. 578, 598 (1909); see also 35 A.B.A. REP. 614, 635 (1910) (listing the instrumental function of procedure as the second principle of the “General Principles of Reformed Procedure” and placing it just behind the “first and most fundamental principle” which endorsed court rulemaking).

34. See, e.g., Manley 0. Hudson, Why Confer Rule-Making Power on Courts?, 7 J. AM. JUDICATURE Soc’Y, 161, 162 (1924) (“Procedure is largely a matter of administration”); Charles H. Paul, The Rule-Making Power of the Courts, 1 WASH. L. REv. 163, 168 (1926) (noting that the legislature is engrossed with political matters and “not the dry details of legal procedure, which is as special a subject as chemistry or physics or higher mathematics”); Pound, supra note 33, at 602 (noting that “it is idle to expect legislatures to take a real interest in anything . . . so technical, and so recondite as legal procedure”).

35. The metaphor of procedure as “machine” or “tool” pervades the rulemaking literature. See, e.g., Edmund M. Morgan, Judicial Regulation of Court Procedure, 2 MINN. L. REv. 81, 84-85 (1918) (likening the judge to a “workman” and procedure to “tools of their trade”); Roscoe Pound, Some Principles of Procedural Reform, 4 ILL. L. REv. 388, 394 (1909) (noting that “the machinery of law must be judged by its success in achieving its end in concrete cases without undue waste or friction or consumption of fuel”); Edson R. Sunderland, The Exercise of the Rule-Making Power, 12 A.B.A. J. 548, 548 (1926) (noting that nineteenth-century codificationists “failed to appreciate the delicate adjustment of machinery necessary to an efficient administration of justice”). The metaphor reinforced the view that procedural design was strictly a matter of instrumental rationality logically dictated by the end to be achieved and properly left to engineers with technical expertise.

36. See Charles E. Clark, The Code Cause of Action, 33 YALE L.J. 817, 836-37 (1924) (referring to “convenience” as the main goal of procedure, as opposed to “policy” which was the business of substantive law).

37. The notion of “efficiency” that reformers had in mind was much narrower than the modern law-and-economics idea of cost-minimization. Nowhere is there any evidence that early twentiethcentury reformers understood the need for trade-offs between process cost and error cost. See infra note 163 and accompanying text. Procedure was efficient in the same way a machine or a business was efficient: it tolerated nothing that was not essential to its proper function, and it was streamlined to create a minimum of friction and waste.

38. Procedure reformers constantly stressed the importance of simplicity and flexibility. See, e.g., Pound, supra note 35, at 400-04. Simple and flexible rules allowed trial judges to use discretion to respond to the procedural needs of particular cases. Moreover, flexibility also meant that the rules themselves could be changed rapidly when experience revealed areas needing improvement. See Robert G. Bone, Mapping the Boundaries of a Dispute: Conceptions of Ideal Lawsuit Structure From the Field Code to the Federal Rules, 89 COLUM. L. REv. 1, 98-104 (1989) (discussing the importance of flexibility and simplicity to the drafters’ vision).

39. See Peter Charles Hoffer, Text, Translation, Context, Conversation: Preliminary Notes for Decoding the Deliberations of the Advisory Committee that Wrote the Federal Rules of Civil Procedure, 37 AM. J. LEGAL HIST. 409, 417 (1993) (describing Sunderland as “arguably the foremost proceduralist of his day”).

40. Sunderland, supra note 33, at 381-82.

41. See, e.g., Hudson, supra note 34, at 162; Pound, supra note 33, at 601-02; John H. Wigmore, All Legislative Rules for Judiciary Procedure Are Void Constitutionally, 23 ILL. L. REv. 276, 278 (1928). 42. The themes of expertise and experience figured prominently in justifications of court rulemaking. See, e.g., Pound, supra note 33, at 602-03; Sunderland, supra note 35, at 548; Edson R. Sunderland, Expert Control of Legal Procedure Thru Rules of Court, 4 CAL. ST. B.J. 216 (1930); Wigmore, supra note 41, at 278.

43. See, e.g., Paul, supra note 34, at 168; Wigmore, supra note 41, at 278. In fact, a number of powerful groups both inside and outside the bar had an interest in procedural matters. See Stephen N. Subrin, How Equity Conquered Common Law: The Federal Rules of Civil Procedure in Historical Perspective, 135 U. PA. L. REv. 909, 943-75 (1987); see also Robert G. Bone, Procedural Reform in a Local Context: The Massachusetts Supreme Judicial Court and the Federal Rule Model, in THE HISTORY OF THE LAW IN MASSACHUSETTS: THE SUPREME JUDICIAL COURT 1692-1992, at 393, 402-16 (Russell K. Osgood ed., 1992) (describing the political opposition to court rulemaking in Massachusetts). Moreover, many of the novel features of the Federal Rules, such as rules providing for broad discovery and injunctions, were of keen interest to powerful insurance companies and labor unions. See Hoffer, supra note 39, at 425. In view of these interests, the prevailing assumption that procedure was a purely technical subject devoid of substantive concern could not have been meant as a description of the real world, but more likely as a claim about the true nature of procedural law. 44. These advantages were summarized in the 1926 Report to the ABA by the Committee on Uniform Judicial Procedure, which recommended a short bill giving the Supreme Court broad rulemaking power:

This short bill is all the legislation at present required. To the student and the thoughtful man it is the key that will unlock the door to a new era of scientific judicial relations. It will set the judges and lawyers free to perfect the machinery of the courts for which they are held solely responsible by laymen. It is the principle adopted by England more than 50 years ago. The united Bench and Bar will cooperate in first constructing and then in gradually perfecting a

simple, correlated, scientific system of rules of procedure and practice in lieu of the present complicated “federal practice.”

Report of the Committee on Uniform Judicial Procedure, 51 A.B.A. Rep. 505, 513 (1926). 45. See Resnik, supra note 7, at 498-507. In fact, the strong interest in reform was fueled in part by fear that business clients, frustrated with the uncertainty, delay and expense of litigation, would abandon traditional adjudication in favor of arbitration, and that the legislature would remove disputes from the courts and place them in administrative agencies. See, e.g., Pound, supra note 35, at 388, 396.

46. See Alexander Holtzhoff, Two Years ‘ Experience Under the Federal Rules of Civil Procedure, 21 B.U. L. REv. 33, 33 (1941) (noting that the Federal Rule reforms were intended to “streamline the administration of justice by weeding out the unessentials, expediting the determination of controversies, and eliminating what has been so felicitously denominated as the `sporting theory of justice’ “).

47. See, e.g., Charles E. Clark, The New Federal Rules of Civil Procedure: The Last PhaseUnderlying Philosophy Embodied in Some of the Basic Provisions of the New Procedure, 23 A.B.A. J. 976, 977 (1937) (justifying discovery and summary judgment as much better devices than pleading to “aid in matters of proof,” making it possible for pleading to perform its proper function). Another reason for the popularity of court rulemaking relates to the influence of philosophical pragmatism on the early twentieth-century reform movement. The idea that procedure was best designed by experts reflecting on experience with procedural rules in practice is quintessentially pragmatic. See Bone, supra note 38, at 86 n.288, 93 n.319 (noting the influence of American pragmatism).

48. See Oakley & Coon, supra note 31, at 1377-78, 1428. The idea of court rulemaking, however, turned out to be more popular than the Federal Rules themselves. See id. at 1369. By the end of the 1960s roughly twenty states had adopted procedural rules modeled closely on the Federal Rules of Civil Procedure. See id. at 1428.

49. Before 1950, civil procedure courses in many law schools taught a combination of code and common law procedure with a heavy emphasis on pleading in the first year. By the early 1950s, however, academics in increasing numbers were calling for a pedagogical shift to the Federal Rules. See Richard C. Donnelly, Book Review, 60 YALE L.J. 377, 380 (1951); Harold R. Medina, Book Review, 53 COLUM. L. REv. 575, 577 (1953). This movement accelerated in 1953 with the publication by

Professors Richard Field and Benjamin Kaplan of Materials for a Basic Course in Civil Procedure, a new first year casebook built around the Federal Rules. See Delmar Karlen, The Place of the Federal Rules in the Teaching of Procedure, 7 VAND. L. REV. 571, 571 (1954) (“If there is any proposition upon which teachers of procedure seem to agree it is that the Federal Rules ought to be a focal point of interest in the study of their subject.”).

50. See generally Benjamin Kaplan & Warren J. Greene, The Legislature’s Relation to Judicial Rule-Making: An Appraisal of Winberry v. Salisbury, 65 HARV. L. REv. 234 (1951); A. Leo Levin & Anthony G. Amsterdam, Legislative Control Over Judicial Rule-Making: A Problem in Constitutional Revision, 107 U. PA. L. REv. I (1958); Roscoe Pound, Procedure Under Rules of Court in New Jersey, 66 HARV. L. REv. 28 (1952).

51. This is not to say that the rulemaking process had completely smooth sailing during this period. In fact, the Supreme Court discharged the advisory committee in 1956 without explanation. See Order Discharging the Advisory Committee, 352 U.S. 803 (1956). Congress filled the void in 1958 by assigning the advisory role to the Judicial Conference, see Act of July 11, 1958, Pub. L. No. 85-513, 72 Stat. 356 (1958), and the Judicial Conference in turn created a committee structure, including a standing committee and advisory committees, similar to those that operate today. 52. Legal Process was made popular during the 1950s and 1960s primarily through the work of Professors Albert Sacks, Henry Hart, and Lon Fuller of the Harvard Law School. See HENRY M. HART, JR. & ALBERT M. SACKS, THE LEGAL PROCESS: BASIC PROBLEMS IN THE MAKING AND APPLICATION OF LAW (William N. Eskridge, Jr. & Phillip P. Frickey eds., 1994). Teaching materials prepared by Hart and Sacks, see id., and circulated widely in draft form, inspired the creation of Legal Process courses at many American law schools. See ROBERT S. SUMMERS, LON L. FULLER 14 (1984); Robert G. Bone, ton Fuller’s Theory of Adjudication and the False Dichotomy Between Dispute Resolution and Public Law Models of Litigation, 75 B.U. L. REV. 1273, 1278 (1995). For historical background on Legal Process, see MORTON J. HORWITZ, THE TRANSFORMATION OF AMERICAN LAW 1870-1960: THE CRISIS OF LEGAL ORTHODOXY 253-68 (1992).

53. See Bone, supra note 52, at 1276-78.

54. These ideas are most clearly expressed in the work of Lon Fuller, one of the leading theorists of Legal Process. See LON L. FULLER, THE MoRALITY OF LAW 150-51 (rev. ed. 1969) (referring to law’s “inner logic”); Lon L. Fuller, The Forms and Limits of Adjudication, 92 HARV. L. REv. 353, 360-61 (1978) (referring to the “rational core of human institutions” and

the “rational elements that inhere in social organization” and that give direction and consistency to institutions).

55. See Bone, supra note 52, at 1301-10. 56. See id. at 1288-1301.

57. Indeed, Fuller believed that lawyers had special skills for institutional design; they were, in his felicitous phrase, “architects of social structure.” See, e.g., Lon L. Fuller, The Lawyer as an Architect of Social Structures, in THE PRINCPLES OF SOCIAL ORDER 264 (Kenneth I. Winston ed., 1981). 58. See Kaplan & Greene, supra note 50, at 252-53; Levin & Amsterdam, supra note 50, at 12-14; William M. Trumbull, Judicial Responsibility for Regulating Practice and Procedure in Illinois, 47 Nw. U. L. REv. 443, 451-53 (1952).

59. Dan Byron Dobbs, Judicial Regulation of Procedure, 9 ARK. L. REv. 146, 152-55 (1955); see Charles W. Joiner & Oscar J. Miller, Rules of Practice and Procedure: A Study of Judicial Rule Making, 55 MIcH. L. REv. 623, 634-36 (1957); Levin & Amsterdam, supra note 50, at 14. But see Pound, supra note 50, at 43-46 (arguing against legislative involvement by stressing the public choice problems with legislation).

60. See, e.g., Joiner & Miller, supra note 59, at 644-53 (assigning most topics to court rulemaking, including party and claim joinder, class actions, discovery, and res judicata); Levin & Amsterdam, supra note 50, at 14-20 (listing a number of topics suitable for court rulemaking with legislative check, including rules about taxing fees as costs and about venue). This is not to say that all criticism during the period was so tame. For example, Justices Black and Douglas frequently objected to proposed rules on the ground that they “determine matters so substantially affecting the rights of litigants in lawsuits that in practical effect they are the equivalent of new legislation which, in our judgment, the Constitution requires to be initiated in and enacted by the Congress and approved by the President.” 374 U.S. 861, 865-66 (1963) (dissenting from the transmittal of proposed Federal Rule amendments to Congress). See also Sibbach v. Wilson & Co., 312 U.S. 1, 16-19 (1941) (Frankfurter, J., dissenting) (objecting to Federal Rule 35 on similar grounds).

61. See STEPHEN C. YEAZELL, FROM MEDIEVAL GROUP LITIGATION TO THE MODERN CLASS ACTION 240-45 (1987) (discussing the manner in which civil rights, consumer protection, and environmental movements affected the class action).

62. See Marc S. Galanter, Why the “Haves” Come Out Ahead: Speculations on the Limits of Legal Change, 9 LAW & SOC’Y REV. 95, 97-110,140-44 (1974); Resnik, supra note 7, at 517-20. 63. See YEAZELL, supra note 61, at 238-66. 64. See Galanter, supra note 62, at 95-151.

65. See Richard L. Marcus, The Revival of Fact Pleading Under the Federal Rules of Civil Procedure, 86 COLum. L. REV. 433 (1986).

66. See generally OWEN M. FIss, THE CIVIL RIGHTS INJUNCTION (1978); Abram Chayes, The Role of the Judge in Public Law Litigation, 89 HAv. L. REv. 1281 (1976); Owen M. Fiss, The Supreme Court, 1978 Term-Foreword: The Forms of Justice, 93 HARV. L. REv. 1 (1979).

67. See Chayes, supra note 66, at 1284,1288-1304 (describing these developments as signs of a new procedural model, the so-called “public law litigation model,” that focuses on public norms rather than private rights).

68. See id. at 1294-95.


71. See Gary J. Miller, The Impact of Economics on Contemporary Political Science, 35 J. ECON. LITERATURE 1173, 1187-88 (1997) (positive political theory showed that procedural rules were “not neutral tools for expediting decision making, but potentially powerful weapons for biasing outcomes one way or the other”).

72. Chief Justice Warren Burger focused attention on these problems in a series of publications and addresses. See, e.g., Warren E. Burger, Agenda for 2000 A.D.-A Need for Systematic Anticipation, in THE POUND CONFERENCE: PERSPECTIVES ON JUSTICE IN THE FUTURE 23 (A. Levin & R. Wheeler eds., 1979); Warren E. Burger, Isn’t There a Better Way, 68 A.B.A. J. 274 (1982). Not everyone agreed with this diagnosis, however. See Mark S. Galanter, Reading the Landscape of Disputes: What We Know and Don’t Know (and Think We Know) About Our Allegedly Contentious and Litigious Societ, 31 UCLA L. REV. 4 passim ( 1983).

73. See, e.g., Wayne D. Brazil, The Adversary Character of Civil Discovery: A Critique and Proposals for Change, 31 VAND. L. REv. 1295 (1978) (discussing problems with discovery); Marcus, supra note 65, passim (describing the debate over notice pleading and the trend toward stricter pleading).

74. 424 U.S. 319, 334-35 (1976).

75. At least if the recipient had access to an evidentiary hearing following termination. See id. at 349. Indeed, the Court indicated that a full-blown pretermination evidentiary hearing would be the exception in such a case; ordinarily the Due Process Clause required much less. See id. at 343. 76. See id. at 334-35.

77. This interpretation is all the more evident when Mathews is contrasted with the Court’s earlier decision in Goldberg v. Kelly, 397 U.S. 254, 264-65 (1970). In Goldberg, the Court embraced the evidentiary hearing as a paradigm of ideal procedure.

78. See Richard B. Stewart, The Reformation of American Administrative Law, 88 HARV. L. REv. 1669, 1677-78 (1975) (describing the traditional rationale for agency decisionmaking). 79. See id. at 1681-88.

80. See Burbank, supra note 29, at 1018, 1020.

81. See Geyh, supra note 11, at 1187-88 (discussing the rulemaking process as vulnerable to interest groups and lobbyists); Richard L. Marcus, Of Babies and Bathwater: The Prospects for Procedural Progress, 59 BROOK. L. REv. 761, 805-12, 819 (1993) (noting and critiquing the politicization of procedural reform).

82. See Geyh, supra note 11, at 1187-88, 1227 (describing a move toward more divisive politics in rulemaking). Lawyer interest groups include the American Bar Association, the American Trial Lawyers Association, Lawyers for Civil Justice, and the American Tort Reform Association. See Jeffrey W. Stempel, Cultural Literacy and the Adversary System: The Enduring Problems of Distrust, Misunderstanding, and Narrow Perspective, 27 VAL. U. L. REV. 313, 336-37 (1993). Client lobbying groups include Public Interest, the Manhattan Project, and various corporate and insurance companies, civil rights groups, and consumer groups. Public policy think tanks such as the Brookings Institution, RAND Corporation, and the Heritage Foundation are influential as well. See id. at 328-32, 337. 83. See Kent Sinclair, Service of Process: Rethinking the Theory and Procedure of Serving Process Under Federal Rule 4(c), 73 VA. L. REv. 1183, 1197-1212 (1987) (recounting the legislative skirmish). 84. See Paul D. Carrington, “Substance” and “Procedure” in the Rules Enabling Act, 1989 DUKE L.J. 281, 283-84 (describing these events). For other examples, see Linda S. Mullenix, Hope Over Experience: Mandatory Informal Discovery and the Politics of Rulemaking, 69 N.C. L. REv. 795, 851-55 (1991) (describing interest group lobbying before the Advisory Committee on Rule 30 and Rule 11 revisions).

85. Judicial Improvements and Access to Justice Act, Pub. L. No. 100-702, 102 Stat. 4642 (codified at 28 U.S.C. 2073(c) (1994)).

86. In the past, the Advisory Committee circulated drafts to bench and bar and invited input, but never seemed to treat participation as a requirement of legitimacy. See, e.g., Edson R. Sunderland, The Regulation of Procedure by Rules Originating in the Judicial Council, 10 IND. L.J. 202, 209 (1934) (noting that the chief safeguard of public interest is not representation by the lay public). But see Geyh, supra note 11, at 1189 n.124 (arguing that the 1988 amendments added little to an already heavily politicized rulemaking process).

87. See, e.g., Chemerinsky & Friedman, supra note 8, at 792 & n.157 (calling for more representatives of user groups); Jeffrey W. Stempel, Halting Devolution or Bleak to the Future: Subrin ‘s New-Old Procedure as a Possible Antidote to Dreyfuss’s “Tolstoy Problem”, 46 FLA. L. REv. 57, 89-97 (1994) (recommending a rulemaking process that involves representatives from the bar and user groups); Laura A. Kaster & Kenneth A. Wittenberg, Rulemakers Should Be Litigators, NAT’L L.J., Aug. 17, 1992, at 15 (calling for more lawyers on the Committee so that rules are more responsive to the needs of the practicing bar). So far these efforts have failed. For example, the original bill that eventually amended the Enabling Act in 1988 included a provision, ultimately deleted, that would have made the Advisory Committee more representative, see Paul D. Carrington, Making Rules to Dispose of

Manifestly Unfounded Assertions: An Exorcism of the Bogy of Non-Trans-Substantive Rules of Civil Procedure, 137 U. PA. L. REv. 2067, 2076 n.50 (1989), and similar bills have been introduced from time to time since then.

88. See, e.g., WINIFRED R. BROWN, FEDERAL RULE-MAKING: PROBLEMS AND POSSIBILITIES (1981); JACK B. WEINSTEIN, REFORM OF COURT RULEMAKING PROCEDURES 106 (1977); Howard Lesnick, The Federal Rule-Making Process: A Time for Re-examination, 61 A.B.A. J. 579 (1975); Russell R. Wheeler, Broadening Participation in the Courts Through Rule-Making and Administration, 62 JUDICATURE 280 (1979).

89. See Stewart, supra note 78, at 1723-60 (proposing a model of interest representation); Peter L. Strauss, From Expertise to Politics: The Transformation of American Rulemaking, 31 WAKE FOREST L. REv. 745, 756-57, 760-66 (1996) (describing this trend).

90. See Stephen B. Burbank, Implementing Procedural Change: Who, How, Why, and When?, 49 ALA. L. REv. 221, 242 (1997) (noting the trend toward assimilating rulemaking to a legislative process and linking it to a greater congressional willingness to intervene); Chemerinsky & Friedman, supra note 8, at 792 n.157 (noting that many argue rulemaking should follow a pluralistic political model); Mullenix, supra note 84, at 832-36 (describing the views of the then-Reporter to the Advisory Committee, Professor Paul Carrington, recorded in a 1988 memorandum to the Committee, which remarks on the disturbing trend toward a legislative model of rulemaking); see also Patrick Higginbotham, Foreword, 49 ALA. L. REv. 1, 2 (1997) (noting that recognition of the need for direct participation of bench, bar and academy in rulemaking is “kindled by the reality that changes in rules of procedure today immediately engage social policy in ways that tax the dichotomy of substance and procedure”).

91. Pub. L. No.101-650,104 Stat. 5089 (1990) (codified at 28 U.S.C. 471-82 (Supp.1998)). 92. See S. REP. No.101-416, at 1-2 (1990), reprinted in 1990 U.S.C.C.A.N. 6802, 6804; H.R. REP. No.101-734, at IS (1990), reprinted in 1990 U.S.C.C.A.N. 6860,6861. 93. 28 U.S.C. 471. 94. See id. 478(b). 95. See id. 472(c).

96. The main provisions of the CJRA were supposed to expire on December 1, 1997. See Pub. L. No. 101-650, 103(b)(2)(A), 104 Stat. 5089.

97. For example, local districts adopted rules that provide for case tracking, early discovery conferences, mandatory discovery disclosure, alternative dispute resolution, and even fee-shifting. See, e.g., JAMES S. KAKALIK ET AL., RAND INSTITUTE FOR CIVIL JUSTICE, JUST, SPEEDY, AND INEXPENSIVE: AN EVALUATION OF JUDICIAL CASE MANAGEMENT UNDER THE CIVIL JUSTICE REFORM ACT 11-21 (1996) [hereinafter RAND STUDY]. The precise status of these local rules after the expiration of the CJRA is uncertain. Presumably, local rules in conflict with the Federal Rules of Civil Procedure should be invalid, see 28 U.S.C. 2071(a); FED. R. CIv. P. 83(a)(1), but given the history of broad local rulemaking, one could safely assume that these local rules are still being used.

98. The CJRA provides that, after a period of experimentation and analysis of the resulting data, the Judicial Conference will consider the local plans and adopt those innovations that seem desirable for the federal court system as a whole. 28 U.S.C. 479. The RAND Institute has completed its initial study of the CJRA reforms, see RAND STUDY, supra note 97, and the Advisory Committee has already started looking at possible revisions to the national rules.

99. See, e.g., Paul D. Carrington, A New Confederacy?: Disunionism in the Federal Courts, 45 DuKE L.J. 929 passim (1996); Chemerinsky & Friedman, supra note 8, passim; Linda S. Mullenix, The Counter-Reformation in Procedural Justice, 77 MINN. L. REV. 375 passim (1992).

100. See 28 U.S.C. 471; S. REP. No. 101-416, at 14-IS (1990), reprinted in 1990 U.S.C.C.A.N. 6802, 6817-18. A Brookings Institution study of the federal courts commissioned by Senator Joseph Biden, Justice for All: Reducing Costs and Delays in Civil Litigation 11 (Brookings Inst. 1989), recommended the “bottom up” approach that became the cornerstone of the CJRA.

101. For more on this idea and its role in the CJRA, see Chemerinsky & Friedman, supra note 8, at 763-64, and Mullenix, supra note 99, at 385-96. 102. See, e.g., Chemerinsky & Friedman, supra note 8, at 764. 103. See Mullenix, supra note 99, at 432-34, 436-38.

104. In particular, the Senate argued that the two goals-increasing access and improving efficiencyinvolved “policies extrinsic to the business of the courts” and affected “constituencies in their out-of-court affairs.” As a result, CJRA reforms were substantive in nature, requiring “the accountability and give and take of the legislative process.” Id. at 432-33. 105. See Burbank, supra note 90, at 231-32 (attributing the more extreme claims to legislative pique at equally extreme claims made by federal judges).

106. See Mullenix, supra note 99, at 429. Indeed, federal judges perceived the threat to their traditional rulemaking role, and the Judicial Conference vigorously opposed the CJRA. See id. at 411-18. The resulting political struggle further strained relations between Congress and the federal judiciary, already taxed by previous legislative forays into procedural reform. See Geyh, supra note 11, at 1207-10; Resnik, supra note 24, at 213-14.

107. The discovery amendments required mandatory disclosure, but allowed local district courts to opt out. In addition, the package sent to Congress also included a revised Rule 11. The Justices were not unanimous in their decision to transmit these amendments, however, and their disagreement surfaced in a highly unusual set of opinions. See Amendments to the Federal Rules of Civil Procedure, 146 ER.D. 401, 501 (1993).

108. See Paul D. Carrington, Learning from the Rule 26 Brouhaha: Our Courts Need Real Friends, 156 ER.D. 295, 307-09 (1994); see also Marcus supra note 81, at 812 (concluding that the discovery amendment controversy reflects “the high level of suspicion that now attends the reform process, a suspicion that probably has resulted in part from focusing on winners and losers”). 109. See Common Sense Legal Standards Reform Act, H.R. 956, 104th Cong. (1995); Attorney Accountability Act, H.R. 988, 104th Cong. (1995); Private Securities Litigation Reform Act, H.R. 1058, 104th Cong. (1995). The Private Securities Litigation Reform Act became law in 1995 over a Presidential veto. Pub. L. No. 104-67, 109 Stat. 737 (codified in various provisions of 15 U.S.C. 77 (1997)).

110. See, e.g., Geyh, supra note 11, at 1187-91; Minutes of the Civil Rules Advisory Committee, Mar. 20-21, 1997 available in 1997 WL 1056240 (committee chair noting the potential threat of congressional interference); Minutes of the Civil Rules Advisory Committee, May 1-2, 1997 available in 1997 WL 1056241 (noting the importance of accommodating interest groups in order to avoid congressional intervention).

111. See, e.g., Resnik, supra note 24, at 136, 185.

112. See Stempel, supra note 2, at 669-74. See also Alternative Dispute Resolution Act of 1998, Pub. L. No. 105-315, 112 Stat. 2993 (1998) (to be codified at 28 U.S.C. 651-58) (requiring local federal districts to make alternative dispute resolution available to litigants). 113. See Sibbach v. Wilson & Co., 312 U.S. 1, 9-10 (1941). 114. See supra notes 78-79, 89 and accompanying text.

115. The Rules Enabling Act expressly limits the scope of court rulemaking to “general rules of practice and procedure” that do not “abridge, enlarge or modify any substantive right.” 28 U.S.C.

2072(a), (b) (1994). This limitation was meant to preserve legislative control over substantive law. See Burbank, supra note 29, at 1106.

116. See Strauss, supra note 89, at 760-62 (noting the increasing importance of political oversight to theories of agency legitimacy during the 1980s and 1990s). Indeed, in recent years, some commentators have focused on executive control, arguing that it helps to ensure political accountability and mitigates some of the ill effects of factionalism. See, e.g., Steven G. Calabresi, Some Normative Arguments for the Unitary Executive, 48 ARK. L. REV. 23 (1995); Lawrence Lessig & Cass R. Sunstein, The President and the Administration, 94 COLUM. L. REV. 1 (1994). But see Cynthia R. Farina, The Consent of the Governed: Against Simple Rules for a Complex World, 72 CHI.-KENT L. REV. 987, 989-1018 (1997) (criticizing this view).

117. See, e.g., WEINSTEIN, supra note 88, at 8.

118. Congress located the Commission, at least formally, in the judicial branch and gave it power over a subject area, criminal sentencing, in which courts had exercised a traditional role. See Mistretta v. United States, 488 U.S. 361, 363-70 (1989) (describing the Sentencing Commission in an opinion upholding its constitutionality). Later in this article, I recommend that the Civil Rules Advisory Committee provide more extensive justifications for Federal Rules, and I analogize to the Sentencing Commission to support the proposal. See infra note 302.

119. See Mistretta, 488 U.S. at 368-69, 374-79. Also, in contrast to the court rulemaking model that I defend in this article, the Sentencing Commission is “fully accountable to Congress” and obliged to provide broad public participation pursuant to the notice and comment provisions of the Administrative Procedure Act. Id. at 393-94. Moreover, that Congress can constitutionally delegate power over sentencing guidelines to an independent commission, as Mistretta held, does not answer the more concrete policy questions concerning the form and scope that the delegation should take. Finally, it is worth noting that the legitimacy of the Sentencing Commission was hardly free from controversy. See id. at 413-27 (Scalia, J., dissenting).

120. See Note, Civic Republican Administrative Theory: Bureaucrats as Deliberative Democrats, 107 HARV. L. REv. 1401, 1405-12 (1994) (reviewing these deliberative theories).

121. See, e.g., WEINSTEIN, supra note 88, at 11 (“Some [rules] . . . have very important substantive and procedural impacts[, yet] they are generally promulgated without warning or public debate”); Burbank, supra note 6, at 1472-73 (arguing that because the effects of procedural rules are not “neutral,” consideration should be given to the political legitimacy of the rulemaking process); Chemerinsky & Friedman, supra note 8, at 785-86 (noting that even rules dealing with “minutiae” can affect substantive rights by influencing outcomes); Subrin, supra note 22, at 2050 (“many individual procedural rules or clusters of rules have an inherently political aspect, in that they favor or disfavor types of cases or litigants”); see generally Marcus, supra note 81, at 773-76 (criticizing the current tendency to characterize rulemaking as “political” because of the differential effects procedural reform has on “winners” and “losers”).

122. See, e.g., Subrin, supra note 22, at 2050. Cf Burbank, supra note 6, at 1472.

123. The fact of different treatment is not in itself problematic. Many procedural rules treat different kinds of cases differently and many more rules contemplate different treatment depending on the type of case. For example, Rule 9(b) of the Federal Rules of Civil Procedure requires particularistic allegations of fraud or mistake but allows general pleading of most other matters. The mere fact that some cases receive different treatment is no more a problem in itself than any of the myriad distinctions the law draws.

124. For example, a wealthy person might better afford devices that make it easier to comply with a common law tort standard of care, yet this does not mean that the common law is unjust or that tort law must be made by the legislature rather than the courts. Indeed, even if the tort law should take greater account of wealth inequality, doing so does not entail a commitment to a politically accountable process.

125. See, e.g., Phyllis Tropper Baumann et al., Substance in the Shadow of Procedure: The Integration of Substantive and Procedural Law in Title VII Cases, 33 B.C. L. REv. 211, 301 (1992); Pamela J. Stephens, Manipulation of Procedural Rules in Pursuit of Substantive Goals: A Reconsideration of the Impermissible Collateral Attack Doctrine, 24 ARIZ. ST. L.J. 1109, 1131 (1993). I am not at all persuaded by the argument from bias. Although abuse of power is always a risk, I am aware of no convincing evidence that this is a serious problem in rulemaking. The rulemaking process has a number of safeguards against abuse, including a requirement of reasoned deliberation in a group setting and multiple levels of review. To believe that abuse is serious, one would have to be extremely skeptical about the good faith of judges and lawyers, and I am not willing to indulge this degree of skepticism without strong evidence. Of course, it is always possible that rulemakers suffer from subconscious bias that favors more privileged groups. See WEINSTEIN, supra note 88, at 7-8. Once again, however, reasoned deliberation and open discussion coupled with review and public justification should go a long way towards mitigating these effects.

126. Notice pleading requires only that a complaint state enough of the facts to put the defendant on notice of what the suit is about. See Conley v. Gibson, 355 U.S. 41, 47-48 (1957). Strict pleading, on the other hand, requires more factual detail and specificity. One of its primary purposes is to screen out frivolous suits early in the litigation. See Marcus, supra note 65, at 450-51. 127. To the extent a decision involves formal application of preexisting law, the concept of error is relatively straightforward. When courts make new law under a system of precedent, however, the concept is a bit more complex, since then it must distinguish between better and worse decisions. Nevertheless, the concept is meaningful in both settings, and the analysis given here applies to both.

128. This is especially important when information about potential liability is difficult for plaintiffs to obtain before filing.

129. For example, if false positives are much more serious than taise negatives, Rule B wout be superior to Rule A, so long as the frequency of false negatives is not too much greater than the frequency of false positives. In an efficiency analysis, we can represent the choice mathematically in the following way: choose the rule that minimizes ph + qk, where p and q are the probabilities and h and k the social costs of false positives and false negatives, respectively.

130. For an analysis of incentives to file frivolous suits and the resulting costs, see Robert G. Bone, Modeling Frivolous Suits, 145 U. PA. L. REv. 519 (1997).

131. The only way to balance false positives and false negatives without valuing substantive interests is to assume that the two kinds of error always create exactly the same harm. In fact, however, the law does not value all substantive interests the same. See infra notes 239-41 and accompanying text. 132. Although a strict pleading rule, insofar as it invites more motions to dismiss, can produce higher satellite litigation costs. See Bone, supra note 130, at 587-89 (analyzing the costs of a strict pleading rule).

133. See, e.g., D. J. GALLIGAN, DUE PROCESS AND FAIR PROCEDURES: A STUDY OF ADMINISTRATIVE PROCEDURES 100-02, 119-22 (1996); Ronald Dworkin, Principle, Policy, Procedure, in A MATTER OF PRINcIPLE 72 (1985).

134. See, e.g., Michael Bayles, Principles for Legal Procedure, 5 LAW & PHn.. 33, 41-57 (1986) (listing nine principles of procedure); Robert M. Cover, For James Wm. Moore: Some Reflections on a Reading of the Rules, 84 YALE L.J. 718, 721-22, 727 (1975) (referring to procedural principles).

135. My point here parallels the general argument in political and legal philosophy against processbased theories of political and constitutional legitimacy. See, e.g., Joshua Cohen, Pluralism and Proceduralism, 69 CHI.-KENT L. REV. 589 (1994); Lawrence G. Sager, The Incorrigible Constitution, 65 N.YU. L. REV. 893, 913-lS (1990). Process-based theories assume that legitimacy can be grounded in general agreement on the proper processes (such as voting and representation) for making social decisions. See also supra notes 69-70 and accompanying text. Critics of process-based theory argue that there is no fundamental distinction between procedural and substantive justice and that procedural justice necessarily requires reference to substantive norms. For example, critics insist that processbased theorists need a substantive ideal of fairness to decide which of the various modes of political representation allocate power in a fair way. See, e.g., Cohen, supra, at 602-03; Sager, supra, at 914.

136. See Stephen B. Burbank, Hold the Corks: A Comment on Paul Carrington ‘s “Substance”and “Procedure” in the Rules Enabling Act, 1989 DUKE L.J. 1012, 1039 (distinguishing “procedural policies” from “substantive policies”); Carrington, supra note 84, at 307-08 (deferring to the reference in Rule I of the Federal Rules of Civil Procedure to “just, speedy, and inexpensive” determinations). 137. See, e.g., John R. Allison, Ideology, Prejudgment, and Process Values, 28 NEW ENG. L. REv. 657, 671-82 (1994); see also SELF STUDY, supra note 5, at 692-95 (listing efficiency, fairness, simplicity, consensus, and uniformity).

138. The same is true of “legitimacy” understood in terms of whether people feel the system is legitimate, since feelings have no normative significance in and of themselves. See generally Robert G. Bone, Rethinking the “Day in Court” Ideal and Nonparty Preclusion, 67 N.Y.U. L. REv. 193, 233-35 (1992) (critiquing this notion of legitimacy).

139. One final point. Some readers might insist that substantive values are not an important aspect ot the rulemaking process for legitimacy purposes, because rulemaking aims directly at procedure not substance and involves substantive values only incidentally. This view is mistaken. The argument from substantive justification does not depend in any way on the “aim” of rulemaking; it depends only on the necessity of controversial value choice. Furthermore, it is a mistake to think that substantive values

are involved only incidentally in rulemaking. Substantive value is central to measuring error costs, and error costs are critical to procedural analysis. The main point of procedure, after all, is to enforce the substantive law and this means enforcing the values the substantive law protects. 140. The most prominent advocate of this view is Professor Paul Carrington, a distinguished procedure scholar and a former Reporter to the Civil Rules Advisory Committee. See Paul D. Carrington & Derek P. Apanovitch, The Constitutional Limits of Judicial Rulemaking: The Illegitimacy of Mass-Tort Settlements Negotiated Under Federal Rule 23, 39 ARIZ. L. REv. 461, 474-82 (1997); Carrington, supra note 87, at 2074-85; Carrington, supra note 84, at 301-10.

141. See, e.g., Carrington, supra note 84, at 282 (rulemaking is for “complex technical issues of judicial practice”). In a recent writing, Professor Carrington argues that rulemaking is proper only if it deals with matters relating to the internal operation of the federal courts in deciding Article III cases and controversies (as opposed, for example, to supervising settlement). See Carrington & Apanovitch, supra note 140, at 474-82.

142. See Carrington & Apanovitch, supra note 140, at 480-81. Indeed, Professor Carrington’s chief concern is to protect the rulemaking process from the blight of factional politics. See, e.g., id.; Carrington, supra note 87, at 2074-79.

143. For example, Professor Carrington treats mandatory disclosure as a purely technical matter proper for court rulemaking. See Carrington, supra note 25, at 54-55. Yet the 1993 amendments to Rule 26 triggered a firestorm of political controversy. See supra notes 107-08 and accompanying text. Moreover, Carrington believes rulemakers acted properly when they included the small claimant class action in Rule 23, see Carrington & Apanovitch, supra note 140, at 491, even though that device is closely tied to substantive interests and is highly controversial. See infra notes 265-72 and accompanying text.

144. See, e.g., Burbank, supra note 9, at 841-42; Laurens Walker, A Comprehensive Refoni for Federal Civil Rulemaking, 61 GEO. WASH. L. REv. 455 (1993); Laurens Walker, Avoiding Surprise From Federal Civil Rulemaking: The Role of Economic Analysis, 23 J. LEGAL STUD. 569 (1994).

145. See Bryant G. Garth, Observations on an Uncomfortable Relationship: Civil Procedure and Empirical Research, 49 ALA. L. REv. 103, 106-13 (1997) (noting the strong tendency to invest empirical work with the power to resolve controversy). 146. See id. at 113-17.

147. These include the absence of controls, the distorting influence of selection effects, the difficulty of obtaining data on settlement amount and other critical factors, and, of course, insufficient funding. See, e.g., Robert G. Bone, The Empirical Turn in Procedural Rulemaking: A Comment on Walker, 23 J. LEGAL STUD. 595, 609-11 (1994); Garth, supra note 145, at 106-13. 148. One example is the response of advocates for case management and alternative dispute resolution to the RAND Institute’s findings, based on its analysis of data collected under the Civil Justice Reform Act, that mandatory arbitration, mediation, and early neutral evaluation have little effect on time to disposition, attorney time, or litigation cost. See Carol M. Rice, The Civil Justice Reform Act Conference: A Reporter’s View, 49 ALA. L. REv. 265, 268-71 (1997).

149. A concrete example is the heated debate over the 1983 amendment to Rule 11 between civil rights proponents and those more concerned about the costs of frivolous litigation. See, e.g., Carl Tobias, Rule 11 and Civil Rights Litigation, 37 BUFF. L. REv. 485 (1989). 150. See, e.g., Burbank, supra note 90, at 235-49; Stempel, supra note 2, at 739-59; Subrin, supra note 4, at 54-55; see also SELF STUDY, supra note 5, at 694 (citing consensus as an important rulemaking value); Minutes of the Civil Rules Advisory Committee, May 1-2, 1997 available in 1997 WL 1056241 (report of Policy and Agenda Committee noting that “[t]he ideal rule change is one that is purely procedural, that `creates peace,’ and is satisfactory to all sides of a dispute”).

151. See Cary Coglianese, Assessing Consensus: The Promise and Performance of Negotiated Rulemaking, 46 DUKE L.J. 1255, 1271-1334 (1997) (reporting results that show that negotiated rulemaking has not achieved its intended purposes); Philip J. Harter, Fear of Commitment: An Affliction of Adolescents, 46 DuKE L.J. 1389,1400-1423 (1997) (taking a more favorable view toward negotiated rulemaking, but also noting its limitations). For more perspectives on negotiated rulemaking, see the other articles collected as part of the Twenty-Eighth Annual Administrative Law Issue, 46 DUKE L.J. 1255 (1997).

152. Professor Stempel instead advocates a civic republican vision of consensus in which rulemakers deliberate in an effort to work out a shared conception of the public interest. See Stempel, supra note 87, at 98-103. Although he is on the right track in anchoring rulemaking to reasoned deliberation rather than interest group accommodation, his references to “civic republicanism” are not helpful. We need a more definite theory of court rulemaking to provide the common ground from which a shared normative conception can grow. I develop such a theory in Part IV below. 153. See SELF-STUDY, supra note 5, at 694 (noting that consensus tends to favor the status quo).

154. See Daniel J. Meador, A Perspective on Change in the Litigation System, 49 ALA. L. REv. 7, IS (1997) (noting that “the history of judicial reform efforts over the last couple of decades does not provide much basis for optimism” that consensus can be achieved). Even advocates of the consensus approach concede these difficulties. See, e.g., Stephen N. Subrin, Uniformity in Procedural Rules and the Attributes of a Sound Procedural System: The Case for Presumptive Limits, 49 ALA. L. REV. 79, 84 (1997).

155. See Coglianese, supra note 151, at 1322-27 (noting that because of bargaining costs consensus rulemaking in administrative agencies is recommended only when the number of interested groups is small and that even then some interested parties are often excluded). In addition, a focus on consensus can intensify later conflict by inviting disputes about the meaning of the consensus reached. See id. at 1322-30. Some may view a failure of consensus as a positive sign, indicating that the particular matter should be left to legislation. But before we cede controversial procedural issues to the legislature, we ought to be confident that legislation is likely to produce good rules. A failure of consensus, by itself, is no reason for such confidence.

156. Significantly, the trend today seems to be in favor of highly general procedural rules that avoid policy choices and defer to the discretion of trial judges. See, e.g., Stephen B. Burbank, Of Rules and Discretion: The Supreme Court, Federal Rules and Common Law, 63 NOTRE DAME L. REV. 693, 715, 718-19 (1988). See generally Alan Schwartz & Robert E. Scott, The Political Economy of Private Legislatures, 143 U. PA. L. REv. 595 (1995) (arguing that the institutional and informational features of “private legislatures,” such as the ALl, favor adoption of general and vague rules). Another response to interest group conflict is to compromise within or across rule proposals. See Chemerinsky & Friedman, supra note 8, at 792 (noting disturbing trend toward political compromise); Stempel, supra note 87, at 73-77 (describing the 1993 mandatory disclosure provisions as a political compromise). 157. See, e.g., Stempel, supra note 87, at 86.

158. See Subrin, supra note 4, at 43-44 (questioning the legitimacy of relatively unfettered trial judge discretion).

159. Case management is extremely popular today, especially in the federal courts. See, e.g., Kesmk, supra note 24, at 203-04, 209-10; Subrin, supra note 4, at 43-44.

160. See Resnik, supra note 24, at 177 (noting that case management is appealing because it seems strictly “procedural” and thus a “superficially safe haven from partisanship”).

161. Part of the appeal comes from its association with some neo-pragmatist accounts of moral and legal reasoning that are increasingly popular today. See generally Symposium on the Renaissance of Pragmatism in American Legal Thought, 63 S. CAL. L. REv. 1569 (1990) (exploring pragmatist theories of law and adjudication).

162. Cf. Ronald M. Dworkin, Pragmatism, Right Answers, and True Banality, in PRAGMATISM IN LAw AND SOCIETY 359, 370-71 (Michael Brindt & William Weaver eds., 1991) (making the same point more generally as an argument against a pragmatic theory of law).

163. See, e.g., Robert D. Cooter & Daniel L. Rubinfeld, Economic Analysis of Disputes and Their Resolution, 27 J. ECON. LITERATURE 1067 (1989); Richard A. Posner, An Economic Approach to Legal Procedure and Judicial Administration, 2 J. LEGAL STUD. 399 (1973).

164. This is, of course, an application of the rational choice model of decisionmaking with uncertain outcomes, assuming risk neutrality. See RICHARD A. POSNER, ECONOMIC ANALYSIS OF LAW 12-13 (4th ed. 1992). In this model, parties choose the course of action with the highest “expected value,” defined as the outcome payoff multiplied by the probability that the outcome will materialize. 165. For example, a legislator concerned about risk-bearing costs might reduce the sanction by changing the substantive law while increasing the likelihood of enforcement by changing the procedural law. For an insightful discussion of these issues, see Louis Kaplow, The Value of Accuracy in Adjudication: An Economic Analysis, 23 J. LEGAL STUD. 307 (1994).

166. This is a controversial issue today, usually cast in terms of whether rules should be “transsubstan tive.” Compare Burbank, supra note 156, at 716-719 (recommending some substance-specific tailoring) and Subrin, supra note 4, at 45-56 (same), with Carrington, supra note 87, at 2079-81 (recommending against explicit substantive tailoring). If procedural rules should not vary at all with substance, then the desirability of a simple uniform system would follow automatically without the need to consider the merits of a default approach, and the only issue remaining would be the one addressed in the next section; that is, who should design the uniform system.

167. See Isaac Ehrlich & Richard A. Posner, An Economic Analysis of Legal Rule-Making, 3 J. LEGAL STUD. 257, 270, 273 (1974) (noting the importance of heterogeneity to rule specificity). If most procedural rules had to be customized, a default system would provide little benefit to justify the cost of creating and maintaining it.

168. For example, shifting from negligence to strict liability is likely to have a much stronger effect on incentives than shifting from strict to notice pleading. Procedure is technically complex, and individuals are likely to have difficulty predicting in advance of an actual dispute how different procedural rules will affect settlement and trial. Moreover, procedure affects ex ante incentives indirectly by affecting error risk, and this can blunt its regulatory effect. See Kaplow, supra note 165, at 312-21, 348-49.

169. See, e.g., Burbank, supra note 6, at 1469 (focusing on the complex suit); ub nn, supra note 4, at 45-47 (recommending offering more specific procedural requirements for the more discovery-prone types of cases).

170. See Subrin, supra note 4, at 28 (focusing on discovery rules). 171. The analysis I employ here is similar to the transaction cost approach to default rules in contract law. See, e.g., POSNER, supra note 164, at 92-94. 172. A rigorous efficiency analysis would have to consider a number of factors, including the incentive benefits of tailoring procedure to substance, the promulgation costs of substance-specific rules, and the relative advantages of clarifying rules ex ante as opposed to relying on trial judge discretion ex post. See Louis Kaplow, Rules Versus Standards: An Economic Analysis, 42 DUKE L.J. 557, 590-96 (1992).

173. Congress does this when it removes a substantive area from the courts and places it in some kind of administrative process. See, e.g., Black Lung Benefits Act, 30 U.S.C. 901 (1988). The classic example, though on the state level, is workers’ compensation.

174. See, e.g., Private Securities Litigation Reform Act of 1995, Pub. L. No. 104-67, 109 Stat. 737 (codified in various provisions of 15 U.S.C. 77 (Supp. 1997)).

175. For an overview of public choice theory, see JERRY L. MASHAW, GREED, CHAOS, AND GOVERNANCE (1997); Miller, supra note 71.

176. Campaign support is only one way interest groups influence legislators. Other ways include offering paid speaking engagements and future employment, and even indirectly manipulating the flow of information to lawmakers and the public. See, e.g., DAVID R. MAYHEW, CONGRESS: THE ELECTORAL CONNECTION 39-41 (1974).

177. For example, logrolling allows different groups to register intensity of preference. See, e.g., MASHAW, supra note 175, at 16-17.

178. See, e.g., DENNIS C. MUELLER, PUBLIC CHOICE II, at 83 (1989) (logrolling can produce inefficient results by imposing externalities on nontraders); MANCUR OLSON, THE LOGIC OF COLLECTIVE ACTION: PUBLIC GOODS AND THE THEORY OF GROUPS 127-28 (1965). See also MASHAW, supra note 175, at 17-18 (noting that majority decision rules can produce inefficiency by excluding organized minorities with intense preferences).

179. See supra Part Hc2. To be sure, trial lawyers have some common interests. But divisions between the plaintiffs’ and defendants’ bar over such matters as discovery, settlement, and summary judgment is often intense. See supra note 82 (listing some of the competing interest groups). Even the federal judiciary has its own lobbying organizations in the form of the Judicial Conference and the Federal Judicial Center. See Geyh, supra note 11, at 1173-74, 1178-79. 180. See OLSON, supra note 178, at 64-65.

181. See generally Geoffrey P. Miller, Some Agency Problems in Settlement, 16 J. LEGAL STUD. 189, 197-215 (1987) (analyzing the attorney-client conflict inherent in settlement). This is not to say that the interests of attorneys always conflict with the interests of plaintiffs. But there is a substantial enough risk of conflict to warrant serious concern. To cite just a few examples, attorneys for plaintiffs have interests that can conflict with the interests of their clients with respect to class actions, settlement rules, and in the case of a fee-for-services practice, rules regarding discovery.

182. See Jonathan R. Macey, Judicial Preferences, Public Choice, and the Rules of Procedure, 23 J. LEGAL STUD. 627 (1994).

183. See id. at 630-32. In this regard, Macey relies on Judge Richard Posner’s account of judicial incentives. See Richard A. Posner, What Do Judges and Justices Maximize (The Same Thing Everybody Else Does), 3 SuP. CT. EcoN. REv. 1 (1993). But see Ronald A. Cass, Judging: Norms and Incentives of Retrospective Decision-Making, 75 B.U. L. REv. 941 (1995) (arguing that professional norms and reputation incentives constrain judicial pursuit of self-interest).

184. See Macey, supra note 182, at 632-41. For example, Macey predicts rules that give trial judges broad control over settlement, because settlement increases the judge’s leisure time and can be used to weed out less interesting cases before trial. See id. at 634-35. 185. See id. at 641-43.

186. It has other problems too. As I discuss below, I believe that Macey exaggerates the role of self-interest in the rulemaking process. See infra note 195 and accompanying text. Moreover, he makes heroic assumptions about the ability of rulemakers to work out the indirect strategic effects of proposed rules. See Macey, supra note 182, at 635-36 (discussing Rule 68 proposal). For a critical analysis of Macey’s argument, see Janet Cooper Alexander, Judges ‘ Self-Interest and Procedural Rules: Comment on Macey, 23 J. LEGAL STUD. 647, 649-61 (1994).

187. Macey simply assumes that Congress is not likely to get involved in rulemaking because it has rarely done so in the past. See Macey, supra note 182, at 628. 188. See supra Part IIc2.

189. In this rulemaking game, each player’s behavior influences the behavior of all the other players. Solving a game like this is an extremely complex matter and beyond the scope of this Article. For an account of game theory, see ERIC RASMUSEN, GAMES AND INFORMATION: AN INTRODUCTION TO GAME THEORY (2d ed. 1994). On the importance of game theory to public choice more generally, see Miller, supra note 71.

190. I focus on a single mechanism of interest group influence-credible threats to go to Congressbecause the other mechanisms are likely to be ineffective in the court rulemaking context. For example, interest groups will not likely make much headway by implicitly promising future employment to federal judges, academics, and distinguished practitioners. Nor are rulemakers likely to worry much about maximizing their budget. Interest groups might be able to exert some influence by inviting judges to prestigious conferences in desirable locations, but I doubt the efficacy of this strategy. There is also a possibility that judges eager for promotion within the federal judiciary might curry favor with powerful groups having some influence over promotion decisions. However, promotion is contingent on so many factors that I doubt judges would invest much in pleasing groups through procedural reform.

191. Marshaling a majority coalition in both the House and the Senate and avoiding a Presidential veto are not easy. Moreover, there is a serious risk, especially for bar groups, of damaging valuable long term relationships with the judiciary. And Congress itself has an interest in avoiding friction with the federal judiciary. See, e.g., Geyh, supra note 11, at 1224-27 (noting cooperation between the federal judiciary and the Senate Judiciary Committee). In game theory terminology, the result of this interaction is probably a mixed strategy equilibrium, in which rulemakers sometimes (but not always) serve their own interests; powerful groups sometimes (but not always) respond by going to Congress, and Congress sometimes (but not always) intervenes. See generally RASMUSEN, supra note 189, at 67-83 (describing the concept of a mixed strategy equilibrium).

192. Also worth mentioning is that Arrow-type cycling problems might be easier to solve in a committee format where limited jurisdiction narrows the range of policy options and constrains the agenda. On cycling, see MASHAW, supra note 175, at 12-15.

193. For example, the plaintiffs’ bar typically prefers a very broad class action rule, while the defendants’ bar prefers a narrower one. Given the sharp difference of opinion on other procedural matters, there is not likely to be much room for bargaining. Some commentators also argue that confining the committee’s jurisdiction to general (transsubstantive) rules applicable to a broad range of cases can help reduce public choice problems. See, e.g., Carrington, supra note 87, at 2068. The idea is that general rules are less likely to invite the attention of interest groups than are rules focused on particular types of cases, where the stakes are clearer. However, it is not obvious that generality makes a big difference in this regard. Interest groups can still predict the differential effects of general rules. A civil rights group, for example, should be able to anticipate that a rule imposing severe sanctions for weak filings may have a disproportionate effect in chilling test cases aimed at creating new civil rights law, and the plaintiffs’ tort bar is likely to anticipate that a rule limiting discovery will have a disproportionate effect on tort plaintiffs with limited access to information. Indeed, much of the interest group controversy in the past has focused on the anticipated differential effects of general rules. In addition, the more general the rule, the greater the number of interest groups affected and thus the more intense the interest group pressure may be. Finally, even if general rules reduce the risk of public choice problems, this benefit must be balanced against the benefit of tailoring rules more closely to substantive stakes. Thus, at least from the perspective of efficiency, there is no reason to be categorically committed to highly general rules.

194. See supra notes 156-57 and accompanying text.

195. See generally Cass, supra note 183, at 970-81 (arguing that reputation incentives and internal

ized norms can constrain judges); Geyh, supra note 11, at 1214-16 (criticizing the self-interest model of rulemaking and arguing that the process selects for rulemakers with a public regarding ethic).

196. See Burbank, supra note 90, at 248-49 (recommending negotiation of a new rulemaking “treaty” between the judiciary and Congress). Without a coherent view of the proper role of courts in rulemaking, the outcome of any such negotiation will simply reflect a compromise of interests and provide no assurance that the resulting process will yield good rules.

197. The argument from public choice might suggest that the legislature should have no role at all in procedural rulemaking, since public choice problems will distort substance-specific exceptions as well as default rules. But this goes too far. Congress has comparative informational advantages in coordinating specific rules with substantive objectives, and the bulk of lawmaking power in our democracy lies in Congress despite public choice problems. A default system might even reduce inefficient logrolling in Congress by confining the policy space over which trades can be made and increasing the costs for groups seeking a legislative exception. And because substance-specific exceptions operate over a more limited domain than a default system, their efficiency costs are likely to be lower.

198. See, e.g., Burbank, supra note 156, at 715; Carrington, supra note 99, at 943; Resnik, supra note 24, at 197-204; Subrin, supra note 4, at 43-44.

199. And because the goal is social (i.e., global) efficiency, the judge must take account of the external effects of his rule choices on other cases (such as adding marginally to overall case backlog and delay).

200. Indeed, one federal trial judge has pointed out how a tendency to split the diterence in resolving discovery disputes can encourage parties to insist on extreme positions and litigate them vigorously. See Robert E. Keeton, The Function of Local Rules and the Tension With Uniformity, 50 U. PIT. L. REv. 853, 867 (1989). 201. See RASMUSEN, supra note 189, at 16-18.

202. One possible cause other than the Prisoners Dilemma involves agency costs. Lawyers hired on a fee-for-services basis have an incentive to engage in excessive discovery when doing so increases fees. However, this explanation is weak to the extent the lawyer and client are in a repeat-play situation and the client can monitor performance by reviewing the bill. Concern about malpractice liability for

overlooking evidence is another possible factor, especially for risk-averse lawyers worried about reputation effects. However, this explanation depends on the probability that a damaging piece of evidence will become public and that the client will learn of its existence, neither of which is terribly likely in the ordinary lawsuit. Yet another reason might have to do with the fact that discovery is usually left in the hands of young lawyers who have little experience and strong ambitions to become firm partners. These lawyers are likely to take the safer route and use broad, boilerplate discovery requests that cover all conceivable possibilities. Whatever the force of these alternative explanations, one thing is clear. Like the Prisoners’ Dilemma, they are likely to respond to centrally promulgated rules imposing clear and strict discovery limits.

203. See Ronald J. Gilson & Robert H. Mnookin, Disputing Through Agents: Cooperation and Conflict Between Lawyers in Litigation, 94 COLUM. L. REV. 509, 514-22 (1994); John K. Setear, The Barrister and the Bomb: The Dynamics of Cooperation, Nuclear Deterrence, and Discovery Abuse. 69 B.U. L. REv. 569 (1989).

204. See, e.g., Burbank, supra note 90, at 238 (noting that federal trial judges “do not react well to rules that limit their discretion”); Resnik, supra note 24, at 199-200 (noting that federal trial judges are “strongly attached to their own individual authority” to manage cases). The recent RAND Institute study of the Civil Justice Reform Act data found that federal trial judges resisted formal case-tracking procedures in favor of individual discretionary case management. See, e.g., James S. Kakalik et al., Just, Speedy, and Inexpensive?: An Evaluation of Judicial Case Management Under the Civil Justice Reform Act, 49 ALA. L. REv. 17, 28, 40, 48 (1997).

205. See Carrington, supra note 25, at 65 (noting that federal judges are reluctant to punish lawyers for discovery abuse); Stempel, supra note 87, at 85 (noting the historical reluctance of federal judges to use Rule 37 sanctions). Indeed, the law of discovery gives trial judges extremely broad discretion at the sanctioning stage. See 8A CHARLES A. WRIGHT ET AL., FEDERAL PRACTICE AND PROCEDURE 2 289, at 614 (2d ed. 1994). And the law sometimes explicitly requires an ex post balancing of actual prejudice and justification. See, e.g., In re Exxon Valdez, 102 F.3d 429, 433 (9th Cir. 1996) (using a balancing test to decide on a dismissal sanction).

206. In fact, judges might change their behavior even in a discretionary system once they understand the benefits of a stricter approach. However, this is much more likely in a system with centralized ex ante rules than one that relies on case-specific discretion.

207. See, e.g., Chris William Sanchirico, The Burden of Proof in Civil Litigation: A Simple Model of Mechanism Design, 17 INT’L REV. L. & ECON. 431, 442 (1997). In the language of game theory, clear and strictly binding rules allow for sequential equilibria that are not subgame perfect. See generally RASMUSEN, supra note 189, at 93-106 (discussing subgame perfection). 208. Many frivolous suits involve a failure to investigate. See Bone, supra note 130, at 550 (noting examples). In these cases, one can justify the hypothetical rule on efficiency grounds. A more forgiving rule that excused plaintiffs whose suits later proved meritorious would encourage plaintiffs to gamble rather than investigate and would also add costs by requiring determinations of merit. See generally id. at 591-93 (discussing penalties as a device to deter frivolous suits).

209. And as a fringe benefit, the judge might not have to impose sanctions on meritorious suits very often, since a credible threat will encourage plaintiffs to investigate before filing.

210. Deciding on the optimal balance between rule and discretion is a complex matter that has received extensive treatment in the literature. See David L. Shapiro, Class Actions: The Class As Party and Client, 73 NOTRE DAME L. REv. 913, 946 n.89 (1998) (collecting some of the sources on rules and standards). My goal here is not to give a comprehensive analysis. My point is that the costs of judicial discretion in the litigation arena are likely to be substantially higher and the benefits significantly lower than commonly supposed.

211. In addition, presumptive rules that set clear limits in advance but allow trial judges to grant exceptions in individual cases might also be inefficient. See Subrin, supra note 4, at 47 (recommending substance-specific discovery rules that set presumptive limits). Parties will litigate for exceptions, and the uncertainty will encourage strategic behavior and undermine the judge’s ability to use credible threats. One might make it difficult to obtain exceptions by requiring a compelling demonstration of need. See id. at 48. But in complex cases, where the problems are most severe, there are likely to be numerous ways to argue for compelling need and strong incentives to make the arguments. Therefore, it appears on this rough cut that strict rules could be superior from an efficiency perspective despite the costs of under- and over-inclusion. 212. See infra notes 273-74 and accompanying text. 213. See supra notes 22-26 and accompanying text.

214. Much of this discussion has been critical of broad local rulemaking power. See Carrington, supra note 99, at 944-52; Chemerinsky & Friedman., supra note 8, at 780-91; Robel, supra note 26, at 1483-85. But see Flanders, supra note 3, at 261-73 (arguing for local rulemaking); Keeton, supra note 200, at 874-75 (noting the advantages of local rules).

215. Although, some commentators have raised questions about even these ministerial rules. See Carrington, supra note 99, at 946; Chemerinsky & Friedman, supra note 8, at 785, 787.

216. See, e.g., Flanders, supra note 3, at 263-64.

217. See Kobayashi, et al., supra note 3, at 24-25, 29 (arguing that local rulemaking can lead to the spontaneous emergence of efficient rules); cf. Larry E. Ribstein & Bruce H. Kobayashi, An Economic Analysis of Uniform State Laws, 25 J. LEGAL STUD. 131, 140-41 (1996) (noting how substantive law uniformity chills local innovation and experimentation).

218. See Chemerinsky & Friedman, supra note 8, at 783-86; Robel, supra note 26, at 1484. In fact, we know very little about local legal culture and its effect on procedural rules. See, e.g., Thomas W. Church, Jr., Civil Case Delay in State Trial Courts, 4 JUST. SYS. J. 166 (1978) (finding strong effects due to local legal culture); Herbert M. Kritzer & Frances Kahn Zemans, Local Legal Culture and the Control of Litigation, 27 LAW & Soc’Y REv. 535 (1993) (finding weak effects of local legal culture on variation in use of Rule 11).

219. See Chemerinsky & Friedman, supra note 8, at 770, 789-91. Experience with the Civil Justice Reform Act “experiments” is indicative. Lack of controls and failure to obtain time-longitudinal data significantly undermined the value of the results. See Kakalik et al., supra note 204, at 47-49 (noting the limitations).

220. See Kobayashi et al., supra note 3, at 24-25, 29.

221. See infra notes 224-27 and accompanying text. Indeed, it is not even clear that procedural efficiency benefits a local district court, as efficient rules will likely attract more cases and increase the burden on the public fisc.

222. See, e.g., Chemerinsky & Friedman, supra note 8, at 783. 223. The data to some extent bears out this prediction. See LOCAL RULES PROJECT, supra note 25, pt. 1, at 1-5 (noting the large number and detailed nature of local rules). 224. See, e.g., Daniel A. Farber, Environmental Federalism in a Global Economy, 83 VA. L. REv. 1283, 1300-06 (1997) (analyzing “race to the bottom” theory); Richard L. Revesz, Rehabilitating Interstate Competition: Rethinking the “Race-to-the-Bottom” Rationale for Federal Environmental Regulation, 67 N.Y.U. L. REv. 1210, 1244 (1992) (same).

225. Moreover, transfer opportunities are limited by the strong presumption in favor of plaintiffs forum choice. See 15 WRIGHT ET AL., supra note 205, 3848, at 375-77. 226. See, e.g., Securities Exchange Act of 1934, 15 U.S.C. 78aa (1994); Clayton Act, 15 U.S.C. 22 (1994).

227. Even under nationally uniform federal rules, there is still some risk of an inefficient race to the bottom between state and federal courts, given the prevalence of concurrent jurisdiction and the availability of removal.

228. Professor Ronald Dworkin has provided the most detailed and rigorous account of a rights theory of adjudication. See RONALD DWORKIN, LAW’S EMPIRE (1986) [hereinafter DWORKIN, LAW’S EMPIRE]; RONALD DWORKIN, TAKING RIGHTS SERIOUSLY (1977) [hereinafter DWORKIN, TRS]. 229. So long as doing so does not impair other rights of comparable moral worth. 230. See, e.g., DWORKIN, TRS, supra note 228, at 90-100; THOMAS NAGEL, MORTAL QUESTIONS 112-15,131-33 (1979).

231. See Dworkin, supra note 133, at 78; see also GALLIGAN, supra note 133, at 100-02.

232. See GALLIGAN, supra note 133, at 114-19. 233. See Dworkin, supra note 133, at 92-93.

234. Briefly, the excuse approach needs a theory of excuses, since it risks collapsing rights into utility if it excuses mistakes whenever society has done its best to maximize overall welfare. The fair treatment approach also runs into problems, for it depends on a distinction between substance and procedure that is hard to defend. If substantive rights are only as good as their enforcement, then presumably procedure should be considered part of the substantive right. But then the substantive right itself would be qualified, and mistakes-for example, in convicting an innocent person-would be no cause for moral complaint. On the latter objection, see GALLIGAN, supra note 133, at 117-19. 235. See id. at 119 (“if some risk [of error] is unavoidable, then it must be distributed fairly amongst those who might be affected”); Dworkin, supra note 133, at 92-93 (civil litigants have a right to a consistent evaluation of the moral harm from error in the procedures afforded them as compared to the procedures afforded others in different civil cases).

236. Cf. Peter H. Aranson et al., A Theory of Legislative Delegation, 68 CORNELL L. REv. 1, 50-51 (1982) (using this to explain FDA decisions to concentrate more on the harm from bad drugs than the harm from delaying good drugs).

237. For example, a frivolous suit that wins big at trial is more likely to capture the headlines than a meritorious suit that is not filed or loses. Of course, there are notable exceptions to this proposition, such as perhaps the O.J. Simpson trial if, as some people assume, Simpson was in fact guilty. 238. See GALL[GAN, supra note 133, at 119-22 (referring to this as the principle of proportionality, that treating people with equal concern and respect means tailoring procedural protection to the relative importance of substantive rights); Dworkin, supra note 133, at 80-84, 89, 92 (assuming that the moral character of the substantive interest affects the degree of moral harm due to error and thus the relative distribution of process resources).

239. Even on such a view, procedure might still have to vary with substance in order to equalize error risk between complex and simple cases-between, for instance, factually and legally complex antitrust and patent suits and relatively simple breach of contract actions.

240. 376 U.S. 254 (1964).

241. See also Gomez v. Toledo, 446 U.S. 635, 638-42 (1980) (holding that state defendants in a 42 U.S.C. 1983 case must allege good faith to invoke the qualified immunity defense because plaintiffs are at an informational disadvantage and Congress meant to favor constitutional claims). Similarly, the strict pleading requirements that Rule 9(b) of the Federal Rules of Civil Procedure imposes on fraud cases in part reflect special concern for the reputation and other substantive harms associated with fraud allegations. See Marcus, supra note 65, at 448. Moreover, Rule 23(b)(2), though framed in general terms, was designed to facilitate civil rights class actions, an objective that in part reflects the special value assigned to constitutional rights. See Benjamin Kaplan, Continuing Work of the Civil Committee: 1966 Amendments of the Federal Rules of Civil Procedure fl), 81 HARV. L. REV. 356, 389 (1967); see also YEAZELL, supra note 61, at 247-48, 264-65 (suggesting that the liberal provisions of (b)(2) reflect the strong moral claim of civil rights protection). Indeed, most state procedural systems direct minor cases to small claims courts with relatively slim procedural protections. See JOHN C. RUHNKA & STEVEN WELLER, SMALL CLAIMS COURTS: A NATIONAL EXAMINATION 1-2 (Nat’l Ctr. for State Cts. 1978). 242. See, e.g., GALLIGAN, supra note 133, at 120 (“how different rights are weighted should be determined by examining the law as a whole”); Dworkin, supra note 133, at 93 (procedural rights give a right to “the consistent application of that theory of moral harm [from error] that figures in the best justification of settled legal practice”).

243. This does not necessarily mean that procedural rules will vary greatly with the substantive interests at stake. For one thing, there are costs to designing substance-specific procedure, and these costs are relevant to the scope of procedural rights claims. Furthermore, the value framework might well end up treating most interests identically either because the legal system actually treats them the same or because epistemic limitations render the distinctions opaque. Still, rulemakers must go through the analysis to determine the relative values, and the nature of this analysis favors a court-based, committee-centered rulemaking process.

244. The idea that judges make decisions in individual cases by reference to principles inferred from precedent through a process of interpretation is, of course, widely shared. See, e.g., DWORKIN, LAW’S EMPIRE, supra note 228, at 225-75 (arguing that judges aspire to principled consistency at a high level); CASS R. SUNSTEIN, LEGAL REASONING AND POLITICAL CONFLICT 13-34, 44-46, 53-61 (1996) (arguing that courts use lower-level principles extracted from precedent often by the use of analogy); Fuller, supra note 54, at 372-78 (giving an account of adjudication as a gradual process of developing principles through precedent). See generally Larry Alexander & Ken Kress, Against Legal Principles, 82 IOwA L. REv. 739, 749 (1997) (“The Dworkinian judicial technique of surveying the past decisions in an area of law and distilling a legal principle from those decisions that is not (necessarily) the stated ground for any of them is so commonplace as to be the standard of orthodoxy.”). 245. Some readers might object to the common law analogy on the ground that the legitimacy of the common law depends on features, such as a focus on individual cases and a retrospective quality, not shared by rulemaking. However, the view of court rulemaking that I defend in this article is quite similar to common law adjudication. For more detail, see infra note 288 and accompanying text.

246. In a rights-based theory, procedural rules must be designed as an integrated system. The system as a whole generates and distributes error risk, and all the individual elements interact in complex ways to produce the overall effect. See Dworkin, supra note 133, at 91. 247. See id. at 93. But see GALLIGAN, supra note 133, at 121-22 (imposing stronger constraints). The legislature has broad power to create substantive rights and to prescribe the procedures for their adjudication, and it can also specify the weight assigned to substantive interests protected by a statutory right.

248. See, e.g., Robert S. Summers, Evaluating and Improving Legal Processes-A Plea for “Process Values”, 60 CORNELL L. REv. I (1974). But see Larry Alexander, Are Procedural Rights Derivative Substantive Rights?, 17 LAw & PHIL. 19 (1998) (arguing against process-based rights distinct from substantive rights). Here I refer to process-based rights and duties that are specific to the institution of adjudication. No one disagrees that generally applicable requirements of background morality also apply to adjudication. For example, the state cannot torture a party in order to extract information for a court proceeding anymore than the state can torture citizens in other settings. One might justify this prohibition on outcome grounds, but those who favor a deontological view believe that torture is wrong even if it yields truthful admissions. 249. See Bone, supra note 138, at 279-85. 250. For an extensive discussion, see generally id. 251. See FED. R. CIv. P. 24(a).

252. See id. 23(b)(1)(B).

253. See Robert G. Bone, Statistical Adjudication: Rights, Justice, and Utility in a World of Process Scarcity, 46 VAND. L. REv. 561, 634-50 (1993) (justifying limitations on participation rights implicit in

case sampling from a rights-based and efficiency perspective); Bone, supra note 138, at 203-32, 260-79 (inferring participation limits from precedent).

254. By “principle” in this context, I mean a moral principle of general application that has weight and can be balanced against other principles in an overall analysis. These principles include utilitarian moral principles associated with the efficiency metric, as well as rights-based moral principles associated with the rights-based and process-based metrics. Thus, my approach differs from Professor Dworkin’s in that my approach allows for utilitarian principles. 255. See DWORKIN, LAW’S EMPtRE, supra note 228, at 49-113. 256. See JoHN RAWLS, A THEORY OF JUSTICE 20-22, 48-53 (1971).

257. For example, our interpreter might start by identifying principles that roughly fit the main features of the practice, and then examine those principles to determine whether they justify the practice in the best possible light. A principle that seems to fit the practice might be objectionable on moral grounds. In such a case, the interpreter must consider whether to modify the principle. Before she modifies it, however, she must check the fit between the principle as modified and the actual practice that the principle is meant to interpret. She should not modify the principle if the practice features that the principle fits are central to the integrity of the practice. For if she did so, then the principle would no longer be an interpretation of that practice.

258. This method is a version of what in philosophical circles is called “constructivism” because it aims to construct principles from practice using a standard of coherence or fit. See DwORKIN, TRS, supra note 228, at 159-68. For an example of the constructivist method applied to constitutional interpretation, see Richard H. Fallon, Jr., A Constructivist Coherence Theon, of Constitutional Interpretation, 100 HARV. L. REv. 1189 ( 1987). 259. See DwORKIN, TRS, supra note 228, at 105.

260. For an ambitious effort to develop procedural principles that fit and justify practice using a constructivist methodology, see generally Lawrence B. Solum, Procedural Justice (1998) (unpublished manuscript on file with author).

261. It bears recalling that rights-based and process-based metrics privilege current practice to some extent, so anyone who believes these metrics are relevant, even partially, has a reason to accept the constructivist approach for some areas of rulemaking. It is also worth mentioning that there are potential problems with the Dworkinian requirement of fit. See, e.g., Alexander & Kress, supra note 244, at 785-86 (arguing for direct application of the best moral principles rather than legal principles that diverge from the best morality due to the requirement of fit). But even those who believe that rulemakers ought to apply the best moral principles unhampered by a fit requirement have reason to support court rulemaking insofar as the deliberative process of a committee is better suited than legislation to moral inquiry. 262. See Bone, supra note 253.

263. See Solum, supra note 260, at 14-19 (making a similar argument by relying on the Rawlsian ideal of public reason).

264. See infra Part VI (discussing the limits of court rulemaking). 265. On small claimant class actions generally, see Jonathan R. Macey & Geoffrey P. Miller, The Plaintiffs ‘Attorney’s Role in Class Actions and Derivative Litigation: Economic Analysis and Recommendations for Reform, 58 U. CHI. L. REv. 1 (1991).

266. See, e.g., Kline v. Coldwell, Banker & Co., 508 F.2d 226, 237-38 (9th Cir. 1974) (Duniway, J., concurring); Report of the Section of Antitrust Law, 110 A.B.A. Rep. 905, 906-09 (1985). This controversy surfaced again recently in connection with the Private Securities Litigation Reform Act of 1995, Pub. L. No. 104-67, 109 Stat. 737, as well as a proposed Rule 23 amendment considered by the Advisory Committee in 1997. See 167 ER.D. 523, 559 (1996) (proposed Rule 23(b)(3)(F)).

267. The Supreme Court held in Eisen v. Carlisle & Jacquelin, 417 U.S. 156, 173-75 (1974), that individual notice and opt-out were always required even though the additional cost might scuttle small claimant class actions. However, insofar as this decision is based on Rule 23 rather than the Constitution, it is no impediment to rulemakers deciding differently. But see Phillips Petroleum Co. v. Shutts, 472 U.S. 797, 812 (1985) (holding that individual notice is constitutionally required, but basing its decision on personal jurisdiction principles).

268. See also Shapiro, supra note 210, at 924-25, 956 (arguing that the small claimant class should be viewed as an entity rather than as an aggregation of individuals with strong individual rights to participate).

269. For example, cases with very small stakes are usually handled in small claims court with relatively informal procedures and limited participation opportunities. See supra note 241. 270. See Macey & Miller, supra note 265, at 101.

271. See 7B WRIGHT ET AL., supra note 205, 1781. However, to avoid over-enforcement, rulemakers should also ask whether the legislature anticipated the small claims problem, and if it did, whether it contemplated private enforcement or instead adjusted other aspects of the statute to compensate for the enforcement deficiency.

272. See, e.g., John C. Coffee, Jr., Understanding the Plaintiff’s Attorney: The Implications of Economic Theory for Private Enforcement of Law Through Class and Derivative Actions, 86 COLUM. L. REv. 669 (1986) (discussing the role of private attorneys general in various situations). For those readers who are still not persuaded, consider another example: the issue of how much to limit discovery. See supra note 211 (discussing the efficiency of presumptive versus mandatory discovery limits). Rulemakers need not be concerned about process-based participation rights in this context. No plausible account of current practice norms would sustain a participation right guaranteeing expansive discovery. The value of discovery is usually measured in terms of outcome quality-better information produces a more accurate result-rather than in terms of dignitary participation. It is certainly possible that a rule eliminating all or almost all discovery could implicate process-based values on the theory that participation is meaningless when access to most of the evidence is denied. However, none of the proposals for limiting discovery go this far.

Thus, rulemakers should focus on the two outcome-based metrics. In this connection, it is important to distinguish between two different approaches to limiting discovery. One approach aims to solve the Prisoners’ Dilemma and achieve a Pareto result for both parties without increasing the error risk for either (or perhaps even reducing the error risk by freeing up resources for more productive litigation investments). This first approach is trivially consistent with a rights-based metric insofar as it leaves the error risk unchanged, and it furthers efficiency goals insofar as it reduces the social costs of discovery.

The second approach imposes stricter limits on discovery in order to save more on litigation costs. Rulemakers using this approach face a more difficult task. They must justify both the tradeoff between process costs and error costs and the resulting distribution of error risk. This tradeoff can be justified within the efficiency metric; indeed, efficiency is likely to be the motivating reason for stricter limits. But the rights-based metric presents a more difficult challenge, requiring careful attention to how current practice balances error and process cost. Even so, there are numerous examples of rules in our procedural system that tolerate greater error for the benefit of reducing process costs. Rulemakers would have to construct from these examples a principled account of the acceptable limits to the process-cost/error-cost tradeoff. Yet there is no obvious reason to think that such an account could not make room for some version of the stricter approach.

273. See supra notes 201-09 and accompanying text. In addition, some latitude for local district experimentation might also be justified on informational input grounds similar to those discussed in the text.

274. Case-specific decisions would also be more valuable for rulemakers if those decisions were subjected to a loose process of stare decisis and perhaps also to more exacting appellate review. These additional features would add costs, of course, but those costs might well be justified by the additional informational benefits that stare decisis and appellate review provide. 275. It is not a sufficient answer to cite complaints about the formality of common law procedure in the late nineteenth and early twentieth century when ideas and beliefs were quite different than they are today. See Bone, supra note 38, at 78-104 (noting how the shift from a formalistic to a pragmatic view of procedure influenced criticisms of common law procedure).

276. In a useful recent discussion of this issue, Professor David Shapiro identifies some additional factors that favor a centralized rulemaking process (and perhaps even a legislative one) for major procedural reform. See Shapiro, supra note 210, at 947-50 (arguing that a centralized process facilitates open debate; that the common law is not well suited to reforms which implicate “fundamental values;” and that the Erie doctrine limits the utility of a common law process in federal court).

277. To be sure, substantive common law rules also interact. A firm deciding how much to invest in pollution control equipment, for example, would consider rules of nuisance and negligence, as well as contract rules if the decision might affect the firm’s existing contractual obligations (in addition, of course, to relevant environmental statutes and regulations). However, these interdependency effects are too complex and varied to be easily anticipated or considered at the time of making common law rules. The substantive common law regulates a multitude of distinct transactions and events that vary in their essential characteristics more than lawsuits do.

278. See George L. Priest & Benjamin Klein, The Selection of Disputes for Litigation, 13 J. LeGAL STUD. 1 (1984) (describing case selection effects). Moreover, the common law process is not completely free from public choice problems, since organized groups are better able to litigate for favorable law. See Einer R. Elhauge, Does Interest Group Theory Justify More Intrusive Judicial Review?, 101 YALE L.J. 31, 66-87 (1991). It is also worth noting that my view of court rulemaking is path dependent to some extent, since institutional history influences the selection of rules. But these adverse effects should be much weaker than those associated with case-by-case common law adjudication.

279. For a useful discussion of how to accommodate conflicting values in the form of hybrid principles, see Paul H. Robinson, Hybrid Principles for the Distribution of Criminal Sanctions, 82 Nw. U. L. Rev. 19, 28-40 (1987).

280. The claim is at least plausible. For example, if the wealthier party is likely to invest more in litigation, the procedural system might deny relief to a poor person with a meritorious claim or defense more often than a wealthy person. Furthermore, because a wealthier party can credibly threaten to invest more if the case does not settle, and also has a superior capacity to bear risk and delay, wealthier parties might be able to extract settlements more favorable than their substantive entitlements.

281. This is hardly self-evident. For example, the availability of contingency fees and fee shifting in some cases can compensate for wealth disparity. Moreover, liberal joinder and class action rules provide opportunities to spread litigation costs, and plaintiffs in mass tort cases can pool resources by hiring the same attorney or organizing groups to share information. See Deborah R. Hensler, Resolving Mass Toxic Torts: Myths and Realities, 1989 U. ILL. L. REV. 89, 96-97 (noting the prevalence of multiple representation); Joseph Sanders, The Bendectin Litigation: A Case Study in the Life Cycle of

Mass Torts, 43 HASTINGs L.J. 301, 309, 354 (1992) (noting the tendency of attorneys to cooperate in complex suits).

282. To condemn accurate procedure on this ground, one must be prepared to argue that procedure can be used to circumvent the substantive law whenever doing so serves the ends of justice. If such a thing is ever proper, it surely must be so only in the most extreme situations. For a thoughtful discussion of this issue, see Cover, supra note 134, at 721-31. 283. See supra Part IVB.

284. For example, the process-oriented argument from injustice relies on the idea that people consent to the unjust effects through a political process. But consent presumes choice, and choice is not meaningful when no feasible alternative exists.

285. See Alyeska Pipeline Serv. Co. v. Wilderness Soc’y, 421 U.S. 240 (1975). Moreover, contingency fees and other methods of litigation financing fall within the exclusive domain of state legislatures, state bar organizations, and state courts with authority over matters of professional responsibility.

286. For instance, adverse selection and common pool problems are likely to produce settlements well below expected value for those high-end plaintiffs with large claims. See, e.g., John C. Coffee, Jr., The Regulation of Entrepreneurial Litigation: Balancing Fairness and Efficiency in the Large Class Action, 54 U. CHI. L. REv. 877, 916-17 (1987). Similarly, because poor plaintiffs are not likely to have the resources to monitor attorneys effectively, transferring risk to large law firms through a contingency fee arrangement can generate agency costs that adversely affect the amount of recovery. See Miller,

supra note 181, at 198-202. To use another example, alternative dispute resolution speeds up litigation, which may benefit poor plaintiffs who suffer from delay. But according to some at least, it also runs the risk of perpetuating skewed outcomes by reinforcing the power imbalance between rich and poor. See, e.g., Eric K. Yamamoto, Efficiency’s Threat to the Value of Accessible Courts for Minorities, 25 HARV. C.R.-C.L. L. REv. 341, 360-61 (1990).

287. See Marcus, supra note 81, at 773-76 (defending a “neutralist perspective” against the view that procedure is political, and defining neutrality as “an honest attempt to fashion rules that will fairly accommodate the concerns of accuracy, participation and efficiency”).

288. Of course, the common law analogy is not perfect. Indeed, it is helpful to my argument only if procedural rulemaking is similar enough to a common law process to share in the latter’s legitimacy. Whether this is so depends on a theory of what makes the common law legitimate. Yet most such

theories focus on some combination of principled reasoning, case-specific decisionmaking, and retrospectivity; and court rulemaking shares each of these features to a significant extent. But see Christopher J. Peters, Adjudication as Representation, 97 COLUM. L. REv. 312, 346-78 (1997) (arguing that the legitimacy of the common law can be based on legislative-type norms of broad participation and interest representation).

289. See, e.g., 19 WRIGHT ET AL., supra note 205, 4509; Burbank, supra note 136, at 1019-20, 1031-1036; Carrington, supra note 84, at 297-321; John Hart Ely, The Irrepressible Myth of Erie, 87 HARV. L. REv. 693, 718-38 (1974); Shapiro, supra note 210, at 952-57; Olin Guy Wellborn III, The Federal Rules of Evidence and the Application of State Law in the Federal Court, 55 Tax. L. REv. 371, 402-06 (1977); Peter Westen & Jeffrey S. Lehman, Is There Life for Erie After the Death of Diversit, 78 MICH. L. REv. 311, 361-64 (1980).

290. The Act authorizes the Supreme Court to “prescribe general rules of practice and procedure” so long as those rules do not “abridge, enlarge or modify any substantive right.” 28 U.S.C. 2072(a)-(b) (1994).

291. Burbank, supra note 29, at 1114; see also id. at 1128 (predictably and identifiably affect substantive rights or create remedial rights that predictably and identifiably affect personal liberty or the use and enjoyment of property). In 1988, Congress reenacted the substantive rights limitation, but despite some discussion in a House Report, added little to further clarify or operationalize the separation-of-powers principles that the limitation was meant to express. See generally Burbank, supra note 136, at 1030-36 (discussing the 1988 amendments and House Report).

292. 480 U.S. 1 (1987). Sibbach v. Wilson & Co., 312 U.S. 1 (1941), made the first stab at a test-whether the rule “really regulates procedure”-and Sibbach’s test was repeated in Hanna v. Plumer, 380 U.S. 460 (1965), notwithstanding its obvious tautology. For other cases in which the Court has upheld Rules of Civil Procedure against challenges based on the Rules Enabling Act proviso, see Business Guides v. Chromatic Communications Enterprises, 498 U.S. 533, 551-54 (1991) (Rule 11); Schlagenhauf v. Holder, 379 U.S. 104, 112-14 (1964) (Rule 35(a)); and Mississippi Publishing Corp. v. Murphree, 326 U.S. 438, 445-46 (1946) (Rule 4(f)). 293. 480 U.S. at 5.

294. Some commentators rely exclusively on a rule’s aim, see, e.g., Ely, supra note 289, at 724-25 (stating that a rule is procedural if it is “designed to make the process of litigation a fair and efficient mechanism for the resolution of disputes”); while others focus more on effects, see, e.g., Burbank, supra note 136, at 1019-20 (noting that only Congress has power to make rules that “necessarily and obviously involve policy choices with a predictable and identifiable impact on the rights claimed under substantive law”); see also Hanna v. Plumer, 380 U.S. 460, 475 (1965) (Harlan, J., concurring) (stating the test for both Rules Enabling Act and Erie purposes as whether the rule “would substantially affect . . . primary decisions respecting human conduct”). The Supreme Court’s analysis in Business Guides v. Chromatic Communications Enterprises, 498 U.S. 533, 551-54 (1991) illustrates the problem. In Business Guides, the Court applied the Burlington Northern test to uphold Federal Rule 11 even though the sanction involved fee-shifting. The Court argued that Rule 11 ‘s purpose was procedural-to deter baseless filings and litigation abuses-and that its structure and effect were not the same as a straightforward fee-shifting rule. See id. at 552-53. But there are problems with this analysis. See id. at 564-69 (Kennedy, J., dissenting) (discussing some of the problems). If purpose or aim controls, then rulemakers should be able to adopt a general loser-pays rule so long as its purpose or aim is to deter frivolous suits. On the other hand, if effects control, then it is not clear why fee-shifting under Rule II is consistent with the Rules Enabling Act. The Court seems to think it matters that Rule I1 is not tied to outcome, but the Court does not clearly explain why this fact is significant. Nor is it clear how the Court’s argument would fare if judges routinely imposed fee-shifting sanctions. See Melissa L. Nelken, Has the Chancellor Shot Himself in the Foot? Looking for a Middle Ground on Rule II Sanctions, 41 HASTINGS L.J. 383, 399 (1990) (observing that fee-shifting dominated sanctions under 1983 Rule 11).

295. Professor Carrington anchors his test of rule validity to an explicitly articulated theory of court rulemaking. See Carrington, supra note 84, at 308. The problem with Carrington’s approach, however, lies in his conception of court rulemaking. He supposes, for example, that rulemakers can fashion sufficiently general rules to avoid political controversy, see id. (noting that a rule does not “abridge, enlarge, or modify” a substantive right if it is sufficiently broad so as not to evoke political controversy), and as we have seen, that view assumes a distinction between substance and procedure that cannot be sustained. See supra notes 140-43 and accompanying text (critiquing Carrington’s view).

296. See supra Parts IVA, IVB, and IVc 297. See supra Part IVD.

298. See supra Part IVD; cf. Business Guides., 498 U.S. at 565 (Kennedy, J., dissenting) (“the further our rules depart from our traditional practices, the more troubling becomes the question of our rulemaking authority”).

299. For example, Rule 23 (the federal class action rule) is not suspect merely because it has “substantive consequences” or implicates substantive value choices-as some have suggested. See, e.g., Minutes of Civil Rules Advisory Committee, March 20-21, 1997 available in 1997 WL 1056240 (noting that Rule 23 raises Enabling Act concerns because it has substantive consequences). Indeed, it is even possible that a suitably limited rule authorizing settlement class actions could pass muster under my justification test. (This does not necessarily mean, of course, that such a rule would be desirable as a policy matter.) The third requirement, that the rule fit a coherent account of existing practice, is likely to create difficulties for a broad rule, especially one that authorizes judicial implementation of settlements concluded prior to filing suit. There is a serious question whether such a rule is even consistent with Article III’s case-and-controversy requirement. Assuming that this hurdle can be overcome, however, rulemakers must still justify a settlement class action rule in terms of its effect on process-based participation rights and the distribution of error risk. While I cannot engage the complex issues here, it is worth noting that justification might be possible in special cases. Cf. Bone, supra note 253 (justifying sampling as an aggregation device consistent with a best interpretation of current practice norms under conditions of extreme scarcity). One thing is clear, however. The analysis of validity under my test is sure to be more nuanced than the somewhat extreme views advanced by some commentators today. See, e.g., Carrington & Apanovitch, supra note 140, at 462-64 (arguing against validity, but without taking adequate account of the benefits of the settlement class action for class members and the society at large); Linda S. Mullenix, The Constitutionality of the Proposed Rule 23 Class Action Amendments, 39 ARtz. L. REv. 615, 628 (1997) (arguing for the validity of a settlement class action rule on the ground that it simply authorizes joinder and can be applied by trial judges without consulting

substantive values, but ignoring the fact that validity turns on whether the rulemakers themselves can justify the rule with all its more problematic substantive-type effects).

300. Thus, the fact that the 1983 version of Rule 11 (imposing strict sanctions for frivolous suits) was meant to-and did in fact-restrict access to federal court does not, in and of itself, present any rulemaking problems, contrary to what some commentators seem to suggest. See Burbank, supra note 9, at 844. In a recent discussion of limits on court rulemaking, Professor David Shapiro proposes a balancing test for determining the validity of rules that raise “significant barriers” to the assertion of claims and defenses. See Shapiro, supra note 210, at 956-57. His test would allow such a rule only if it is “clearly warranted by the need to run a judicial system efficiently and fairly” and has only a marginal effect on the assertion of just claims. Id. Professor Shapiro’s test is similar to mine in focusing on a rule’s justification and it has the desirable property of permitting some rules with adverse effects on litigating ability. Still, the test is too restrictive. For example, running a judicial system efficiently might well require a strict pleading rule that substantially impedes the filing of just claims, and fairness might require discovery rules that also create opportunities for strategic imposition of costs with much more than a marginal effect on the ability of resource-poor litigants to litigate.

301. In fact, some of the current Federal Rules, such as Rules 9 and 23, draw claim-specihc distinctions expressly or implicitly, and many more direct trial judges to take account of factors, such as complexity, that correlate strongly with the substantive nature of the case. Indeed, unless court rulemakers have power to adopt substance-specific rules, they could not require strict pleading requirements for particular claims, contrary to what the Supreme Court assumed in Leatherman v. Tarrant County Narcotics Intelligence and Coordination Unit, 507 U.S. 163, If 9$ (1993).

302. A useful analogy is the commentary that accompanies federal sentencing guidelines adopted by the Sentencing Commission. See Stinson v. United States, 508 U.S. 36 (1993). The Committee might also consider using hypothetical illustrations along the lines of those in ALI Restatements, coupled with clear explanations of the reasoning as well as the result. Illustrations would be particularly useful as paradigms to guide trial judges in the exercise of their case-specific discretion.

303. The three requirements of my test are necessary conditions for valid rulemaking, but are they also sufficient? The answer may be no for a state claim if federalism values are also strongly implicated. But what about more generally? To see the problem, consider an amendment to Federal Rule of Civil Procedure 68 that would allow two-way fee shifting in the event a party who rejects a settlement offer

does not improve on the offer at trial. See, e.g., Stephen B. Burbank, Proposals to Amend Rule 68-7ime to Abandon Ship, 19 U. MICH. J.L. REFORM 425 (1986) (discussing such a proposed amendment, which the Advisory Committee considered in the early 1980s). On my test, the rule’s validity depends on the nature of its justification. The problem, however, is that a fee-shifting rule, even one conditioned on settlement and intended to serve procedural goals, directly adds to recovery and so may just seem too “substantive” on its face. But I can see no reason to invalidate the rule on this ground alone. Indeed, such a view would strip judges of any power to impose fee-shifting sanctions for litigation abuse, contrary to accepted doctrine. See Business Guides v. Chromatic Communications Enters., 498 U.S. 533 (1990). At the same time, however, the justification for the rule must take account of all its consequences, including effects on the availability of relief, and this might bar application of the rule in certain circumstances. See Burbank, supra, at 430-40 (arguing that an amended Rule 68 would interfere with fee-shifting statutes for civil rights suits). Those who still insist that a fee-shifting rule is just too “substantive” on its face must explain why, and must do so in a way that generalizes to all possible cases, recognizing that every procedural rule affects outcome in some way.


* Professor of Law, Boston University School of Law. I am grateful for input received on an earlier draft from Boston University School of Law, Columbia Law School, and Western New England College of Law faculty workshops. I especially wish to thank Ron Cass, Mike Dorf, Fred Lawrence, Gerry Lynch, Rick Marcus, Tom Rowe, Larry Sager, David Shapiro, and Allan Stein for their helpful comments, and Patty Washienko and David Schur for their invaluable research assistance

Copyright Georgetown University Law Center Apr 1999

Provided by ProQuest Information and Learning Company. All rights Reserved