Unintended Consequences of Mandatory Disclosure Samuel Issacharoff* & George Loewenstein** The recentlyadoptedmandatory disclosurerule has sparked a contentious debate among the members of the judiciary and academia. Professors Issacharoff and Loewenstein take a fresh look at the wisdom of mandatory disclosure by focusing on whether or not the rule is effective in performing itsprimary task-decreasingthe cost of litigation. The Article assesses the likely consequences of mandatory disclosure using both an economic approach and an assessment of the potential litigation strategies created by the new rule. The authors begin by establishing a formal economic model of settlement to examine the impact of mandatory disclosure on the timing and frequency of settlement. The authors then modify their model by considering various aspects ofpsychology that challenge the traditionalassumptions of economic theory. Specifically, the authors note that persons involved in litigation negotiatein a stepwise manner (taking each step as it comes), ratherthan engagingin backward induction. Furthermore, rather than reach convergent expectations regarding the outcome of litigation, litigants tend to process information in a selfserving manner that diverges the expectations of the parties. Finally, whereas many economic models assume thatpartiesignoresunk costs, litigantsinfact tend to become desensitized to further losses. The authors conclude that mandatory disclosure will have the opposite of its intended effect. Because it increases the up-front costs of litigation, cases that would ordinarilysettle without discovery have increasedsystematic costs. These costs are not ameliorated by any savings under mandatory disclosure because the cost of filing routine discovery is not substantial. Furthermore,the increasedup-front costs, rather than decreasing the amount of suits, will likely give greater vitality to strike suits because the defendantis the party that bears the costs of early discovery. Finally, the authors note that early discovery will probably not promote settlement because the disclosed information is likely to be subject to self-serving judgments that may block efficient settlements.

* Charles Tilford McCormick Professor of Law, The University of Texas School of Law. The authors thank Lisa Bernstein, Theodore Eisenberg, Douglas Laycock, Geoffrey Miller, Daniel Rubinfeld, Steven Shavell, Carl Tobias, Patrick Woolley, and Adrian Zuckerman for helpful comments on drafts of this Article. The authors are, of course, solely responsible for the positions taken in this Article. We also thank Vanessa Clem, Jon Dyck, and Patricia Saydah for valuable research assistance. This Article was originally presented at the John M. Olin Conference on the Law and Economics of Litigation Reform held at the Georgetown University Law Center on October 28-29, 1994. We would like to thank Warren Schwartz and the Conference participants for their comments as well. ** Professor of Economics, Department of Social and Decision Sciences, Carnegie Mellon University.

754

Texas Law Review

[Vol. 73:753

I.

Introduction ................................

II.

Effects on Settlement ........................... 758 A. Bifurcation of the Discovery Period .............. 761 B. The Asymmetrical Distributionof Disclosure Costs .... 768 1. Strike Suits ............................ 769 2. PsychologicalFactors .................... 770 C. Escalation............................... 772 D. The Effect of Disclosure on Parties'Expectations ..... 772 E. Summary ................................778

I.

Litigation Strategies ............................

754

IV. Conclusion .................................

I.

779

785

Introduction

The adoption of mandatory disclosure under amended Rule 26 of the Federal Rules of Civil Procedure inaugurated a sweeping regime of reform

aimed at curtailing litigation costs.

Despite widespread (indeed almost

complete) opposition, 1 and without the benefit of any systematic review of

the experience in pilot jurisdictions,2 the amended procedures that went into effect on December 1, 1993 mark a significant alteration of the

litigation process.3 Almost immediately following the initiation of a lawsuit, and without any formal requests, the new procedural rules require

each party to provide the other with large categories of information. The central target of the reforms are the costs associated with discovery.' Because an estimated ninety-five percent of cases in the 1. See, e.g., Griffin B. Bell et al., Automatic Disclosurein Discovery-The Rush to Reform, 27 GA. L. REV. 1, 29 nn.l10 & 111 (1992) (recounting nearly unanimous opposition to mandatory disclosure when the proposal was first presented by the Federal Rules Advisory Committee). 2. See Amendments to the Federal Rules of Civil Procedure, 146 F.R.D. 402, 511 (1993) (Scalia, J., dissenting) ("It seems ... most imprudent to embrace such a radical alteration that has not... been subjected to any significant testing on the local level.'); Carl Tobias, In Defense of Experimentationwith Automatic Disclosure,27 GA. L. REV. 665, 668-69 (1993) (advising against the adoption of mandatory disclosure prior to significant experimentation). See generally Linda S. Mullenix, Hope Over Experience: MandatoryInformalDiscovery and the Politics of Rulemaking, 69 N.C. L. REV. 795, 814-19, 828-30 (1991) (reviewing mixed early experiences in pilot districts and proposals for the systematic study of discovery costs and discovery reform). 3. For an overview of the curious history leading to the amendment of Rule 26 to provide for mandatory disclosure, see Carl Tobias, Improving the 1988 and 1990 JudicialImprovements Act, 46 STAN. L. REv. 1589, 1611-14 (1994). 4. See PRESIDENT'S COUNCIL ON COMPETITIVENESS, AGENDA FOR CIVIL JUSTICE REFORM IN AMERICA 7 (1991) (reporting the recommendations of a committee chaired by Vice President Dan

1995]

Mandatory Disclosure

federal system are resolved either through settlement or otherwise prior to

trial,5 the likeliest sources of excessive expense are those incurred prior

to trial, and the costliest element of pretrial practice is discovery. Thus, discovery is a prime target for curtailing the costs of adjudication. 6 The disclosure requirement seeks to reduce the costs of discovery by eliminating the need for formal requests that had, prior to the reform, preceded all exchanges of information. As expressed by the Rules Advisory Committee, "A major purpose of the revision is to accelerate the exchange of basic information about the case and to eliminate the paper work involved in requesting such information. "I Proponents of mandatory disclosure also

claim that it is likely to limit discovery abuse' by fostering a less combative style of litigation in general and of discovery in particular.9 The primary practical alterations under the new rule are twofold.

First, the amended rule eliminates the need for filing formal discovery requests for the information that is now covered under mandatory disclosure. The categories of information subject to mandatory disclosure are quite broad. At the outset of litigation, each side is responsible for

Quayle to lower the costs of discovery); Bell et al., supra note 1, at 2 n.3 (summarizing the concerns over discovery costs). 5. See Marc Galanter & Mia Cahill, "Most Cases Settle': JudicialPromotion and Regulation of Settlements, 46 STAN. L. REv. 1339, 1339, 1387 (1994) (repeating the 95% figure for pretrial resolution of disputes and observing that the component of those cases resolved by settlement is increasing). 6. See Linda S. Mullenix, Discovery in Disarray: The Pervasive Myth of Pervasive Discovery Abuse and the Consequencesfor Unfounded Rulemaking, 46 STAN. L. REv. 1393, 1395 (1994) ("If there is a keystone in the arch of civil litigation, it is the pretrial discovery process. The media's constant repetition of the theme of American litigiousness has made discovery a prime target for reformers."). 7. FED. R. Civ. P. 26(a) advisory committee's note. 8. The concept of "discovery abuse" is itself quite complicated. In general, there is far greater cost in complying with a discovery request than in making the discovery request. As a result, there is a strong temptation to inflict harm on one's adversary by seeking additional information for which the adversary will have to incur the cost. This creates what economists would term a "moral hazards" problem in which a party is absolved of the cost consequences of its conduct. See generallyTHE NEW PALGRAVE: A DICTIONARY OF EcONOMics 549 (John Eatwell et al. eds., 1987) (defining "moral hazard" in economic terms). Thus, some commentators have tried to define abusive discovery requests in terms of the relative costs imposed compared to the benefits obtained from additional information. See, e.g., Robert D. Cooter & Daniel L. Rubinfeld, An Economic Model of Legal Discovery, 23 J. LEGAL STUD. 435, 450 (1994) (defining discovery abuse as knowingly making discovery requests with which "compliance [will cost] more than the expected increase in the value of the requesting party's claim"). On the other hand, discovery as a part of litigation produces public benefits in terms of deterring wrongful conduct that may reach beyond the immediate costs to the parties, thereby justifying a more generous discovery regime. See Bruce L. Hay, 23 J. LEGAL STUD. 481, 506-10 (1994) (comparing privately versus socially desirable levels of discovery). 9. See, e.g., William W Schwarzer, The FederalRules, the Adversary Process, and Discovery Reform, 50 U. Prrr. L. REV. 703, 722 (1989) (advocating a "prompt disclosure" requirement to eliminate "game playing" and to prevent "harassment or obstruction in the disclosure phase of discovery"). For a dramatic challenge to the empirical bases of claims of discovery abuse as part of a generalized litigation explosion, see Mullenix, supra note 6, at 1397-1418.

Texas Law Review

[Vol. 73:753

identifying the names of all individuals with relevant information and the content of the information they possess, copies or descriptions of all documents bearing on the subject of the litigation, a computation of all damages suffered, and copies of all insurance agreements that may underwrite all or part of an eventual judgment." While there is no evidence that the costs associated with filing the largely boilerplate requests for such information are a particularly significant source of litigation expenditures, presumably the reform does mildly reduce the costs of requesting the preliminary information necessary for litigation to advance.

Second, the new rule advances the time in which discovery information must be produced and, consequently, accelerates the costs associated with the early stages of litigation. All disclosures are to be made within ten days of a preliminary conference of the parties," which is now required to be held "as soon as practicable" after the commencement of the litigation. 2 A failure to disclose witnesses or information available to a party at the opening salvoes of litigation may result in the preclusion of the use of such witnesses or information at trial. 3 The threat of evidence preclusion will almost inevitably push litigants to accelerate the production of information beyond what would have occurred prior to the reform.1 This Article examines the likely consequences of the procedural changes associated with Rule 26(a). Our goal is to engage the rather

contentious debate over the wisdom of mandatory disclosure that currently engages many of the federal courts around the country. 15 We do so by 10. FED. R. Civ. P. 26(a)(1). 11. Id. 12. FED. R. Civ. P. 26(f). 13. See FED. R. Civ. P. 37(c)(1) ("A party that without substantial justification fails to disclose information ... shall not, unless such failure is harmless, be permitted to use as evidence at a trial, at a hearing, or on a motion any witness or information not so disclosed."). 14. For a discussion of the incentives for parties to produce favorable information in order to favorably alter the settlement demands of their adversaries, see Steven Shavell, Sharing ofInformation Priorto Settlement or Litigation, 20 RAND J. ECON. 183 (1989). 15. One survey found that 57 of the 94judicial districts had opted out of the mandatory disclosure requirements of amended Rule 26. Memorandum from Kathleen L. Blaner, Pepper, Hamilton & Scheetz, to Alfred W. Cortese, Jr., Pepper, Hamilton & Scheetz 3 (Oct. 28, 1994) (on file with the Texas Law Review). Of the 37 districts that implemented the rule, 18 limited its application to only certain classes of cases. Id. at 6. Another national survey showed similar results. See Bruce H. Kobayashi et al., The Process of Procedural Reform: Centralized Uniformity Versus Local Experimentation (Oct. 17, 1994) (unpublished manuscript, on file with the Texas Law Review). Among the four judicial districts in Texas, for example, there is no uniformity of practice. Only the Southern District has implemented amended Rule 26 without modification. Blaner, supra, at 6-8. The Northern District has delegated authority to mandate disclosure to individual judges, and the Western District has adopted its own mandatory disclosure plan. Id. at 3-5. The Eastern District employs a hybrid six-track plan that is applied selectively by each judge in the district. UNITED STATES DIsT. COURT FOR THE E. DIsT. OF TEX., CIVIL JUSTICE EXPENSE AND DELAY REDUCTION PLAN 1 (1991). As the Texas experience suggests, it is difficult to get a comprehensive picture of the current state of practice because some jurisdictions have formulated their local discovery rules through local

1995]

Mandatory Disclosure

relying primarily on formal models of litigation behavior drawn from three areas: economics, psychology, and what may be called strategic theory. Although the combination of methods may be unconventional, each provides an analytic tool to grapple with the problem of the incentives created by the procedural reform of information production. Undoubtedly, some readers may find discrete parts of our analysis to stray from their particular concerns. Nonetheless, broad legal reforms not based on proven successes in generating their claimed benefits should be subjected to as robust an analytic examination as possible. Toward this end, we turn to analytic techniques drawn from social and behavioral sciences. We begin by using a formal economic model of settlement to analyze the likely consequences of the procedural reforms on the timing and frequency of settlement. We show that the introduction of mandatory discovery creates two potential periods for settlement of the dispute when previously there was only one, but that the incentives for settlement in each of these periods are likely to be smaller than those that prevailed in the single period under the old procedural rules. Based on this model, we show that under a wide range of conditions, mandatory disclosure is likely to alter patterns of settlement in a manner that increases rather than decreases litigation costs. In Part I, we examine the strategic implications for litigation that may result from the addition of mandatory disclosure. This Part of the Article will focus on the further unintended consequences issuing from the addition of the new Rule 26 to the litigation arsenal. In addition to identifying the likely consequences of amended Rule 26, this Article seeks to demonstrate the potential for enriching the economic analysis of legal disputes with insights from psychology. 16 In the recent past, legal scholars have turned increasingly to economic analyses in their attempts to predict the impact of proposed rule changes or to provide a theoretical framework for examining the effect of existing rules.17 These efforts bring a systematic, rigorous analytic framework to problems with

rules, while others have done so through standing orders of the court, civil justice reform plans, or other orders or informal practices of individual judges. 16. For a recent discussion of the application of experimentally derived behavioral models to the economic analysis of tax law, see Edward J. McCaffery, Cognitive Theory and Tax, 41 UCLA L. REV. 1861, 1862-74 (1994). 17. See, e.g., Richard A. Posner, An Economic Approach to Legal Procedure and Judicial Administration, 2 J. LEaAL STUD. 399, 421-29 (1973) (arguing that the Federal Rules of Civil Procedure, by allowing more extensive pretrial discovery, have changed the economic incentives of the parties to a lawsuit); Robert D. Cooter & Daniel L. Rubinfeld, Economic Analysi ofLegal Disputes and Their Resolution, 27 J. ECON. Lr'rATuRE 1067 (1989) (surveying law and economic literature); Susan D. Caorle, Note, A HazardousMix: Discretion to Disclose andIncentives to Suppress Under OSHA's Hazard Communication Standard, 97 YAs LEJ. 581, 585 (1988) ("Although economic analysis cannot illuminate many important aspects of policy decisions, it adds to our understanding of how institutions work and helps to predict the feasibility and effects of proposed changes in legal

rules.").

Texas Law Review

[Vol. 73:753

wide-ranging social and economic ramifications. However, many of these analyses make unrealistic assumptions about how people reason and behave in the real-world setting of litigation. Economic analyses of strategic behavior have had only modest success at predicting behavior, even in the most stylized settings such as the repeated prisoner's dilemma 8 or the shrinking pie game.19 Similarly,

experimental simulations of legal cases have produced a variety of anomalies that cannot be explained by standard models of litigant behavior in the law and economics literature.2' We bring psychological insights to bear on the effects of mandatory disclosure in part to show why economic models of litigant behavior are insufficient to predict the likely effects of changes in legal rules.

II. Effects on Settlement In assessing the effect of mandatory disclosure, and specifically its

success in terms of reducing litigation costs, it is necessary to examine its impact on the timing and prevalence of case settlement. The importance of settlement in federal litigation cannot be gainsaid. Somewhere on the order of ninety to ninety-five percent of the cases filed in federal court terminate prior to trial, a large portion through settlement.2 1 In addition, the most extensive surveys of litigation practices estimate that half of the cases fied in the federal court system conclude without any discovery at all,' and another quarter are resolved with only token discovery.' This 18. See R. Selten & R. Stoecker, End Behavior in Sequences of Finite Prisoners' Dilemma Supergames, 7 J. ORGANIZATIONAL BEHAV. 47-70 (1986). 19. See Jack Ochs & Alvin E. Roth, An ExperimentalStudy of SequentialBargaining,AM. ECON. REv., June 1989, at 355, 355-59 (1989) (observing the widely varying results of experimental studies of sequential bargaining problems). 20. See, e.g., George Loewenstein, Samuel Issacharoff, Colin Camerer & Linda Babcock, SelfServing Assessments of Fairnessand PretrialBargaining, 22 J. LEGAL STUD. 135, 153-55 (1992) (demonstrating the inefficient impasses in settlement resulting from the biased incorporation of common information); Linda R. Stanley & Don L. Coursey, EmpiricalEvidence on the SelectionHypothesis and the Decisionto Litigate or Settle, 19 J. LEGAL STUD. 145, 161-64 (1990) (discussing the way in which litigants' strategic behavior, expectations, and psychological motivations may cause a departure from that behavior predicted in experimental models). 21. In the 1990 statistical year, only 8.5% of the cases terminated in federal district court were terminated by trial. ADMINISTRATIVE OFFICE OF THE U.S. COURTS, 1990 FEDERAL COURT MANAGEMENT STATISTICS 167 (1990). Roughly the same figure appears to hold in state court litigation. See David B. Rotman, Tort Litigationin the State Courts: Evidencefrom the Trial Court Information Network, STATE CT. J., Fall 1990, at 4, 6 (reporting the results of a study of state court tort litigation that found a pretrial resolution rate of over 95%). 22. E.g., David M. Trubek et al., The Costs of OrdinaryLitigation, 31 UCLA L. REV. 72, 90 (1983) (citing PAUL R. CONNOLLY ET AL., FEDERAL JUDICIAL CTR., JUDICIAL CONTROLS AND THE CIVIL LITIGATIvE PROCESS: DISCOVERY 29 (1978)). The same pattern holds in state court litigation. See Susan Keilitz et al., Is Civil Discovery in State Trial Courts Out of Control?, 17 STATE CT. J. 8, 10 (1993) (finding that 42% of cases in trial courts of 4 states settled with no formal discovery). 23. C. Trubek et al., supra note 22, at 90 ("Rarely did the records reveal more than five separate discovery events.").

19951

Mandatory Disclosure

759

leaves only twenty-five percent of cases filed in which there is any significant discovery at all-of which only a fraction falls into the "abusive" or "exorbitantly expensive" category. Other, more informal estimates suggest that two-thirds to three-quarters of cases are resolved without extensive discovery.' The net effect of mandatory disclosure on litigation costs will depend not only on its impact on discovery costs when discovery occurs, but also on the rate of settlement prior to discovery. Any effect of mandatory disclosure on rates of settlement prior to discovery will have a major impact on such costs. If mandatory disclosure reduces the settlement rate, causing more cases to proceed through discovery to trial, then the rule will increase net litigation costs. The likely effect of mandatory disclosure on settlement can be assessed by comparing stylized representations of settlement negotiations both before and after amended Rule 26(a). Prior to mandatory disclosure, as illustrated in Figure 1, there were two major points at which the parties could settle: prior to discovery, and after discovery but prior to trial. Figure 1

to negotiate before any Thus, the parties had a single period in which discovery. in Figure 2, the first resourcesas through illustrated depletion of their aggregate disclosure, prior to Under mandatory settlement periods: one two into "bifurcated" settlement period is but prior to discovery. disclosure, and the other following disclosure Figure 2

discovery e that two-thirds eoit to estimate or n sanders's Barefoot Judgelittle 1, at 41 (quoting note probably supraases, more, have al., civil 24. See Bellofetour indicating that less studies to three-fourths 6, at 1434-35 (summarizing Mullenix, supra note requests). mattersU). See generally discovery than 10 involv eo than 10% of all ases

Texas Law Review

[Vol. 73:753

The consequences of this seemingly minor change in procedure can be separated into four components that are discussed separately below: 1) The division of the discovery process into pre- and postmandatory disclosure periods will make a failure to reach agreement in each of the new bargaining periods created by mandatory disclosure less costly to the parties than was a failure to settle in the unified bargaining period that prevailed prior to mandatory disclosure. In the first of the bifurcated periods, the costs of failing to settle are those associated with disclosure that will be incurred under amended Rule 26(a). In the second period, the costs of disclosure have already been incurred. The costs of failing to settle at this point are those associated with the remaining portions of discovery-which are lower than pre-amended Rule 26(a) because the disclosure costs have already been incurred. Because the central perceived incentive for settlement at any stage of litigation is the cost associated with moving to the next stage, bifurcating the settlement zone is likely to affect the rate and timing of settlement. 2) The costs consequent on failing to settle during the first of these bargaining periods-the costs of mandatory disclosure itself-will be borne almost exclusively by the defendant. This asymmetry will affect the plaintiff's willingness to initiate suits, and especially "strike suits," which are suits of limited economic viability brought exclusively for the purpose of negotiating a settlement out of the amount that the defendants would have had to expend in defense of 25 the litigation.

3) The relatively low costs associated with disclosure and the asymmetric burden of such costs will produce low settlement rates prior to disclosure. Thus, earnest negotiations will often begin only following disclosure-after defendants have already incurred costs. This front-loading of costs may influence subsequent settlement rates if the parties fail to treat costs associated with disclosure as "sunk," that is, as irrelevant to their subsequent decisionmaking. 4) The initial revelation of information associated with mandatory disclosure may produce a convergence or divergence of the parties' expectations concerning the likely outcome of proceeding to trial. Again, such convergence or divergence is likely to influence the rate of settlement prior to discovery. In tracing out the consequences that follow from each of these effects, we contrast the likely rate of settlement in the presence and absence of

25. See Villiam M. Lafferty & W. Leighton Lord III, Towards a Relaxed Summary Judgment Standardfor the Delaware Courts of Chancery: A New Weapon Against 'Strike" Suits, 15 DEL. J. CoRP. LAW 921, 923 (1990) (describing strike suits as "blackmail by litigation").

1995]

Mandatory Disclosure

mandatory disclosure. We take as our point of departure a conventional economic analysis of settlement. However, we modify some of the assumptions of litigant behavior implicit in economic models to consider the implications of three psychological factors that are not part of standard economic analyses. First, there are important limitations in human information-processing abilities that prevent people from "solving"

bargaining problems in the manner prescribed by economics. Instead of engaging in a process of backward induction (as called for by conventional game-theoretic analyses'), there is evidence that people proceed through negotiations in a more stepwise fashion, taking each stage as it comes.

Second, contrary to economic prescriptions to ignore sunk costs, parties who incur costs in early stages of a dispute are likely to become desensitized to further losses and thus to become more intransigent. Finally, disputants process information in a self-serving fashion such that

information sharing may well produce a divergence rather than convergence of expectations. Even in the absence of these psychological factors, our economic analysis suggests that mandatory disclosure will most likely

increase rather than decrease costs. Each of the psychological factors we examine only reinforces this conclusion. A. Bifurcation of the Discovery Period The main effect of mandatory disclosure is to bifurcate what has been conventionally a single period of bargaining prior to discovery into two smaller bargaining periods, one preceding and the other following the

mandatory disclosure of information.

Because, in our view, the main

incentive for settling at any stage of litigation is the cost associated with proceeding to the next stage of litigation, the bifurcation of discovery costs

introduced by mandatory disclosure is likely to affect the prevalence and timing of settlement and hence the litigation costs incurred by the two parties.27

26. E.g., R. Selten, Reexaminationof the Perfectness ConceptforEquilibriumPoints in Extensive Games, 4 INT'L J. GAME THEORY 25 (1975). 27. Modeling the two-stage bargaining process created by prediscovery and pretrial negotiations (which we use to depict the practice in the absence of mandatory disclosure), and the three-stage process created by predisclosure, prediscovery, and pretrial negotiations, is extremely complicated. Numerous and often controversial simplifying assumptions are required to make the analysis tractable. Key among these is the question of what parties expect to happen in the next phase if they fail to reach agreement in the current one. One possible assumption is that parties expect to settle in the next phase of litigation at their estimated award value if they fail to settle in the current one. This assumption is clearly unrealistic because even extremely overoptimistic litigants will almost certainly acknowledge that there is some risk of going to trial. A second possible assumption would be that parties expect to not settle in the next phase if they fail to settle in the current one. This errs on the side of pessimism, because settlement prior to trial is the rule rather than the exception. A third would be that parties calculate the conditional probability of not settling in the next phase based on having not settled in the

Texas Law Review

[Vol. 73:753

The basic question raised by the bifurcation of costs is whether two settlement periods that each offer weaker incentives for settlement are more likely to produce settlement and control costs than one settlement period with a commensurately larger settlement zone.' To analyze this question, we adopt a modified version of the economic analysis of settlement first advanced by Professors George Priest and Benjamin Klein. The PriestKlein hypothesis attempted to resolve the question of why cases proceed to trial because the joint wealth of the parties would necessarily be dissipated by litigation expenses. They posited that cases that do not settle are those

in which the parties lack sufficient information about either law or facts, such that they are unable to agree on the likelihood of the plaintiff prevailing or the potential award amount, or those in which parties err as to their estimate of the likely outcome of the case.29 Consistent with

Priest-Klein, we assume that parties estimate the likely outcome of the case, A, with some degree of error, and that parties always settle when there is a positive settlement zone-when there are sufficiently convergent estimates of the expected value of the plaintiff's claim that allow the parties to divide the surplus created by anticipated litigation costs and to increase their aggregate wealth.' ° However, contrary to Priest-Klein, but consistent with our own prior research, we allow for the possibility that parties

may have divergent expectations of the outcome of litigation not simply as a result of random deviations but rather as a result of biased expectations concerning the likely outcome of the case. 3 Our own previous research

current one. While entailing the highest level of rationality of the three assumptions, the third may be the least realisticbecause it is predicated on an unmanageably high degree of calculation. It requires a litigant to estimate the adversary's expectations and the likelihood of settlement in the next period as a function of both parties' award expectations, the next period's costs, and other considerations, and then to factor these together in some fashion into an estimate of settlement probability. For simplicity, and because we believe it is the closest approximation of reality, we make the first of these assumptions. 28. The formal analysis of settlement based on cost avoidance does not take account of the short duration of the settlement period prior to mandatory disclosure and thus probably overstates the likelihood that the parties will settle during this period. Mandatory disclosure significantly compresses the initial period in which the parties may bargain. Because of the short time frame for producing information that must be disclosed, and the severe sanctions for its nondisclosure, the process of information production must begin almost immediately, unless there is quick agreement that the case will indeed be settled. 29. In their seminal work, Professors Priest and Klein sought to identify the conditions under which cases would proceed to trial despite the aggregate reduction in wealth of parties to litigation. George L. Priest & Benjamin Klein, The Selection of Disputesfor Litigation, 13 J. LEGAL STUD. 1 (1984); see also George L. Priest, Reexamining the Selection Hypothesis: Learningfrom Wittman's Mistakes, 14 J. LEGAL STUD. 215 (1985). 30. Priest, supra note 29, at 216-21. 31. We set out this thesis in two articles based on experimental observationsof parties to simulated litigation settings. Linda Babcock, George Loewenstein, Samuel Issacharoff& Colin Camerer, Biased Judgments of Fairnessin Bargaining,Am. ECON. REV. (forthcoming 1995); Loewenstein, Issacharoff, Camerer & Babcock, supra note 20.

1995]

Mandatory Disclosure

points to an "egocentric bias" in legal disputes-so that each party forms exaggeratedly optimistic expectations concerning the likely outcome of an adjudicated settlement. 2 For simplicity, we assume that both parties draw their expectations (A, and Ad) of the value of proceeding to pretrial (postdiscovery) negotiations from a normal distribution with means, E(A) (for the plaintiff) and Ed(A) (for the defendant), and common standard deviations of a. In assessing the likely economic consequences of mandatory disclosure, we can ignore the postdiscovery settlement period because trial costs are likely to be equivalent with or without mandatory disclosure and because the outcome of the case is likely to be the same under either system of rules if the case proceeds to trial. Figure 3 depicts the incentives to settle in the absence of the mandatory disclosure rule: Figure 3 No Mandatory Disclosure

Settle?

NoDiscovery

L4-I , -A

-

LDjI

[,-s] With no mandatory disclosure, the aggregate cost of not settling the dispute prior to discovery is equal to DP + D. (and DP and Dd are the costs of discovery for the plaintiff and defendant respectively), and settlement will occur as long as the aggregate cost of not settling exceeds the difference in the parties' expected award amounts, or, in mathematical terms,33 when E,(A)

-

Ed(A) < D, +

d

32. See Loewenstein, Issacharoff, Camerer & Babcock, supra note 20, at 153 (concluding that experimental data indicates that parties to a legal dispute hold significant self-serving biases). 33. Setting z = AP - A,-D - Dd, the probability that z is less than zero-i.e., P(z < 0)-can be determined. The random variable z has a normal distribution with mean, E(z) = EP(A) - E (A) - Dp 4 Dd, and variance, 20

[z -

N(E(A)

-

Ed(A) - D, - DI, 202)].

This assumes that A, and A, are

independent. If the parties estimates of the consequences of going to trial are unbiased, then E(A) E(A) = 0 and the probability of settling is P(z < 0) given that z - N(-D - Dd, 20). However, if there is a self-serving bias (E(A) 5 A), such that the plaintiff expects a higher judgment than does the defendant, then E,(A) - Ed(A) = B > 0, in which case the probability of settlement is P(z < 0) given thatz - N(B - DP - Dd, 202).

764

Texas Law Review

[Vol. 73:753

With mandatory disclosure, as depicted in Figure 4, we assume that all costs of disclosure, D,, are borne by the defendant.' Figure 4 Mandatory Disclosure

Sete No

Mandatory

[SI,41~ ]

Settle?

[A-DP, -A-D I-D 2 ]

NODiscovery

I,"-

d]

With mandatory disclosure, settlement will occur prior to disclosure as long as the difference between the expected values to the plaintiff and defendant of the initial disclosure period is less than the cost of disclosure; in other words, when E (A)- E (A) < DdI However, the prospects of settlement are affected if the parties have biased expectations of the expected outcome of the case, as denoted by the term (B). If E (A) - Ed(A) = EP(A) - Ed(A) = B, that is, if the parties' expectations prior to mandatory disclosure are identical to what they would have been prior to discovery in the absence of disclosure, then the probability of settlement is P(B - Dd < 0).11 Note that D1 , the cost of mandatory disclosure for the defendant will almost surely be less than that of conventional discovery were the case to be fully litigated, and that the process of complying with mandatory disclosure will reduce the defendant's subsequent discovery costs. In calculating the probability of settling after mandatory disclosure but before discovery, therefore, we make the simplifying assumption that the defendant's subsequent discovery costs are reduced by exactly the amount spent complying with mandatory disclosure and that the plaintiff's discovery costs are unaffected by mandatory disclosure. Thus, parties who have gone through mandatory disclosure will settle prior to discovery if and only if the difference between the expected values to the plaintiff and defendant of proceeding past the disclosure stage

34. This assumption is based on the general rule of thumb that, in litigation, information concerning liability is generally in the hands of the defendant, while information concerning damages is likely to be held by the plaintiff. This is in part a reflection of the many forms of civil cases in the federal courts in which liability turns on the state of mind of the defendant-claims such as securities fraud, civil rights violations, antitrust conspiracy, et cetera. 35. This assumption is reasonable, because in both cases no significant expenses have been incurred at this point.

1995]

Mandatory Disclosure

is less than the cost of proceeding past discovery; in mathematical terms, when

E(A -E(A) < Dd -DI+ DP If the exchange of information associated with mandatory disclosure does not cause the parties to change their expectations concerning the outcome of trial, then settlement will occur prior to discovery only if the difference in the two parties' expectations is greater than Dj but less than Dd - D, + DP. The probability of this happening is simply P(B - Dd + Dd - DP < 0) - P(B - DP < 0). At the opposite extreme, if the exchange of information causes the two parties to re-create their estimates from scratch, then the probability of settling prior to discovery but after mandatory disclosure is simply P(B2 - Dd + D, > 0), in which the superscript on the bias term, B, indicates the possibility that the exchange of information could lead to a systematic convergence or divergence between the parties' expectations. Between these two extremes of complete independence or complete dependence, the probability of settling prior to discovery but after mandatory disclosure will vary monotonically with the serial correlation of each party's pre- and postdisclosure estimates.36 To derive empirically the likely consequences of mandatory disclosure, we assume some specific values for disclosure and discovery costs. We assume, arbitrarily, that discovery costs are equal to 1 unit for each party, that mandatory disclosure costs the defendant 0.5 units, and that, as discussed above, this disclosure cost is subsequently deducted from the defendant's discovery costs. None of these exact numbers are particularly important; all of the results derived below are robust to changes in one or all of these costs. Figures 5a-5c depict the total costs incurred under the two rules-that is, with and without mandatory discovery-as a function of the magnitude of the two parties' degree of bias. In each figure, the dotted line represents the expected aggregate costs of discovery (the likelihood of the occurrence of discovery multiplied by the cost of discovery) in the absence of a mandatory disclosure rule, the solid line represents the combined costs of disclosure and discovery assuming that the two parties do not revise

36. Serial correlation (between pre- and postdisclosure estimates of award amounts) is equal to one when parties do not revise their estimates (or all parties do so by a constant that reflects the degree of convergence or divergence). Serial correlation is zero when parties generate entirely new estimates of the award amount following disclosure (again, possibly shifted by a constant that reflects convergence or divergence). If the probability of settling prior to discovety depends on the assumed degree of serial correlation so that, for example, it is higher with a serial correlation of one than of zero, then monotonicity implies that this probability is, more generally, an increasing function of the degree of serial correlation.

Texas Law Review

[Vol. 73:753

their estimates of the consequences of going to trial following mandatory disclosure, and the dashed line represents the combined cost of disclosure and discovery under the assumption that both parties make a completely new estimate of the consequences of going to trial following mandatory disclosure. The three figures differ in the assumption they make concerning the dispersion of the possible distribution of trial outcomes. By dispersion, we mean the degree of consensus (actually, the lack thereof) that we would expect to observe in estimates of the likely outcome of trial by parties in the same role. Figure 5a assumes a very narrow dispersion, which signifies that most defendants would arrive at similar estimates of the likely outcome of a trial as would most plaintiffs (although the two groups might reach disparate estimates because of bias). Figure 5b assumes an intermediate level of dispersion and 5c assumes a wide dispersion. To give a qualitative sense of the magnitude of these dispersions, Figure 5a assumes that the standard deviation of estimates of trial outcomes is equal to the defendant's costs associated with mandatory disclosure, while Figure 5c assumes that the standard deviation is 2.5 times as large as the sum of the two parties' discovery costs in the absence of mandatory disclosure. Figure 5a: Narrow Dispersion ......

No Mandatoy

1.6 -

-

Mandatory&Dependent Mandatory& Independent

1.4

7

1.2

0.8

7

0.6 0.4

"

0.2

100

0

0.5

I

Bias

1995]

Mandatory Disclosure Figure 5b: Intermediate Dispersion

0

0.5

1

1.5

Bias

1.2

Figure 5c: Wide Dispersion

1

0.8 00

0.6

0.4

........

No Mandatory & Dependent Mandatory & Independent

-Mandatoty

--

0.2

i

Bias

Texas Law Review

[Vol. 73:753

Figure 5a shows that, with low dispersion of trial costs, anticipated

litigation costs with mandatory disclosure will be higher than those in the absence of mandatory disclosure, regardless of the correlation between predisclosure and postdisclosure estimates., The same is true of moderate dispersion (Figure 5b), although, in this case, the costs associated with mandatory disclosure are relatively similar to those arising from the no mandatory disclosure case, assuming that disputants generate entirely new estimates of the likely outcome of a trial following mandatory disclosure. Finally, in the high dispersion case, the absence of mandatory disclosure is intermediate in terms of costs between the two polar cases of mandatory disclosure." The relative costs of mandatory disclosure as compared with no mandatory disclosure, therefore, depend importantly on the dispersion or variance among the estimates of possible trial outcomes and on the dependence or independence between the parties' predisclosure and postdisclo-

sure estimates of trial outcomes. If the assumption of dependence is closer to reality than the assumption of independence (as we think is likely), then

mandatory disclosure will almost inevitably increase aggregate costs. In other words, mandatory disclosure will reduce costs only if it is a necessary precondition to the parties being able to revise their assessments of the case, and if we may assume that parties to litigation are indeed

capable of such a fresh revision of their predisclosure predilections-an unlikely occurrence. This is a rather disturbing result, given that the main purpose of mandatory disclosure is to reduce such costs. B.

The Asymmetrical Distributionof Disclosure Costs

As noted in the previous subpart, the costs of mandatory disclosurethe costs of failing to settle prior to mandatory disclosure-will be borne in large part by the defendant.38 In the mathematical analysis presented

37. The reason for the effect of dispersion is as follows: As the dispersion of the parties' estimates of the outcome of trial increases relative to the magnitude of discovery costs and relative to the magnitude of any bias, settlement will come to be determined almost exclusively by whether the plaintiff's estimate of the outcome of the trial is larger than that of the defendants. As dispersion increases, the probability of settlement in any negotiation will converge on 50%. Without mandatory disclosure, therefore, there is simply a 0.5 chance of incurring discovery costs of 2, which has an expected value of 1. With mandatory disclosure, under the assumption that predisclosure and postdisclosure estimates are statistically independent, there is a 0.5 chance of incurring disclosure costs of 0.5, and a 0.52 chance of incurring discovery costs of 1.5, for a total expected cost of 0.625. 38. There is more than a little irony in comparing this fact to the original arguments in favor of discovery reform. For example, Vice President Dan Quayle in his well-known address to the American Bar Association proclaimed, "Unnecessary document requests and depositions can disrupt or put on hold a company's entire research and development program, and the very idea of limits on discovery is outdated.... Worse yet, discovery can be a virtually cost-free weapon for the requesting party. That is what we want to change." Dan Quayle, Address at the Annual Meeting of the American Bar Association (Aug. 13, 1991) (transcript available in LEXIS, News Library, Script File).

1995]

Mandatory Disclosure

above, we have represented this asymmetry by the stylized assumption that all costs of mandatory disclosure are borne by the defendant. Because the results illustrated in Figures 5a-5c are robust even when this assumption is relaxed, this assumption is not particularly important in economic terms and is thus not critical to the results just described. For two reasons, however, the significance of the asymmetry is probably understated by such

an economic analysis. First, the foregoing analysis does not take account of the potential for strike suits, and the asymmetry of costs has important

implications for the initiation of such suits. Second, the economic analysis just presented does not incorporate a potentially important psychological factor that is likely to come into play as a result of the asymmetry in disclosure costs. 1. Strike Suits.-In general, frontloading litigation costs should serve a gatekeeping function to litigation. To the extent that litigants expect initial expenditures to be high, they should be deterred from entering the litigation process. Most simply, an elevated filing fee will deter litigants

with claims falling below the level of the filing fee from commencing suit. But mandatory disclosure produces an elevation of early cost only for the

defendant, leaving the plaintiff's decision to file suit unaffected. There is, consequently, no gatekeeping function served by the early costs associated with mandatory disclosure. By imposing costs almost exclusively on the defendant and by providing sanctions for a lackluster provision of information, 9 the new statute

greatly increases the attractiveness of strike suits. Strike suits are suits in which the expected value to the plaintiff is zero or negative as a result of

a low probability of success together with the anticipated costs of litigation, but in which the settlement value, defined as the midpoint of the settlement zone, is greater than zero.' In such circumstances, the sole incentive for

bringing suit is the expectation of a negotiated wealth transfer. Strike suits

39. The sanctioning provisions are found in FED. R. CIV. P. 37(a)(2)(A), 37(c). 40. A conventional economic analysis of a strike suit can be expressed as the conjunction of two conditions: that the expected value to the plaintiff of trial is zero or negative (EP(A) - D : 0), but that the parties are nevertheless likely to settle on a positive transfer from defendant to plaintiffif they settle prior to trial. This condition will be met, on average, when [E4 (A) + D] - [EP(A) - DA]12 > 0. Both conditions are met when E(A) + Dd ; EP(A) - Dp 0 Thus, strike suits are likely to be successful 0. when the defendant's disclosure costs are high and when the defendant is pessimistic concerning the likely outcome of an adjudicated settlement. See Lucian A. Bebchuck, Suing Solely to Extract a Settlement Offer, 17 3. LEGAL STUD. 437, 440-41 (1988) (postulating that a defendantwill likely settle a frivolous suit if he erroneously believes that he may lose at trial and incur substantial litigation costs); Avery Katz, The Effect of Frivolous Lawsuits on the Settlement of Litigation, 10 INT'L REV. L. & ECON. 3, 4 (1990) (attributing the settlement of strike suits to defendants' uncertainty about the merits of plaintiffs' claims and about anticipated trial costs); David Rosenberg & Steven Shavell, A Model in Which Suits Are Brought for Their Nuisance Value, 5 INT'L REV. L. & ECON. 3, 4-5 (1985) (developing the asymmetric cost theory of nuisance suits).

Texas Law Review

[Vol. 73:753

are encouraged by mandatory disclosure because it is possible for the potential initiator of a strike suit to credibly threaten to force a potential defendant to incur disclosure costs, even in the relative absence of a strong legal case. The threat is credible because if the defendant fails to comply with mandatory disclosure requirements or discloses information in a cursory fashion, the resultant preclusion of evidence might strengthen the plaintiff's hand, turning a nonexistent or weak case into a stronger one. To avoid incurring such disclosure costs, the potential defendant may be willing to pay the plaintiff to avoid proceeding with suit. Note that this option is much less viable in the absence of mandatory disclosure because initial discovery costs are more evenly distributed between the two parties; discovery has a downside potential for both parties. 2. PsychologicalFactors.-Thesecond negative potential consequence of the asymmetry of disclosure costs is that serious plaintiffs-those not engaged in strike suits-may view such disclosure as costless to themselves. In fact, after the defendant has incurred the costs of mandatory disclosure, the settlement zone will shift in the defendant's direction. In other words, the plaintiff should expect to negotiate a smaller settlementthe difference being approximately half of the defendant's disclosure costs-after such disclosure has taken place. In the economic analysis just described, this anticipated shift creates an equivalent incentive for both the plaintiff and defendant to negotiate in good faith prior to mandatory disclosure. However, economic models, such as the one presented in the previous section, typically exaggerate the rationality of players. Research in psychology points to significant limitations in people's abilities to process and store information."1 Of special relevance to legal disputes is the finding that people do not "solve" economic games in the manner prescribed by economics. Economic analyses of multistage bargaining (of which legal disputes are an example) assume that players solve the game by a process of "backward induction." 2 Applied to the decision tree depicted in Figures 2 and 3, backward induction implies that people base their bargaining in the third settlement bargaining phase on an analysis of what they expect would happen if they went to trial, their bargaining in the

41. See, e.g., Herbert A. Simon, A BehavioralModelofRational Choice, 69 Q.J. ECON. 99, 10310 (1955) (arguing that individuals do not act with maximum rationality because of their inability to predict the exact nature of outcomes and their corresponding inability to determine definite payoffs for each possible outcome). 42. See, e.g., Ariel Rubinstein, Perfect Equilibriumin a BargainingModel, 50 ECONOMETRICA 97, 98 (1982) (examining the bargaining problem through the axiomatic method in which one states as axioms several properties that would seem natural for the solution and then discovers that the axioms actually determine the solution uniquely).

1995]

Mandatory Disclosure

second phase on what they expect to be the outcome of the third phase and if they went to trial, and so on. Contrary to this view, empirical research on behavior in games has found that people actually proceed in a more myopic fashion, "looking ahead" at most by a single stage; they do not make use of information about later phases of bargaining even when such information is costless to acquire.' The failure to look ahead implies that plaintiffs may fail to anticipate or take account of the effect of mandatory disclosure on the settlement zone. Mandatory disclosure will appear to the serious plaintiff and to the strike suit plaintiff alike as a free good, a chance to obtain costless information about potential weaknesses in the defendant's case. If plaintiffs ignore the negative consequence of disclosure to themselves, they will be less likely to negotiate in good faith at the outset and much more likely to delay serious negotiations until after disclosure. In other words, limited rationality may create the illusion that it is costless to require the other side to disclose information. If this is the case, then the economic analysis presented above overstates the probability that the parties will settle prior to mandatory disclosure and, consequently, understates the costs associated

with mandatory disclosure.

43. Colin F. Camerer et al., Cognition and Framing in Sequential Bargainingfor Gains and Losses, in FRONTIERS OF GAME THEORY 27-47 (Ken Binmore et al. eds., 1993). In response to these findings and to a general growing recognition that standard models assume too high a level of rationality, a variety of "learning models" that assume very limited rationality on the part of agents have been proposed and tested via simulation. See James Andreoni & John H. Miller, Auctions with Artificial Adaptive Agents, 8 GAMES & ECON. BEHAV. (forthcoming 1995) (examining systematic bidding errors in auction bidding strategies in the context of adaptive learning); James Andreoni & John H. Miller, Rational Cooperationin the FinitelyRepeated Prisoner'sDilemma:ExperimentalEvidence, 103 ECON. J. 570 (1993); Alvin E. Roth & Ido Erev, Learningin Extensive-Form Games: Experimental Data andSimple Dynamic Models in the IntermediateTerm, 8 GAMES & ECON. BEHAV. (forthcoming 1995) (using simple learning models to track the behavior observed in experiments concerning three extensive form games with similar perfect equilibria); Ken Binmore et al., Learning to Be Imperfect: The Ultimate Game, 8 GAMES & ECON. BEHAv. (forthcoming 1995). There is also increased recognition that agents' utility functions have been misspecified in past theorizing. For example, there is by now a vast amount of evidence that people care as much (or more) about relative payoffs than about absolutepayoffs. See GeorgeF. Loewensteinet al., Social Utility andDecisionmakingin Interpersonal Contexts, 57 J. PERSONALrY & SOc. PSYCHOL. 426, 426 (1989) (suggesting that disputants are generally concerned not only with outcomes they receive, but also with the outcome of their opponents). This can explain, for example, the high prevalence of rejections of positive but unequal offers in the ultimatum game and, consequently, the high rate of equal or near equal offers. See Werner Glith & Reinhard Tietz, Ultimatum Bargaining Behavior: A Survey and Comparison of ExperimentalResults, 11 J. ECON. PSYCHOL. 417, 424 (1990) ("[Players... are often enough willing to sacrifice quite a monetary amount in order to punish [the other] player... for making an 'unfair' proposal."); Richard H. Thaler, Anomalies: The Winner's Curse, J. ECON. PERsp., Winter 1988, at 191, 192-201 (positing that when faced with competition, businesses often make higher bids for a purchase than is economically rational).

Texas Law Review

[Vol. 73:753

C.

Escalation Another basic assumption of economic models is that people treat costs incurred in the past as "sunk"-that is, as irrelevant to subsequent decisionmaking." Decisionmakers are assumed to decide between options solely on the basis of prospective costs and benefits. However, psychological research on "escalating commitment" suggests that people who have incurred costs in the past are more likely to continue to invest in an enterprise even when prospective expected returns fall below those offered by other options. Applied to bargaining in multistage games, this tendency to "throw good money after bad" suggests that people who have incurred costs in early stages of a game are likely to be more intransigent in later stages than would be predicted by a narrow economic analysis. The most common explanation for escalating commitment attributes it to decreasing sensitivity to losses as losses mount.' For example, the psychological difference between losing nothing and losing $100 is much greater than the difference between losing $1000 and losing $1100.' Because the main incentive for settling at any point in a dispute is the threat of future losses, if one or more of the parties becomes desensitized to losses (as a result of having incurred prior losses), then they will have less incentive to settle the dispute. Mandatory disclosure, by leading to an early outlay by defendants, may decrease their subsequent willingness to settle. Again, the economic model, without enlightenment by psychology, understates the potential negative consequences of mandatory disclosure for settlement. D.

The Effect of Disclosure on Parties' Expectations

The final potential consequence of mandatory disclosure on settlement that we examine stems from the impact of mandatory disclosure on the parties' expectations of what will happen if they go to trial. Economic models of settlement typically assume that people process new information in an unbiased fashion. In the long run, this implies that the receipt of common information by parties involved in a lawsuit will lead to a convergence of expectations.47 Thus, the costs associated with discovery can be

44. Barry M. Staw, The Escalation of Commitment to a Course ofAction, 6 AcAD. MGMT. REV. 577, 578 (1981); Richard Thaler, Toward a Positive Theory of Consumer Choice, 1 J. ECON. BEHAV. & ORO. 39, 47 (1980). 45. See Daniel Kahneman & Amos Tversky, ProspectTheory: An Analysis of Decision Under Risk, 47 ECONOMETRIcA 263, 278 (1979) ("[Tlhe marginal value of... losses generally decreases with their magnitude."). 46. Id. 47. The economic equilibrium models used in Bayesian theory imply that giving two parties common information should, on average, cause their expectations to converge. Bruno de Finetti,

Mandatory Disclosure

1995]

justified, in part, by the presumed propensity for joint information to promote settlement as parties rationally assess the likely outcome of litigation based on common facts." If such expectational convergence and its effect on settlement is sufficient, its benefit will outweigh the additional costs associated with the production of information. Taken one step further, the expected convergence of views following less expensive mechanisms of exchanging information should in turn lead to an increased efficiency of settlement under mandatory disclosure. Although mandatory disclosure is more likely to be efficient if it does, in fact, lead to a convergence of the two parties' expectations concerning the outcome of going to trial, its efficiency also depends on the degree of bias prevailing in a particular case. Both of these effects are depicted in Figures 6a and 6b. Figure 6a Dependent Estimates: Difference Between Mandatory Disclosure and No Mandatory Disclosure

-0.5

Convergence/ Divergence

Foresight:Its LogicalLaws, Its Subjective Sources, in STUDIES IN SUBJECrIVE PROBABILITY 93, 152 (Henry E. Kyburg, Jr. & Howard Smokier eds. & Henry E. Kyburg trans., 1964). 48. See RICHARD A. POSNER, ECONOMIC ANALYSIS OF LAW 557 (4th ed. 1992) ("[A] full exchange of the information in the possession of the parties is likely to facilitate settlement by enabling each party to form a more accurate, and generally therefore a more convergent, estimate of the likely outcome of the case .... ).

774

Texas Law Review

[Vol. 73:753

Cost Difference

toq

to

tr! to•

IYo

o

900

go

0

19951

Mandatory Disclosure

These Figures show the anticipated aggregate cost difference between a regime with and without mandatory disclosure as a function of the amount of convergence or divergence that results from information exchange and of the aggregate degree of bias prevailing in the case. The two horizontal axes of the graph, the floor of the charts, represent varying magnitudes of bias (from 0 to 2) and the degree of convergence or divergence (from -1, indicating that the difference in the parties' estimates decreases by one unit following disclosure, to 1, indicating that it increases by one unit). The vertical axis of each graph shows the difference in costs associated with the use and nonuse of mandatory disclosure. The vertical axis is key for our analysis because a rise above the floor represents a projected increase of costs for mandatory disclosure relative to no mandatory disclosure. A drop below the floor level of 0 represents an efficiency gain as a result of mandatory disclosure. In other words, positive y-values in the figure designate situations in which mandatory disclosure is more costly than no mandatory disclosure; negative y-values designate the opposite situation. Figure 6a depicts the case of complete serial correlation-the parties' estimates of the consequences of going to trial are unchanged by the exchange of information that results from mandatory disclosure, except for a fixed amount of divergence or convergence resulting from the information produced through mandatory disclosure. Figure 6b depicts the case of complete independence-a situation in which both parties make entirely new estimates of the consequences of going to trial after disclosure occurs, but the average degree of bias is either decreased or increased by the amount specified on the axis labeled "convergence/divergence." In both figures, the relative advantageousness of mandatory disclosure increases with the postdisclosure convergence of views, as we would expect. Two other features of the figures, however, are more surprising. First, in both figures, the directional effect of bias depends on the degree of convergence or divergence caused by mandatory disclosure. There is a striking interaction between bias and convergence or divergence in their determination of the advantageousness of mandatory disclosure. When disclosure leads to divergence, then mandatory disclosure is more inefficient under conditions of high bias than low bias. This is because, with high bias, cases are likely to settle only when high nonsettlement costs overcome the impediment associated with the bias. Mandatory disclosure establishes an initial settlement zone with very small costs during which settlement is unlikely and then exacerbates the bias while lowering aggregate costs during the prediscovery period. When mandatory disclosure leads to convergence, however, its relative efficiency is greater in regimes with high bias rather than low bias. This is because, with low bias, cases are very likely to settle even without the convergence potentially

Texas Law Review

[Vol. 73:753

produced by disclosure, so disclosure adds little to the prospects of settlement while introducing an additional costly stage during which settlement is less likely. With high bias and convergence, cases are unlikely to settle prior to disclosure, but the benefits of convergence are more likely to offset the costs of disclosure itself. The second surprising feature of the figures is the small range of situations in which mandatory disclosure does, in fact, increase efficiency. Given the background parameters we have specified and the range of parameter values depicted in Figure 6a, it can be seen that, in the case of dependent estimates, mandatory disclosure is always inefficient-that is, it raises aggregate litigation costs. With independent estimates, the figure shows that mandatory disclosure is likely to produce efficiency gains when it leads to a convergence of expectations but not otherwise. Again, the effect of bias on aggregate costs depends on whether disclosure leads to a convergence or divergence of expectations. Whether disclosure leads to divergence or convergence of the parties' expectations, therefore, is potentially critical for evaluating the impact of mandatory disclosure. With divergence, mandatory disclosure will almost certainly result in higher aggregate costs; however, with convergence, the relative efficiency of mandatory disclosure depends on the degree of dependence or independence between the parties' predisclosure and postdisclosure judgments of the likely outcome of trial. In reality, both extreme examples (complete dependence or independence) are unlikely to capture the complexities of the integration of new information in the litigation context. Reality probably lies somewhere in between the extremes of no revisions of predisclosure positions depicted in Figure 6a and the completely new evaluation of the positions depicted in Figure 6b. What is clear, however, is that even at the polar extremes, the only condition under which mandatory disclosure can be expected to reduce litigation costs significantly is one in which there is a great deal of systematic bias and in which the disclosed information leads to a considerable convergence in the positions of the parties. Thus, the uniquely advantageous situation for mandatory disclosure is when there is substantial bias and the disclosure leads to meaningful convergence. Which then is a more realistic assumption: convergence or divergence? Contrary to the conventional assumption in the law and economics literature,4 9 the few studies that have examined this question empirically have obtained evidence of a divergence rather than a convergence of views following the exchange of information.' In one classic study, people

49. See id. (assuming that a full exchange of information leads to convergent views of the likely outcome of a case). 50. See, e.g., Colin F. Camerer& George Loewenstein, Information, Fairness, and Efficiency in Bargaining, in JUSTICE (Barbara A. Mellers ed., 1993); Albert H. Hastorf& Hadley Cantril, They Saw

1995]

Mandatory Disclosure

who were either in favor of or opposed to the death penalty reviewed two empirical studies, one which concluded that the death penalty deterred crime and the other which reached the opposite conclusion.5 ' Subjects found fault with the methodology of the study that reached a conclusion different from their own views, but endorsed the methodology of the study whose conclusion agreed with their views.'2 As a result, even though opponents and proponents of the death penalty read exactly the same two studies, their views diverged rather than converged. 3 In a somewhat different context, we ran an experiment in which two subjects were assigned the roles of plaintiff and defendant in a legal dispute arising from a collision of an automobile with a motorcycle.' The two parties read the same case materials and were told that an independent judge had also read the materials and arrived at a judgment. The subjects then recorded what they believed was a fair settlement to the case, predicted the judge's decision, and attempted to negotiate a settlement consisting of a payment to the plaintiff from the defendant (who received extra money at the outset). If the parties failed to settle, the decision of the judge determined the value of the payment and legal costs were imposed on both parties. In one condition, subjects were not told their role (plaintiff or defendant) until after they had read the case; in the other, they were told their role prior to reading the case. Even though both parties read exactly the same case materials, the expectations of plaintiffs and defendants who knew their roles before reading the case diverged markedly in favor of the value of their assigned positions. These divergent expectations led to high rates of nonsettlement, even though there were strong financial disincentives to allowing the case to be resolved by an arbiter. 55 While experimental studies do not take account of the learning from error that would presumably accrue from the engagement of repeat actors, such as lawyers, there is reason to believe that even the presence of more detached litigators would not significantly mediate this effect. In a recent study performed under the auspices of the American Bankruptcy Institute, for example, experienced bankruptcy lawyers and judges were asked to

a Game: A Case Study, 49 1. ABNORMAL & SOC. PSYCHOL. 129, 132 (1954) (reporting that after watching a football game, fans interpreted controversial plays in different ways depending on their team allegiance); Charles G. Lord et al., BiasedAssimilation andAttitude Polarization:The Effects of Prior Theories on Subsequently ConsideredEvidence, 37 J. PERSONALUrY & SOC. PSYCHOL. 2098, 2105 (1979) (concluding that when identical information is given to subjects with opposite viewpoints, beliefs polarize further rather than converge). 51. Lord et al., supra note 50, at 2100-01. 52. Id. at 2102-04. 53. Id. at 2104. 54. Loewenstein, Issacharoff, Camerer & Babcock, supra note 20. 55. Id. at 145-53.

Texas Law Review

[Vol. 73:753

assess the efficiency and propriety of local bankruptcy practices with regard to the awarding of attorneys' fees.5 6 The survey revealed systematic selfserving biases, with judges and lawyers overestimating the performance of their group relative to the assessment by the other group. Together, these studies, and many others that document an "egocentric bias" in judgment, suggest that, contrary to the standard assumption of convergence, information sharing between disputants is fully capable of producing a divergence of expectations. As explained above, such divergence, in turn, suggests that mandatory discovery is likely to increase aggregate litigation costs. E.

Summary

Whether mandatory disclosure increases or decreases costs, therefore, appears to hinge on three main factors: (1) the degree of dispersion of the parties' estimates of the outcome of a trial; (2) their tendency to retain or revise such estimates as a result of the information exchange associated with mandatory disclosure; and (3) whether disclosure, on average, leads to a convergence or divergence of expectations. Our analysis shows that there are circumstances in which mandatory disclosure is likely to result in lower aggregate expenditures-namely, when there is considerable uncertainty concerning the outcome of a trial, when disclosure leads parties to produce a completely fresh estimate of the likely outcome of going to trial, and when disclosure leads to a convergence of the parties' expectations. However, none of these three conditions alone is sufficient for mandatory disclosure to outperform nonmandatory disclosure; all must be satisfied simultaneously. Although the first condition is, in fact, often likely to be satisfied-the outcome of going to trial is often highly uncertain-the second two are much less likely to hold. First, it seems highly unlikely that the information exchange produced by mandatory disclosure is likely to lead to a wholesale revision of the parties' estimates of the consequences of going to trial, particularly for defendants who, prior to disclosure, are already in possession of most of the evidence that is disclosed. Second, the experimental evidence reviewed above suggests that there is no certainty that mandatory disclosure will produce convergence as opposed to divergence of expectations. Thus, a relatively strong conclusion emerges from our analysis: that mandatory disclosure will most likely increase rather than decrease aggregate litigation costs.

56. Theodore Eisenberg, Differing Perceptions ofAttorney Fees in Bankruptcy Cases, 72 WASH. U. L.Q. 979 (1994).

1995]

Mandatory Disclosure

III. Litigation Strategies In addition to the strategic misincentives created by mandatory disclosure, there are a number of what may be termed structural problems

raised by the amended discovery procedures. First, as many commentators have pointed out, there is an inherent vagueness to the command for the production57of names of witnesses and identification of documents that are "relevant" to the case.58 In light of the liberal standards of notice

pleading,59 reaffirmed only last year by the Supreme Court,' it is by no means self-evident what the full scope of disclosure should be. Opponents of mandatory disclosure have pointed to the vague claims in products liability cases, for example, in which the complaint would state only that a defendant had caused a defective product to be placed in the stream of commerce.

61

Under such circumstances, the search for relevant information or

documents could be open-ended. 62 To the extent that the attorney for a producing defendant narrows the request to that which is truly relevant to the issue of liability, for example, the information to be disclosed is not 57. This is the language of the new, amended disclosure rule. FED. R. CIV. P. 26(a). 58. This was one of the primary arguments raised by opponents of mandatory disclosure prior to the amendment of Rule 26. See, e.g., Jeffrey J. Mayer, Prescribing Cooperation:The Mandatory PretrialDisclosureRequirement ofProposedRules 26 and37 of the FederalRules of Civil Procedure, 12 REv. Lri. 77, 112-16 (1992) (noting that the parties to a dispute will often have differing conceptions of what information is relevant). 59. See Conley v. Gibson, 355 U.S. 41, 48 (1957) (remarking that the Federal Rules of Civil Procedure intend pleading to facilitate determination on the merits, not to determine the outcome of a dispute based on the formality of the pleadings). 60. Leatherman v. Tarrant County Narcotics Intelligence & Coordination Unit, 113 S. Ct. 1160, 1163 (1993). 61. See, e.g., Bell et al., supra note 1, at 42 (arguing that in products liability cases defendants have no basis for determining what or how much information they should disclose). 62. This problem led to a wonderful parody of an attorney's letter to a client that would presumably be required under the new discovery rules: We will need to schedule a meeting with you at the earliest possible opportunity to review all facts tending to support the plaintiff's claims for relief. I will immediately providethis information to the plaintiff's attorney. If the plaintiffhas failed to pursue any appropriate claims against your company, we should gather together and pass on to plaintiff's counsel all facts which he would want to know in order to amend his complaint and pursue such additional theories against your company. If you or your employees are aware of any facts which might support a claim for punitive damages, please organize that information in a format which will be easily understandable to plaintiff's counsel. Although no theories supporting punitive damages have been raised, the plaintiff's attorney would surely be interested in learning any such facts and would undoubtedly be anxious to amend the complaint to include a prayer for such relief. It is our obligation to exercise our best efforts to collect any appropriate data which the plaintiff's attorney would want and I am sure punitive damages would be high on his list. Bruno de Finetti, The Disclosure Rule Is a Mistake, MARicOPA LAw., Aug. 1992, at 6, quoted in RICHARD L. MARCUS Er AL., C IL PRocmnuRE 69-70 (Supp. 1994).

Texas Law Review

[Vol. 73:753

simply an abstract category of documents, but rather is a categorization of documents that has been filtered by the producing attorney. In effect, the disclosure rule allows a plaintiff to secure not simply information responsive to a request of the plaintiff's formulation, but information with "value added by the litigation lawyer" of the opposing side.' This First, by treating the situation creates two significant concerns. information to be disclosed as if it were a prepackaged good, the mandatory disclosure rule ignores the costs associated with production-and all the attendant problems we have already discussed. The inescapable bottom line is that "information is simply not free. " ' Second, at the point of forcing the disclosing attorney to create a value-added package, the disclosure rule inevitably threatens to commandeer the work product exception carefully protected by the Supreme Court in Hickman v. Taylor' and subsequently incorporated into Rule 26(b)(3).' The confrontation with the work product privilege is somewhat alleviated by the final language of Rule 26(a)(1)(C), which requires mandatory disclosure only as to those facts set forth with particularity.67 Even when such specificity is met, it is unlikely that parties seeking the information will be satisfied to accept only the proffered disclosure. Because the determination of what should be disclosed is made by the party providing information, mandatory disclosure provides a source of newly contested litigation for three reasons. First, the looseness of both the initial pleading and disclosure standards will undoubtedly require ample discovery to test the sufficiency of the information produced at the automatic disclosure stage. The new rule then provides an incentive for an aggressive second round of discovery to determine whether information that should have been disclosed was instead withheld. 68 Second, because the amended Rule carries with it a generous preclusion of evidence sanction, there is a strong incentive for parties to engage in escalated discovery to

63. Kobayashi et al., supra note 15, at 8. 64. Id. at 11. 65. 329 U.S. 495 (1947). 66. See James Holmes, Note, The Disruption of MandatoryDisclosure with the Work Product Doctrine:An Analysis ofa PotentialProblem anda ProposedSolution, 73 TEX. L. REV. 177 (1994) (exploring the potential conflict between the demands of mandatory disclosure and the protections offered by the federal work product doctrine). For a thorough analysis of the history, operation, and justification of the federal work product doctrine, see CHARLES A. WRIGHT, LAW OF FEDERAL COURTS 589-98 (5th ed. 1994). See also Alex W. Albright, The Texas Discovery Privileges:A Fool's Game?,

70 TEX. L. REV. 781, 784-98 (1992). 67. But see Mayer, supra note 58, at 114 (arguing that restricting the scope of disclosure to facts pleaded with particularity saps the disclosure rule of any vitality). 68. Although there is not, as yet, any case law to assess whether lawyers will add the new discovery sanctions to their litigation arsenal, there are some early blips on the horizon. See, e.g., Scheetz ex rel. Handeland v. Bridgestone/lFirestone, Inc., 152 F.R.D. 628 (D. Mont. 1993) (challenging the sufficiency of mandatory disclosure under a local variant of the new Federal Rule 26).

1995]

Mandatory Disclosure

expose omissions in the production of information in order to gain a strategic advantage at trial. Third, the amended rule could easily lead to a new round of motion practices identified by Justice Scalia,' in which the quality of the automatically disclosed information is tested under the potential preclusion sanction. In response to this problem, there is an impulse to diminish the opentextured quality of mandatory disclosure by specifying the exact information that needs to be produced at the beginning of each case.' To the extent that this is accomplished, however, mandatory disclosure begins to resemble a prearranged series of interrogatories that must be answered at the beginning of each lawsuit." At this point, some justification must be given for taking the request for information out of the hands of litigants who can cheaply tailor them to the particular needs of individual cases. If these interrogatories are truly of such obvious and universal need in litigation, they are most likely already part of the standard discovery repertoire of any modestly experienced attorney. It is therefore unclear, in the first instance, that requiring production in the form of mandatory disclosure-rather than through service of a standard set of interrogatories and requests for production of documents-entails any significant cost savings at all .' Moreover, there is absolutely no evidence that mandatory

69. See Amendments to the Federal Rules of Civil Procedure, 146 F.R.D. 402, 510 (1993). 70. This more limited form of disclosure is reflected in the final version of amended Rule 26, which specifies certain confined categories of information subject to mandatory disclosure. Specifically, the disclosure obligation extends to the names of persons and documents "relevant to disputed facts alleged with particularity in the pleadings," evidence of damage calculations, and evidence of insurance policies. FED. R. CIv. P. 26(a)(1)(A)-(D). This language corresponds to the Advisory Committee's desire to focus mandatory disclosure on "certain basic information" in the possession of each party. See FED. R. Civ. P. 26(a) advisory committee's note. 71. In fact, the Advisory Committee refers to the mandatory disclosure provisions "[as the functional equivalent of court-ordered interrogatories... of four types of information that have been customarily secured early in litigation through formal discovery." FED. R. Civ. P. 26(a)(1) advisory committee's note. 72. We leave aside a number of other criticisms of mandatory disclosure, some of which may exacerbate the problems we identify in this Article. There are two basic categories of such additional problems. The first is the balkanization of procedure by having categorical rules that apply to all cases, but having a different set of rules in each local jurisdiction. Mandatory disclosure opt-out provisions continue the practices begun under the Civil Justice Reform Act. Compare FED. R. CIv. P. 26(a)(1) advisory committee's note (noting that a district court may "reject all [mandatory disclosure] requirements for the present") with 28 U.S.C. § 471(a)(4) (Supp. V 1993). While an increase in uncertainty and related costs is to be expected under any altered legal regime, the new discovery rules are likely to compound this short-term problem by allowing local jurisdictions to either bail out from the new procedures or modify them through local rulemaking. As a result, the expected processes of litigation that produce more clearly defined rules of application (the informational externalities identified in POSNER, supra note 48, at 548) will be frustrated by the divergent legal regimes in the various local federal courts. The information benefits of litigation for future disputants will be reduced because they apply only to specific local jurisdictions. An additional immediate concern arises from the disruption of the attorney-client relationship occasioned by the requirement that parties undertake discovery of themselves at their own initiative.

Texas Law Review

[Vol. 73:753

disclosure will reduce expenses in the "mega" cases that most readily evoke the specter of litigation costs run rampant.Y These foreseeable problems lead back to a more fundamental question in civil procedure, that of the tension between "trans-substantive"74 rules and broader standards. As formulated by Professor Hart, all systems of regulation must, of necessity, [C]ompromise between two social needs: the need for certain rules which can, over great areas of conduct, safely be applied by private individuals to themselves without fresh official guidance or weighing up of social issues, and the need to leave open, for later settlement by an informed, official choice, issues which can only be 7properly 5 appreciated and settled when they arise in a concrete case. The amendment of Rule 26 is an example of addressing a specific problem found in a decided minority of cases while retaining a rule of general application for all cases in the federal courts.76 The question remains whether the allure of trans-substantive procedure is worth the anticipated problems with mandatory disclosure.' In other words, why choose to

See Bell et al., supra note 1, at 46 n.175 (citing one lawyer's concern that mandatory disclosure may make clients see their lawyers as suspicious agents of the judicial bureaucracy rather than as trusted advocates). The concern here is that in high-stakes litigation, in which the problem of excessive costs was thought to be of greatest import, the new discovery rules may create greater agency costs by complicating the attorey-client relationship. It is entirely foreseeable that placing the burden of discovery on anticipated trial counsel may lead litigants to retain additional nontrial counsel to protect against overvigilant discovery by their own attorneys. 73. See Anne Y. Shields, The Utility ofDisclosureas a Reform to the PretrialDiscoveryProcess, 67 ST. JOHN'S L. REv. 907, 916-17 (1993) ("Since discovery would almost always be necessary in complex litigation, the general disclosure system would do nothing more than add an extra layer of expense and delay." (footnote omitted)). 74. We use the term trans-substantive to mean procedural rules that operate regardless of the substantive legal dispute at issue. 75. H.L.A. HART, THE CONCEPr OF LAw 127 (1961). 76. Indeed, one of the key conclusions that may be derived from the criticism of mandatory disclosure is that the procedure must become increasingly nuanced to the costs and complexity of the underlying dispute. With the increasing complexity of matters now brought before the federal judiciaries, it may be that, as Adrian Zuckerman notes, the search for a global process regardless of the nature of the dispute, is a romantic notion that is simply unaffordable. A.A.S. Zuckerman, QualityandEconomy in Civil Procedure:The Casefor Commuting CorrectJudgmentsfor Timely Judgments, 14 OXFORD J. LEGAL STUD. 353 (1994).

77. Our critique of mandatory disclosure therefore echoes extensive criticisms in the scholarly literature concerning the fading allure of trans-substantive rules. See, e.g., Stephen B. Burbank, The Costs of Complexity, 85 MICH. L. REV. 1463, 1474 (1987) ("Many of the Federal Rules authorize essentially ad hoc decisions and therefore are trans-substantive in only the most trivial sense.') (reviewing RICHARD L. MARCUS & EDWARD F. SHERMAN, COMPL~x LITIGATION (1985)); Judith Resnick, Failing Faith: Adjudicatory Procedure in Decline, 53 U. CHI. L. REv. 494, 547 (1986) (concluding that trans-substantive rules are "unworkable" and suggesting an alternative, contextual approach). There is a further question whether the rules in practice are indeed trans-substantive in any meaningful sense. See, e.g., Carl Tobias, Public Law Litigation and the Federal Rules of Civil Procedure,74 CORNELL L. REv. 270, 334 (1989) (arguing that "public law litigation appears to differ

1995]

Mandatory Disclosure

proceed through a categorical rule rather than a more attenuated standard that focuses on the capacity for abuse in any given case?78 As a general matter, both broad-gauged rules and context-specific exceptions have benefits and costs. Rules of general application are typically more predictable and have fewer costs associated with their case-by-case utilization. 9 But predictability exacts its own costs in terms 'of misapplication to nuanced situations. As Professor Schauer explains, "Instead of allowing decision-makers to scrutinize a large, complex, and variable array of factors, rules substitute decision based on a smaller number of easily identified, easily applied, and easily externally checked factors. " '° However, as soon as rules become highly attenuated to context, they necessarily devolve into standards that impose costs associated with the uncertainty of their application in each particular case. No systematic effort has been made to determine what the advantages of mandatory discovery will be even in cases that were previously deemed

at risk of discovery abuse.81

Nor has there been any effort made to

selectively police the discovery process through more active intervention of judges or magistrates. This failure is troubling in light of the general

trend in civil procedure toward more active case management by the judiciary.'

The mandatory disclosure rule also appears to violate the

from private disputes in ways that call for distinctive considerationunder certain and perhaps all of the rules"); Samuel Issacharoff& George Loewenstein, Second Thoughts About Summary Judgment, 100 YALE LJ.73, 89-90 (1990) (noting the special reliance on summary judgment in defamation and toxic tort cases). 78. There is a decided trend away from trans-substantive rules in favor of more nuanced roles for judges through the local rulemaking process in developing context specific application of the rules. See Carl Tobias, The Transformation of Trans-Substantivity, 49 WASH. & LEE L. REV. 1501, 1504-05 (1992) (summarizing the contributions of Congress, the federal judiciary, and the Manualfor Complex Litigation to the erosion of the trans-substantive premise of the Federal Rules). 79. See Louis Kaplow, Rules Versus Standards:AnEconomic Analysis, 42 DUKE LJ.557 (1992) (determining that rules are more effective than standards at facilitating prediction of consequences of behavior). 80. FREDERICK SCHAUER, PLAYING BY THE RULES: A PHILOSOPHICAL EXAMINATION OF RULE-

BASED DECISION-MAKING IN LAW AND LIFE 152 (1991). Professor Schauer describes the process of deciding on the level of specificity in commands as a tool "for the allocation ofpower." Id. at 159 (emphasis in original). 81. Indeed, the early experiences with mandatory disclosure were decidedly mixed. In reviewing the data that do exist, Professor Carl Tobias concluded: "The above examination of automatic disclosure's implementation indicates that it undermined uniformity, simplicity, and trans-substantivity, while increasing judicial discretion and expense and delay." Tobias, supra note 3, at 1615 (footnote omitted). Whether the early effects are moderated as lawyers and judges become more familiar with the new procedures remains to be seen. 82. For example, the expanded power of the court to manage each case under Rules 16 and 26(f) gives courts the capacity to mold the application of the rules to the needs of the parties in any particular dispute. In particular, the existence of Rule 26(f) since 1980 gives the federal courts the power to manage the discovery schedules of individual cases and to tailor discovery practices to the actual needs of the litigants.

784

Texas Law Review

[Vol. 73:753

experimental direction charted by the Civil Justice Reform Act of 1990,' which provided for differential management of litigation based on casespecific considerations," judicially supervised case management conferences, and, when possible, voluntary discovery arrangements."' Perhaps the single most important feature was the tailoring of discovery to the particular case under judicial supervision,' which accelerates the power for case management intervention by the federal courts under Rule

16; indeed, this prospect of judicial oversight of discovery was thought particularly promising by litigants.," For example, the Eastern District of Texas has developed a system of

differential case management that requires early court review of the case in order to assign it to one of six discovery tracks. 9 The tracks range from no discovery at all, to prearranged disclosure rules, to disclosure plus limited additional discovery, to specifically tailored discovery programs for a particular case.' Furthermore, there is already an additional rule of procedure

authorizing judicial management of discovery on a case-by-case basis91

83. 28 U.S.C. §§ 471-482 (Supp. V 1993) (requiring each federal district court to formulate and implement a Civil Justice Expense and Delay Reduction Plan). 84. See id. § 473(a)(1) (requiring consideration of case complexity, case preparation time, and the judicial resources necessary to dispose of the case). 85. Id.§ 473(a)(3). 86. Id. § 473(a)(4). 87. Id. § 473(a)(2)(C). 88. Se&Robert L. Haig & Warren N. Stone, Does All This Litigation "Reform" Really Benefit the Cent?, 67 ST. JOHN'S L. REv. 843, 850-52 (1993) (advocating case management plans as an effective means of expediting litigation). 89. For an overview of phased discovery and the specific use of "tracks" to tailor discovery to particular cases, see Edward F. Sherman, A ProcessModel andAgenda for Civil Justice Refoms in the States, 46 STAN. L. REv. 1553, 1566-70 (1994). In addition, the Manualfor Complex Litigation (Second) recommends the use of "waves" of discovery in order to bring discipline to the process and to contain costa in large-scale litigation. MANUAL FOR COMPLEX LITIGATION (SECOND) § 21.421 (1985). 90. The six tracks are: Track One: No discovery. Track Two: Disclosure Only. Track Three: Disclosure plus 15 interrogatories, 15 requests for admission, depositions of the parties, and deposition on written questions of custodians of business records for third parties. Track Four: Disclosure plus 15 interrogatories, 15 requests for admissions, depositions of the parties, depositions on written questions of custodians of business records for third parties, and three other depositions per side (i.e., per party or per group of parties with a common interest). Track Five: A discovery plan tailored by the judicial officer to fit the special management needs of the case. Track Six: Specialized treatment and program as determined by the judicial officers. Order Amending Civil Justice Expense and Delay Reduction Plan, No. 93-13 (E.D. Tex. Sept. 2, 1993). 91. FED. R. Civ. P. 26(f). But see Stephen C. Yeazel, The Misunderstood Consequences of

1995]

Mandatory Disclosure

and one governing discovery abuse that remains virtually unused.' For example, a review of Lexis's "Genfed" database identified fewer than eighty cases that involved Rule 26(g) between 1983 and 1989.94 By contrast, Rule 11, which had fallen into desuetude prior to its amendment in 1983, enjoys a vigorous life-indeed perhaps too much vitality.9' The point is not to extol Rule 11 sanctions as much as to point out that a corresponding standard for policing discovery abuse remained untested at the time of the decision to embark upon sweeping, rule-based reform. Undoubtedly, case-by-case examination of claims of discovery abuse will result in greater demands upon the managerial competence of judges.' It is perhaps because of a reluctance to police discovery practices that go largely unreviewed by either judges or magistrates that the Supreme Court so grandly opted for global reform. 7 Nonetheless, there is little reason to believe that the broad-scale rule approach of Rule 26(a) will yield either systemic.efficiency or greater settlement rates. IV. Conclusion This Article is intended to show the difficulties with responding globally to problems that emerge in a limited minority of actual cases. The mandatory disclosure rule was adopted to address the problems presented graphically in cases in which pretrial discovery costs have spiraled out of control. It is our contention that the likely consequences of this reform should be measured against the purposes of the amended rule. Judged from the vantage point of costs and timing of costs, there are several critiques of the Rule that must be addressed:

Modem Civil Process, 1994 Wis. L. REv. 631, 651 n.66 (noting that the 1993 amendments to the Federal Rules eliminate direct judicial supervision over the initial pretrial conference). 92. FED. R. Civ. P. 26(g). 93. See Lauren Robel, FracturedProcedure: 7he Civil Justice Reform Act of 1990, 46 STAN. L. REV. 1447, 1457 (1994) (observing that judges have power under Federal Rules 16 and 26(f) to manage discovery or force parties to devise cost-effective discovery approaches to particular cases). 94. Melissa L. Nelken, Has the Chancellor Shot Himself in the Foot?Looking for a Middle Ground on Rule 11 Sanctions, 41 HASTINGS L.J. 383, 384 n.8 (1990). 95. See, e.g., Laurens Walker, Avoiding Surprisefrom Federal Civil Rule Making: The Role of Economic Analysis, 23 J. LEGAL STUD. 569, 570-71 (1994) (reporting that prior to the 1983 amendments there were no Rule 11 cases, whereas "today there are more than a thousand reported decisions involving Rule 11 sanctions"). 96. A more case-sensitive approach is found in the experimental practices in the Eastern District of Texas; this approach necessarily imposes greater requirements on the court to appropriately manage each case. See supra notes 89-90 and accompanying text. 97. This approach is consistent with Professor Macey's provocative claim that procedural reforms are largely inspired by the desire of the judiciary to provide increased mechanisms for getting rid of undesirable cases without having to consider the substantive issues involved. See Jonathan R. Macey, JudicialPreferences,Public Choice and the Rules of Procedure,23 J. LEGAL STUD. 627 (1994).

Texas Law Review

1) Cost saving under mandatory disclosure is not that significant because the cost of mechanical filing of routine early discovery is not substantial. 2) Mandatory disclosure is likely to drive up systemic costs as selfdiscovery is commenced under time pressure, even in cases that would otherwise likely settle with little, if any, discovery. 3) Increased up-front costs are unlikely to serve any kind of gatekeeping function because it is the defendant rather than plaintiff who is the primary cost-bearer of this type of early self-discovery. 4) Greater early costs borne by the defendant may give greater economic vitality to marginal suits or even to strike suits. 5) Front-loading costs reduces the disincentives to proceeding once those early costs are borne. By creating these sunk costs, mandatory disclosure reduces the marginal costs of proceeding after early discovery and may adversely affect settlement. 6) The increased early costs of mandatory disclosure may not promote settlement if they do not lead to a convergence of parties' assessments of the merits of the case. This problem is compounded if the disclosed information is subject to self-serving biased judgments that may, in turn, block efficient settlements. Based upon the examination in this Article, it is difficult to escape the conclusion that the mandatory disclosure procedure is misguided. We have examined this discovery reform using models drawn from economics, psychology, and the structures of litigation. Each of these approaches independently counsels against the adoption of mandatory disclosure. Together, they overwhelmingly suggest that the likeliest consequence of mandatory disclosure will be an increase in litigation costs, an increase in strike suits, and a greater difficulty in reaching efficient settlements. This does not bode well for a reform aimed at reducing the costs of litigation.