Litigation realities redux.
Both summarizing recent empirical work and presenting new observations on each of the six phases of a civil lawsuit (forum, pretrial, settlement, trial, judgment, and appeal), the author stresses the needs for and benefits from understanding and using empirical methods in the study and reform of the adjudicatory system's operation.
INTRODUCTION I. FORUM A. Forum Selection B. Forum Effect II. PRETRIAL A. Steps to Termination 1. Pleading 2. Disclosure 3. Discovery 4. Conference 5. Settlement 6. Motion B. Time to Termination III. SETTLEMENT A. Importance of Settlement B. Rate of Settlement IV. TRIAL A. Decline of Civil Trial B. Trial by Judge or Jury V. JUDGMENT A. Win Rates B. Foreigner Effect VI. APPEAL A. Affirmance Effect B. Anti-Plaintiff Effect CONCLUSION
A half-dozen years ago, Ted Eisenberg and I started our article entitled Litigation Realities (1) with a quotation of Louis XVI: his journal entry for July 14, 1789, was "Nothing." (2) Our point was that the modern lawyer who ignores empirical research, even though law has long ignored empirical methods, risks giving in retrospect the very same impression as the French king gave. From that starting point, our article tried to explain empirical methods and map an empirical agenda.
Well, a new age has since dawned. The recent years have seen tremendous advances in empirical studies. Much remains to do, of course. But, as the even earlier French proverb put it (although arguably with inaccuracy under some circumstances), "Something is better than nothing." (3)
I propose in this Article to discuss anew what all of us are now learning about litigation, thanks to this increasing use of empirical methods. I shall again treat separately the six phases of a lawsuit: forum selection, pretrial practice, settlement process, trial practice, judgment entry, and appellate practice. For each, I shall describe what I see as important insights from recent empirical publications, while also providing new data on the realities of that phase. As I shall demonstrate, the last half-dozen years have altered our earlier article's understanding of some features of litigation, which was to be expected given that any initial steps into a new field of study must be tentative ones. Thus, the emphasis in this redoing of Litigation Realities, as compared to the original, will appropriately be less on yesterday's news of empirical methods (4) and more on recent empirical results.
A. Forum Selection
The name of the game is forum shopping, as many have observed elsewhere. (5) Lawyers all know this and have lived by it forever. The contribution of recent empirical research, besides confirming the existence of the phenomenon, has been to show that all of those lawyers were not wasting their clients' money on forum fights--because, in fact, forum matters. Forum is worth fighting over because outcome often turns on forum, as I shall explain in the next subpart.
Forum selection accordingly remains extraordinarily important in the American civil litigation system. Today, after perhaps some initial skirmishing, most cases settle, while few cases reach trial. (6) Yet all cases entail forum selection. The plaintiff's opening moves include shopping for the most favorable forum, be it some state's courts or the federal system, and be it any particular venue within the jurisdiction. (7) Then, the defendant's parries and thrusts might challenge the plaintiffs choice of forum and also might include some forum shopping in return, possibly by removal from state to federal court (8) or by a motion for change of venue. (9) As a consequence, the parties frequently dispute forum. Federal litigators, for example, deal with many more change-of-venue motions than trials. (10) When the dust settles, the case typically does too--but on terms that reflect the results of the shopping and skirmishing. Thus, forum selection is a critical step for litigators, and any fight over forum can be the critical dispute in the case.
When all these individual incentives cumulate, forum selection also becomes a critical concern of the legal system as a whole. Forum selection is very important not only to the litigator, but also to the office lawyer who is drafting contracts with an eye toward possible future litigation. Not surprisingly, there exists an entire treatise devoted to the subject of forum selection. (11) Moreover, the transactional costs of forum shopping, and its effects on outcome and so on justice, should be important to society.
Removal provides a good illustration of forum selection. (12) For background, suppose the plaintiffs commence in a state court an action that they could instead have started in a federal district court. All the served defendants acting together may then seek removal, subject to a few exceptions. (13) The defendants must promptly file, in the federal district court sitting in the same locality, a notice of removal. (14) The defendants must give the plaintiffs and the state court notification of the filing. (15) By this activity solely on the part of the defendants, removal is complete. (16) The state court can proceed no further with the action unless and until the federal district court remands it to the state court, as upon a finding that it was by law not removable. (17) A decision to remand, however, is typically not appealable. (18)
Normally, the defendants can remove only a case that the plaintiffs could have brought in federal court but instead chose to bring in state court. Thus, the group of removed cases are ones where both the plaintiffs and the defendants had a choice of court, but the defendants preferred federal court and possessed the power to trump the plaintiffs' initial choice of state court. The obvious story behind removal, then, is one of forum selection.
What forces drive the parties' choice between state and federal forum? It is not substantive law, as the same substantive law will apply after removal. However, there are many other considerations that might affect choice, according to empirical studies of attorneys' preferences. (19) Most of these considerations group under four general headings: expected bias against a litigant; logistical and practical concerns; perceived disparity in quality and other characteristics between state and federal judges and between state and federal juries; and the different procedures offered by one or the other court system. (20)
What about the numbers? Although the overwhelming majority of all U.S. cases are state cases, (21) a surprising number of those cases are removed to federal court. Consider the data on removal presented in Figure 1, which come from a federal database that I shall be using for all six figures in this Article. (22) The graph shows removal over the thirty-seven-year period for which computerized data exist and are available. (23) The graph focuses on a particular head of federal jurisdiction--diversity jurisdiction (24)--not because the pattern appears only there, even if it is particularly salient there, but because a single context makes expression of analysis clearer. The upper line shows the proportion of diversity cases that originated as removals. The lower line shows the proportion of those removed cases that the district court remanded. (25)
In the original Litigation Realities, we took the data through 2000. The resulting graph (26) exhibited, in its two ascending lines, a surprising time trend. It suggested a removal story of increasing use of removal as a forum selection device and possibly increasing abuse of removal that required more and more remands, a story that nicely conformed with anecdotal impression. (27)
Then, Professors Eisenberg and Morrison extended the graph through 2003, showing that the numbers were staying at their elevated levels. (28) They observed that during the recent period in which state tort filings noticeably decreased, the numbers of both tort and other diversity cases that rested on removal were steadily or even markedly increasing, with their combined percentage of the whole diversity docket mounting well over 30%. (29) Meanwhile, the percentage of those removed cases that the federal courts remanded to state court had climbed toward 20%, which raised even sharper concerns. (30) These remands, by definition, involved erroneous removal, the correction of which very often involves difficult questions of fact and law and almost always involves considerable time and expense that represent a deadweight loss to the system and the parties. Out of a belief that the increase in erroneous removal might entail an increase in abusive removal, they ended by suggesting possible reform that would provide for more frequent fee-shifting against the remanded defendant. (31)
A call to action may have been premature. Figure 1 now goes through 2006. It shows that a more modest upward trend persists for removal (32) but the remand trend has unexpectedly reversed. (33) This very recent dive in remand rate is hard to explain. Perhaps the removing defendants are adjusting to the new regime, so that remand rates can return to their historic level of around 10% of the cases removed.
[FIGURE 1 OMITTED]
B. Forum Effect
Let me return to the effect of forum selection on the outcome of cases, and let me define "win rate" as the fraction of plaintiff wins among all judgments for either plaintiff or defendant. (34) Application of empirical methods can then reveal the effect of forum in the context of removal, (35) as well as in the analogous context of transfer of venue between federal district courts. (36)
Our removal article showed that plaintiffs' win rates in removed cases are very low, compared to original cases in federal court and to state cases. For example, the win rate in original diversity cases is 71%, but for removed diversity cases only 34%. (37) The explanation could be the ready one based on the purpose of removal: the defendants thereby defeat the plaintiffs' forum advantage and shift the biases, inconveniences, court quality, and procedural law in the defendants' favor. Alternatively, the explanation might lie not in forum impact but instead in case selection: removed cases may simply be a set of weak cases (1) involving out-of-state defendants who have satisfied or settled all but plaintiffs' weakest cases or (2) involving plaintiffs' attorneys who have demonstrated their incompetence by not avoiding removal. Our analysis indicated that both forum impact and case selection are at work. After a regression controlling for many case variables, which is a statistical technique that helps to account for differences among the cases and thus to neutralize the case-selection effect, the impact of removal remains sizable and significant. Forum really does affect outcome, with removal taking the defendant to a much more favorable forum. The statistical analysis indicates a residual removal effect for diversity cases that would reduce 50% odds for plaintiff to about 39%. (38) This eleven point reduction from even odds represents the impact of a federal forum on the case--the removal effect. (39)
We also studied the transfer effect, whereby the win rate drops markedly after transfer of venue between federal districts. Plaintiffs' win rate in all federal civil cases drops from 58%, calculated for cases in which there is no transfer, to 29% in transferred cases. (40) For transfer, the loss of a favorable forum, with the result of a strongly shifted balance of inconveniences and a shift of local biases, seems to be the primary explanation, because we were able more easily to discount explanations based on differences in the strength of nontransferred and transferred cases. (41) That is, the win rate declines largely because the plaintiffs have lost a forum advantage. The plaintiff's odds would drop after transfer of venue from 50% to 40%, after controlling for all available variables. (42)
The comparison of removal and transfer suggests a consistent forum effect, whereby the plaintiffs' loss of forum advantage by removal or transfer reduces their chance of winning by about one-fifth. This empirical finding is important, even if it seems an unsurprising confirmation of what most lawyers already knew: the name of the game indeed is forum shopping.
Policymakers, in addition to practitioners, obviously take interest in this finding. For example, Congress recently enacted the Class Action Fairness Act (CAFA), (43) the single most important piece of class action legislation in the nation's history. The Republican Congress gave CAFA a broad scope covering interstate class actions, with the expressed intent of defeating plaintiff lawyers' manipulation of state courts. (44) When Republican President George W. Bush signed it into law, he declared that it "marks a critical step toward ending the lawsuit culture in our country." (45) The statute's method was to funnel more class actions away from the state courts and into the federal courts, and perhaps thereby to discourage class actions.
Most important for present purposes was CAFA's expansion of federal subject matter jurisdiction for class actions. Congress bestowed original jurisdiction on the federal district courts for sizable multistate class actions, generally those in which the plaintiff class contains at least 100 members and their claims aggregated together exceed $5 million, exclusive of interest and costs; the statute does not require complete diversity, but rather minimal diversity, which means that only one plaintiff member of the class must differ in citizenship from any one defendant. (46) Congress further provided that any defendant can remove a class action from state court to the local federal district court if the action would be within the original federal jurisdiction; the statute goes on to say that the removing defendant can be a local citizen and need not seek the consent of the other defendants, and that any decision as to remand is immediately appealable. (47)
We recently conducted an empirical study of all online judicial opinions on CAFA in order to gauge judicial activity and receptivity in regard to this legislative attempt to tilt the field in defendants' favor. (48) We found that CAFA has produced an unusually large amount of litigation in its short life. (49) The cases were varied, of course, but most typically the federal decision involved a removed contract case, with the dispute turning on CAFA's effective date or on federal jurisdiction. (50) More interesting, we saw wise but value-laden resistance to CAFA by federal trial and appellate judges. (51) By an almost two-to-one rate, the judges construed or applied it in a way that narrowed rather than broadened it, although Republican male judges stood out as bucking the trend. (52) In general, the federal judiciary has not warmly embraced the statute, dampening the early hopes of overly enthusiastic removers.
A. Steps to Termination
The pretrial phase divides into case exposition and case disposition. These two involve both devices driven in the main by the parties' own efforts and also devices driven in the main by forces external to the parties. Case exposition's internal devices are pleading and disclosure, while its external devices are discovery and conference. Case disposition's internal devices comprise the processes of settlement, and its external devices comprise the methods of summary adjudication by motion. I shall run quickly through those six steps of pretrial.
The content of the pleadings was long a controversial subject in Anglo-American practice. (53) The older view held that pleadings must accomplish a great deal, laying out the issues in dispute and stating the facts in considerable detail. (54) But holders of this view asked too much of the pleading step, which consequently became the center of legal attention, ended up all too often mired down in battles over technicalities, and provided the vehicle for monumental abuse. (55) Modern practice has shifted forward most of the former functions of pleadings, moving them into the steps of disclosure, discovery, pretrial conference, summary judgment, and trial. The motivating theory was that these later steps can more efficiently and fairly handle functions such as narrowing issues and revealing facts, and thus the whole system can better deliver a proper decision on the merits. (56) Accordingly, most people came to accept that the main task of pleadings is to give the adversary (and the court and the public) fair notice of the pleader's contentions. (57)
Despite this modern prevalence of notice pleading, disagreement on the critical question of how pleading should proceed is now recommencing. (58) Lower federal courts of late have been requiring greater detail in pleading, or so-called heightened pleading. (59) The Supreme Court has twice batted down this change for being unauthorized by the Federal Rules. (60) Yet change keeps coming--even though there have been no empirical studies whatsoever on the virtues of case exposition through pleading. (61)
The shot-in-the-dark adjudication in Bell Atlantic Corp. v. Twombly (62) nicely represents recent "reform" of pleading. In that case, telephone and internet subscribers brought a class action against the telecommunications giants, claiming an illegal conspiracy in restraint of trade. (63) Under antitrust law, however, parallel and even consciously identical conduct unfavorable to competition is not illegal if it involves only independent actions by competitors without any agreement. (64) The complaint alleged parallel conduct in great detail, explaining how each company sought to inhibit upstarts in its own region and refrained from entering the other major companies' regions. (65) But it alleged an agreement mainly in conclusory terms upon information and belief, because the plaintiffs had no proof yet in hand. (66)
The legal system's concern in this big complex case was obviously with opening the door to the plaintiffs' expensive discovery. So, the Supreme Court ordered dismissal on a pre-answer motion for failure to state a claim upon which relief could be granted, holding that the complaint failed to show its allegation of agreement to be "plausible." (67) According to the Court, the defendants' alleged behavior was merely what each company would have naturally done in pursuit of its own interests. (68) The plaintiffs needed to give factual detail to make their complaint plausible, but they "mentioned no specific time, place, or person involved in the alleged conspiracies." (69) Dismissal followed for these plaintiffs who had "not nudged their claims across the line from conceivable to plausible." (70)
In so ruling, the Court invented a plausibility test for the pleading stage. This gatekeeping move represents the Court's first unmistakable step backward from the modern conception of notice pleading. The Court did not step in the direction of simply requiring heightened detail in allegations, but instead it instituted a judicial inquiry into the pleading's convincingness.
Justice Stevens, joined in relevant part by Justice Ginsburg, dissented. He saw the decision as a "dramatic departure from settled procedural law," and an unjustified one because it should have come if at all by amendment to the Federal Rules or by statute. (71) He lamented that by imposing a plausibility test on pleadings
the Court succumbs to the temptation that previous Courts have steadfastly resisted.... Here, the failure the majority identifies is not a failure of notice--which "notice pleading" rightly condemns--but rather a failure to satisfy the Court that the agreement alleged might plausibly have occurred. That being a question not of notice but of proof [courts will now have] to engage in armchair economics at the pleading stage [in order to ascertain somehow whether the complaint's pleaded facts adequately show liability]. (72)
Bell Atlantic, during its short life, has triggered tremendous confusion in case (73) and commentary. (74) What exactly it meant is clearly open to dispute, as is the wisdom of imposing, with no forewarning or public discussion, any sort of plausibility test on pleading. But of importance here is the fact that the Court acted with no empirical support that a problem existed, and with no exploration of the dimensions of that problem or the efficacy of the Court's newfangled cure.
Empirical work on the effects of Bell Atlantic is just getting started. In the first study, (75) a law student examined the reported cases and found that the courts are frequently and widely applying the case, in fields way beyond antitrust. (76) But courts do not seem to be dismissing cases at a significantly higher rate, except for civil rights cases. In that latter category, he showed that the rate of granting dismissal jumped by eleven points. (77)
As to case exposition through disclosure, "reform" came instead through rule amendment. Consequently, the adoption of mandatory disclosure was more comprehensible, but it came equally without an empirical basis. It proved to be one of the most controversial pretrial reforms of recent times. (78)
The federal rulemakers introduced mandatory disclosure in 1993. (79) Parties now must disclose certain core information, elaborating on the pleaded facts without awaiting a discovery request. Under Federal Rule of Civil Procedure 26(a), there are three distinct types of disclosure (80): initial disclosures, expert information, and pretrial disclosures. In particular as to the so-called initial disclosures, the adopted Rule 26 (a) (1) required disclosure of routine evidentiary and insurance matters. These matters comprised (1) witnesses "likely to have discoverable information relevant to disputed facts alleged with particularity in the pleadings," (2) documents and things "in the possession, custody, or control of the party that are relevant to disputed facts alleged with particularity in the pleadings," (3) computation of claimed damages, and (4) insurance agreements that might cover part or all of an eventual judgment. (81) However, federal districts by local rule could alter these initial disclosure obligations. Indeed, almost half the districts opted out of the standard scheme by diminishing initial disclosures to some degree. (82)
The federal rulemakers' introduction of disclosure aimed at achieving some savings in expense and delay, and also at moderating litigants' adversary behavior in the pretrial process. (83) The rulemakers credited as their inspiration the anecdotal advocacy (84) in a law review article by Professor Wayne Brazil (85) and another by Judge William Schwarzer. (86) However, critics claimed that disclosure, in its routine operation and by the inevitably ensuing disputes, would actually increase expenses and delays; (87) also, the critics argued that disclosure would counterproductively clash with the prevailing adversary system and with the notice pleading scheme. (88) After the rulemakers' introduction of disclosure, the unabating controversy prompted them finally to commission empirical studies, by both the Federal Judicial Center (FJC) (89) and the RAND Institute for Civil Justice (RAND). (90)
The FJC reported a survey of 2000 attorneys involved in 1000 general civil cases terminated in 1996 that were likely to have had some discovery activities, a survey with a 59% response rate. (91) Most of the responding attorneys felt that initial disclosure had had no effect on delay or fairness, but among those who detected effects, more attorneys believed the effects to be positive rather than negative. (92) Also, the respondents rarely reported fears of increased satellite litigation. (93) Finally, by statistical analysis of its small sample of cases, the FJC found that the use of initial disclosure tended to shorten actual disposition time. (94)
The RAND report used its preexisting data to compare a small group of district courts with local rules requiring some type of disclosure during 1992-1993 to another small group with no such rules. (95) The data included the attorneys' subjective measure of satisfaction and sense of fairness, as well as objective measures of attorneys' hours worked and case disposition time. (96) RAND found no significant effect of disclosure on fairness sensed, hours worked, or disposition time. (97) But mandatory disclosure did markedly lower attorney satisfaction. (98)
In 2000, based on these two imperfect studies, the rulemakers profoundly amended Rule 26(a)(1). Although they now prohibited the district courts from opting out of the requirements, the rulemakers exempted eight specified categories of proceedings from initial disclosure and, most importantly, reduced the scope of the initial disclosure. (99) Henceforth, a party needed to disclose only those witnesses, as well as those documents and things in the party's custody or control, "that the disclosing party may use to support its claims or defenses." (100) However, such disclosures of favorable information no longer needed to be triggered by "disputed facts alleged with particularity in the pleadings." (101)
Professor Kuo-Chang Huang, when still a graduate student at Cornell Law School, recognized the shortcomings of the two previous studies and then performed his own clever study of disclosure by using Administrative Office data. (102) Among other statistical analyses, he "vertically" compared disposition time in the years before a district court required initial disclosure with disposition time after adoption of such disclosure. (103) He also "horizontally" compared district courts that required initial disclosure with district courts that had opted out of such disclosure. (104) By multivariate regression, Professor Huang showed that adoption of disclosure tended slightly but significantly to slow down disposition. (105) He concluded that with almost no practical effects, this controversial device has no justification. (106) Thus, the rulemakers in 2000 would have been better advised just to eliminate initial disclosure. (107)
As to case exposition's external device of discovery, the story is much the same as for disclosure. Rule amendments over recent decades have been remarkably frequent, but unremarkably reliant on logic and anecdote alone. (108) In fact, over the course of its existence, despite the revolution worked by it, the discovery scheme has seen very little in the way of systematic empirical study. (109)
Nevertheless, this situation may be brightening. Again, Professor Huang in a recent study (110) shined some light on Taiwan, (111) which, at the least, reflected to the United States. He looked at Taiwan's settlement rate before and after its major procedural reform of 2000 that introduced the concept of discovery. (112) In this first-ever vertical study of an introduction of discovery, he found that the settlement rate, which had been steadily decreasing for some reason through 2000, significantly reversed its direction to increase steadily after 2000. (113) After controlling for various variables, he came to think that introducing discovery could be the cause of the new trend and that similar reform should raise other civil law countries' traditionally low settlement rates--with discovery presumably contributing to settlement by decreasing informational asymmetry as to trial evidence. (114)
As to case exposition's other external device, pretrial conference, the story stays the same. Rule amendments over recent decades again have been frequent, but still reliant on guesswork. Despite some promising early empirical work, the system's continuing reliance on conferences has seen virtually nothing in the way of systematic empirical study. (115)
The lesson here is an obvious one. Not only do practitioners and students need to attend to empirical methods, but so do commentators on the legal system ranging from academics to journalists. Most of all, empirical studies must be put before those who govern the system. Indeed, there is a "compelling need for public policymakers to commission expert, independent evaluations that systematically gather, analyze, and synthesize dependable empirical data." (116) The data might come from archival research of some sort, or they might even come from experimental research such as field experiments conducted under the authority of new local rules. (117) Then the policymakers must "closely consult and carefully apply the material assembled when reforming civil justice." (118) That is, there is a demand-side problem as well as a supply-side problem with empirical studies: almost nobody in power pays attention to the few studies that do exist. (119) Official reformers have proceeded largely on the basis of intuition in overhauling pleading and motion practice, while adding disclosure, discovery, conference, and settlement mechanisms.
On case disposition through settlement, we know neither how much settlement is optimal, nor how much settlement we are experiencing. Nevertheless, because settlement is so important, and because the settlement processes extend in time before and after the pretrial phase, I shall discuss them separately in the next Part.
As to case disposition through pretrial motions, the device of greatest interest in modern times has been summary judgment. (120) It is the important tool for determining whether trial is necessary, a tool that nicely complements the foregoing pretrial scheme, which features notice pleading and extensive discovery. (121)
Summary judgment allows the court to decide legal disputes, without trial, when there are no genuine and material factual disputes. Under Federal Rule 56, summary judgment will be given to a movant "entitled to judgment as a matter of law" if "there is no genuine issue" as to any fact that is material to the case (or as to any material application of a legal standard to the facts). (122) A "genuine" factual dispute equates to a triable one, which would require the motion to be denied and trial awaited. The principal inquiry on the motion is therefore whether any such factual disputes truly exist, never how to resolve factual disputes that do exist.
In determining whether there is a genuine issue as to any fact, the court construes all factual matters in the light reasonably most favorable to the party opposing the motion and then asks whether reasonable minds could differ as to the fact's existence. (123) That is, summary judgment can be granted if, looking only at all the evidence that is favorable to the opponent of the motion and also the unquestionable evidence that is favorable to the movant, the judge believes that a reasonable factfinder could not find for the opponent. Under this standard, disputes on the papers as to objective facts can sometimes be resolved by overwhelming evidence that removes all reasonable doubt, but disputes that turn on credibility cannot. (124) Accordingly, it is easier to obtain summary judgment against the party who will bear at least the burden of production at trial, although even a party who will bear the burden of both production and persuasion at trial can sometimes properly obtain summary judgment with a sufficiently strong showing.
Despite summary judgment's importance, our knowledge of its workings has always been scanty. (125) Much uncertainty existed about how much summary judgment activity was going on. Further details, some as important as the plaintiffs' and defendants' relative success rates on such motions, remained even more obscure. This ignorance provided the backdrop for ongoing academic dispute over the appropriate standard for granting summary judgment. (126) Most agreed that providing for summary judgment is a good idea, especially in today's strained procedural system, because it allows weeding out those cases that do not require trial at all. But, of course, the system must avoid an overuse of summary judgment that would undercut the right to trial. So, how tough should the standard for summary judgment be? Some worried academics would have restricted summary judgment by toughening the prevailing standard (127)--which knocked out cases when one side was being irrational in disputing the facts--while others would have loosened it in the name of efficiency. (128)
Just three weeks before Bell Atlantic, the Supreme Court stepped in to encourage use of summary judgment as another way for judges to short circuit litigation, with the Court taking a very activist role in drawing inferences from the record in order to reverse a denial of summary judgment. Scott v. Harris (129) was this shot in the dark. Again, the Court seemed to rely on logic and anecdote, rather than on an accurate sense of how often parties were making and winning summary judgment motions, to rein in the perceived excesses of today's litigation.
The case was a civil rights action complaining of the conduct of a police officer in pursuing an automobile, which he had clocked at seventy-three mph in a fifty-five mph zone. (130) The officer ultimately bumped the car, causing it to crash and thereby grievously injuring the plaintiff driver. (131) The Fourth Amendment required that the police behavior be objectively reasonable, and so the summary judgment question boiled down to whether the defendant could, as a matter of law, defeat a finding of violation of the Fourth Amendment. (132) The Supreme Court, after viewing a videotape of the chase, held that the plaintiff's conduct posed a risk of imminent harm to others substantial enough to justify the police conduct. (133) That is, the plaintiff's "version of events is so utterly discredited by the record that no reasonable jury could have believed him." (134)
Justice Stevens, the sole dissenter, objected to "this unprecedented departure from our well-settled standard." (135) He contended that the videotape "surely does not provide a principled basis for depriving the [plaintiff] of his right to have a jury evaluate the question whether the police officers' decision to use deadly force to bring the chase to an end was reasonable." (136) Indeed, he found the video ambiguous, (137) and also pointed out that the majority was purely speculating as to matters such as what would have happened if the police had simply ceased their pursuit. "In my judgment, jurors in Georgia should be allowed to evaluate the reasonableness of the decision to ram respondent's speeding vehicle in a manner that created an obvious risk of death and has in fact made him a quadriplegic at the age of 19." (138)
By contrast, the Federal Judicial Center has recently released the premier published study of summary judgment. (139) It looked at a sample of federal civil cases (excluding prisoner, Social Security, and benefit repayment cases) in six districts from 1975 to 2000. (140) While emphasizing that summary judgment practice varies considerably with locale and case type, (141) it found overall that the percentage of cases involving one or more summary judgment motions increased from 12% in fiscal year 1975 to 20% in calendar year 2000; (142) the court granted such a motion in full or in part in 6% or 12%, in those respective years, of all cases in the sample; (143) and grant of summary judgment resulted in termination of 3.7% or 7.8%, respectively, of the sample. (144) It suggested that the modern ascendancy of summary judgment dates from the upswing in the late 1970s of judicial case management and its emphasis on motion practice. (145) Incidentally, among all the summary judgment motions over the whole time period studied, viewed on a motion level rather than a case level, 72% were motions by defendants (with a 49% rate of granting in full or in part in 2000), while 28% were plaintiffs' motions (with a 36% success rate in 2000). (146)
To conclude this discussion of pretrial practice on a happier note, the federal rulemakers are now readying to amend Rule 56, with a projected effective date of December 1, 2010. (147) They would include new procedures requiring parties to state the facts assertedly uncontested and to respond thereto; the rule also would clarify judicial options when a party fails to respond to a motion and would recognize the practice of moving for partial summary judgment. (148) The encouraging development here is not the detailed proposal, which very well may be undesirable, but the fact that the rulemakers are heavily relying on the Federal Judicial Center to get empirical support in advance of amending the rule. (149)
B. Time to Termination
As the number of pretrial steps might suggest, the pretrial phase is the lengthiest phase of litigation. Naturally, then, it has been the focus of recent reform efforts to speed up litigation.
It is not surprising that reformers focus on delay in litigation, whether in the pretrial phase or in the other phases of a lawsuit. "Delay in the courts is unqualifiedly bad." (150) Justice delayed is justice denied, after all. And there is plenty of judicial delay for everyone. But caution!
Although Figure 2 might show delay, it does not support a view that the problem has increased recently, even though here I have extended it to cover the most recent years. The upper dashed line shows the slightly increasing average time from filing to termination for all federal civil cases in which the procedural progress code indicates resolution during or after trial. More importantly, the lower solid line shows the time to termination for the much more numerous cases resolved before trial begins (over the thirty-seven-year period, the increasing percentage that are pretrial terminations has averaged ninety-six percent). These untried cases do not take that long, and the length has not increased over the years despite the considerable increase in the courts' caseloads shown by the dotted line.
Moreover, there are other good reasons to proceed with wariness before accepting the truth of either old maxims about delay or the potential of new reforms based merely on intuition. Both recent theoretical work and recent empirical study argue for such caution.
Theoretical work contends that delay is not necessarily an evil. (151) Delay is an unavoidable feature of life, and it is not an evil in itself. The only evil is excessive delay, where excessive means that the costs of delay outweigh its benefits. The costs of figuratively queuing to try a case tend to be exaggerated, because we overlook that the parties can engage in other pursuits while waiting. Meanwhile, queuing in fact has some benefits, such as lowering the demand for expensive trials.
[FIGURE 2 OMITTED]
Another study, (152) both theoretical and empirical, shows that many obvious reforms simply have not worked and will not work to reduce delay. (153) The basic insight is that any reduction in delay increases the incentives to litigate and reduces the parties' incentives to settle, with the consequent increase in caseload offsetting the reduction in delay. Therefore, most attempts at reform, such as heeding the constant call for adding judges, will only increase the number of dispositions, rather than decreasing the time to disposition. Adding judges to the system to reduce congestion is similar to expanding the number of lanes on a freeway, an improvement that would draw traffic off the side streets and from public transportation. More cases might flow into the system, and the lesser burden of litigating might reduce the subsequent incentives to settle rather than litigate, so the increased number of judges would be able to adjudicate basically the same percentage of cases. Indeed, the author of the study postulated a "congestion equilibrium hypothesis": almost all reform attempts to accelerate litigation will be largely offset by increases in the amount of litigation. (154)
Pure empirical work in this area is rather rare because of the scarcity of data and the inherently complex nature of the relevant research questions. It is unclear even what to measure, no less how to measure in a controlled way. But the empirical work that exists, while suggesting that delay is neither that lengthy nor increasing recently, is otherwise consistently discouraging for the persistent reformer. One study used state data to demonstrate that particular processes, such as alternative dispute resolution (ADR), do not correlate with shortened disposition times--meanwhile, the factors that do so correlate, such as forum locale and case category, are beyond the reach of process-oriented reform. (155)
In sum, assumptions about delay are risky, making empirical study a necessity. But the caution of conducting empirical studies warrants yet more caution. Some related empirical work that Professor Eisenberg and I have done counsels this doubled caution.
Once again using the Administrative Office database of federal civil cases, but now limited to sizable tort and contract categories that clearly involved a choice between jury and judge trial, we showed that while the actual jury trials themselves may proceed twice as slowly as bench trials conducted by a judge without a jury, over their lives on the docket such judge-tried cases last significantly longer than jury-tried cases: the median judge-tried case spends 619 days on the district court docket, compared to the median jury-tried case terminating in 566 days. (156) That is, although most commentators assume that the wait in the jury queue is uniformly longer than the wait for a judge's trial and decision, (157) the reality in federal courts is the opposite. After regression and other analyses, we found the most likely explanation to be that the press of other duties leads judges to interrupt bench trials, to postpone issuing their eventual decision, and thereby to slow down judge-tried cases. (158) Consequently, a reform aimed at restricting jury trials in order to reduce delay is apt to be counterproductive. (159)
Here the new development since the original Litigation Realities was learning that current data on tort and contract cases from state courts of general jurisdiction suggest that jury cases do in fact last significantly longer than bench cases: the median disposition time in the state courts for a jury-tried case was 21.7 months and for a judge-tried case only 16.1 months. (160) For that result, the researchers looked at one year's trials in the nation's seventy-five largest counties. (161) Although, of course, some unperceived selection effect could be at play, any difference between the data sets that would explain this federal/state difference is not readily apparent. For example, the data sets comprised completed trials in comparable case categories occurring in comparable proportions, and the federal/state difference was quite consistent across those case categories. (162)
The resulting contrast between federal and state courts demonstrates that before drawing conclusions, one must either ensure that most pieces of the empirical puzzle are visible or restrain one's conclusions accordingly. An examination of the underlying state database confirms that state jury trials start much later in a case's life than state bench trials do, which is not the case in federal courts. (163) Another important observation is that state judges do not delay nearly as long after the end of bench trial before issuing a decision as do federal judges. (164) It therefore seems that the state courts, unlike the federal courts, are imposing waiting costs upon those who wish a jury trial and not on those who agree to a bench trial, with the effect of discouraging jury trials. (165) Not all states follow this practice, as some adopt a more neutral approach or the federal approach. (166) Nonetheless, although ultimately a matter of local culture, most states act, whether intended or not, in ways that tend to discourage jury trials--while federal courts do not.
A. Importance of Settlement
Most lawsuits do not make it all the way through the pretrial practice I have just examined. Indeed, most disputes do not even become lawsuits in the first place. Injured persons abandon or settle the overwhelming majority of grievances at some point along the line. (167)
A useful mental image is the so-called grievance pyramid, which from its broad bottom of the whole realm of human experience narrows upward to injurious experiences, grievances, claims, and then to disputes, a set that in turn produces the subset of litigated cases. (168) As one progresses up the steps of the pyramid, most cases peel off through avoidance, nonperception, acceptance, settlement, or some other alternative to litigation. For example, only a subset of grievances ripen into claims when the aggrieved voices the grievance to the injurer--most aggrieved persons accept their injury, viewing it as part of life or just figuring that no remedy is available. Similarly, most disputants never make it to a lawyer, much less to a courthouse. Thus, infinite experiences produce countless disputes, which yield few cases.
From the viewpoint of the civil justice system, settlement fills a critical need. Ours is a slow and expensive procedure. The system simply would not be able to adjudicate all cases conceivable or even all those filed. We depend on the parties finding alternatives to using the system. Accordingly, reformers are constantly seeking ways to increase the settlement rate (which is a loose term that here measures the percentage of filed cases leaving the sides of the grievance pyramid, whether by abandonment, concession, or privately negotiated settlement or by ADR such as arbitration, mediation, and conciliation). Many reformers contend that the settlement rate in the United States today lies below the optimum. (169)
Nevertheless, the system must adjudicate some cases in order to pronounce the law. These cases set the standards under which the parties negotiate settlement of their disputes. The parties can thereby "bargain in the shadow of the law" to reach outcomes that generally conform to the law and thereby further the law's purposes. (170) Thus, the settlement rate could conceivably become too high. If parties settled all cases, there would be big gaps in the law that is supposed to be setting the standards for settlement. (171) If courts adjudicated some cases, but still too few, the gaps might be smaller but the law would remain not only inefficiently fuzzy but also insufficiently conformed to social purposes. But at some lower settlement rate, the law would be optimally set so that further adjudication would be wasteful. (172)
Shifting from the viewpoint of the system to that of the disputants, settlement is also of critical importance. For them, in the usual course, settlement is our system of justice (and for their "trial" lawyers, negotiation of settlements--and pursuit of other alternatives to litigation--is what their profession primarily entails). Alternatives to litigation usually offer procedural and substantive advantages to the disputants. (173) Again, however, some optimal settlement rate exists, above which the increases in external pressure to settle would impose undesirable costs on party autonomy. (174)
B. Rate of Settlement
Despite its undeniable importance, we do not know much about the actualities of settlement. (175) It is hard to observe. Litigation, with its judgments, is much more observable than settlement. Empirical studies tend to focus on the readily observable.
It is obvious, nevertheless, that the settlement rate is high. Alternatively put, the slope of the sides of the grievance pyramid is quite gentle, so that a huge percentage of situations leave the pyramid at each step upward. A telephone survey in January 1980 of more than five thousand households indicated that during the previous three years just over a third of the households had perceived one or more grievances of certain litigable types; 71.8% of those grievances produced a claim informally; 62.6% of those claims met an initial rebuff to produce a dispute; and 11.2% of those disputes resulted in filing a lawsuit. (176) Indeed, these percentages are exaggeratedly high, because the survey limited its inquiries to grievances involving $1000 or more. But even for such substantial grievances, litigation is by no means a knee-jerk or common reaction in the United States, as overall only 5% of the survey's grievances ultimately resulted in a court filing. (177)
Do we have a better idea of the settlement rate for filed cases? Everyone knows that in this world of litigation at the top of the pyramid, the sides' slope remains gentle. Of the relatively few filed cases, only a small percentage make it through the procedural system to a contested judgment. Seeking greater specificity, the original Litigation Realities divided all federal civil cases among four sets of disposition methods (settlement, pretrial adjudication, trial adjudication, and other). (178) Since then, however, another researcher, in a difficult article, made the important point that the Administrative Office codes relevant to settlement are ambiguous and hence treacherous, and even more so in regard to data drawn from different years. (179) While her research counsels against making fine distinctions or drawing strong conclusions from the Administrative Office data as to settlement, it also demands my reallocating the codes to refine the four sets of disposition methods. (180) After reallocating, I can take as an illustration all the 271,753 federal civil cases terminated in all federal districts during fiscal year 2005. (181) Of these, approximately 67.7% were coded as settled in one way or another; around 20.7% were adjudicated at the pretrial phase, as by a motion under Federal Rule 12 or 56; about 1.3% were adjudicated at the trial phase; and the other 10.3% of the cases fell (182) into a welter of other classification codes, predominantly remand or transfer to another court, whereby most would result as a matter of probability in an eventual settlement rather than a final adjudication.
I can then combine all these rough numbers with the visual presentation of the grievance pyramid. From the experiential infinitude, imagine that 1000 sizable grievances arise. This typical thousand will decrease to 718 claims, 449 disputes, 50 filed cases, 12 litigated judgments and 1 decided appeal. (183) Thus, I advisedly described the pyramid's sides by saying that their slope is gentle.
I can now also redraw and extend Figure 3 beyond what appeared in the original Litigation Realities. The coding reallocation raises the rate of pretrial adjudication while lowering the rate of "other" dispositions, compared to the original version of the graph. But the percentage of these other dispositions still increases with time. Because most of these other dispositions will result eventually in agreed settlement, if not immediate abandonment, the overall rate of settlement is holding about constant. Thus, Figure 3 continues to tell the same basic story of the continuing dominance of settlement, against the backdrop of a growing role for pretrial adjudication and a diminishing role for civil trial.
A. Decline of Civil Trial
As nontrial terminations of various sorts have increased, the civil trial has all but disappeared. (184) Many have noted this trend of late, (185) although there is less agreement on cause. (186) The phenomenon does not appear to be limited to civil cases, (187) or to federal courts, (188) or to the United States for that matter. (189)
[FIGURE 3 OMITTED]
Interestingly, judge trial fell even more precipitously than jury trial in federal civil cases. This development is especially mysterious because both queues for trial pass through the regulation of the same person, the trial judge. (190) Perhaps the explanation lies in judicial distaste for a consuming task like bench trial; or, as the disincentives to any trial have increased, those litigants who prefer jury trial have proved to be the more determined group. But a major contributing factor is that, as already explained, (191) the federal courts do nothing to discourage jury trial as they neither make litigants wait longer nor impose any special user fee for this costly form of trial, but rather discourage judge trial by protracting it. Parties, either of which can opt for jury trial, act upon those economic incentives.
Extended through 2006, Figure 4 presents some suggestive data on these time trends. (192) The solid and dashed lines show jury and judge trials, respectively, as a percentage of all federal civil terminations. These percentages have decreased with the passing years, and the judge line fell more sharply than the jury line until 2000. The dotted line shows the result in the form of an increasing ratio of jury trials to judge trials. For example, in 1979 there was one jury trial for every two judge trials, and by 2006 there were more than two jury trials for every judge trial. A lot more analysis remains necessary to get a secure take on the causes, or even on the real size of the declines given a changing legal environment. (193) And then there would remain the contentious issue of the normative implications of the vanishing civil trial and bench trial. (194)
Nevertheless, some stories are becoming apparent. Figures 2, 3, and 4, taken together, describe an adjudicatory system in equilibrium. In the mid-1970s a butterfly beat its wings in Brazil or, more directly, the population of the United States grew, and so court filings started going up. The lawmakers increased the system's capacity, but not enough, partly because having more judges induced more filings. (195) The judiciary reacted with increased attention to judicial management and a new emphasis on motion practice, so that the granting of summary judgment rose. (196) Parties refusing to settle had to wait slightly longer for trial, and even longer for bench trial. This disincentive, by lowering the expected value of a tried judgment, decreased the number of trials. But the disincentive was not enough to reverse the increase in case filings, and so the cycle continued. More cases came, of which more ended early and fewer reached trial. The last three decades' drastic increase in caseload would necessarily correlate with drastic changes elsewhere in the system, such as the observed drastic decline in civil trials. But all the changes offset one another, so that the equilibrium held as far as delay goes. Indeed, the time to termination for the vast majority of cases stayed level. (197)
[FIGURE 4 OMITTED]
In this kind of dynamically interactive system, it makes little sense to speak of cause and effect. But people do, producing a chorus of voices identifying an array of causes. (198) Most of those voices stress to some degree the courts' increasing workload. And in some sense, it is sound to say that the increase in caseload resulted in the decline of the civil trial. (199) The number of trials had to decrease, and the number of summary judgments had to increase, so keeping the delay in handling the growing caseload at its natural level. But in another sense, this cause is illusory. The system could have prohibited summary judgment, for purposes of illustration, and another equilibrium position would emerge with increased trials and settlements and decreased filings. Thus, one could speak of the failure to prohibit summary judgment as the cause of the decline of the civil trial. The better way to speak is of a dynamically interactive system.
B. Trial by Jury or Judge
The classic work on jury and judge differences was by Professors Harry Kalven and Hans Zeisel. (200) They addressed the reliability (the ability to treat like cases alike) of jury decisionmaking, as opposed to validity (correctness). Their questionnaires to presiding judges in some 4000 actual state and federal civil jury trials nationwide in the 1950s--asking the judges how they would decide those same cases, a decision supposedly formulated before the verdict but reported afterwards--yielded data showing a 78% agreement between judge and jury on liability. (201) The rate of agreement is more impressive than it first appears. The cases that reach trial are close cases. When compared to other human decisionmakers, this 78% agreement rate proves better than the rate of agreement on dichotomous decisions between scientists doing peer review, employment interviewers ranking applicants, and physicians diagnosing patients, and almost as good as the 79% or 80% rate of agreement between judges themselves making sentencing decisions on custody or no custody in an experimental setting. (202)
Incidentally, when judge and jury did disagree in the Kalven and Zeisel study, they exhibited no distinct pattern other than the juries' very small tendency to favor plaintiffs more than judges did. (203) The jury but not the judge found for the plaintiff in 12% of the cases, while the judge but not the jury found for the plaintiff in 10% of the cases. (204)
A quarter-century later we performed the first large-scale comparison of plaintiff win rates and recoveries in federal civil cases that actually went to trial before either juries or judges. (205) Unlike Kalven and Zeisel, we took case outcomes in the stream going through jury trial and compared them to outcomes after bench trials, the two streams of course comprising different cases. The cases all came from sizable tort and contract categories that clearly involved a choice between jury and judge trial. (206) In two of the most controversial areas of modern tort law, product liability and medical malpractice, the win rates substantially differ from other categories' win rates and in a surprising way: plaintiffs in these two areas prevail after trial at a much higher rate before judges (48%) than they do before juries (28%). (207) Furthermore, in medical malpractice but not in product liability, the mean recovery in judge trials is higher than the mean recovery injury trials. (208) These empirical results proved resistant to all simple explanations, such as differences in the size of award explaining differences in win rates. (209)
So we considered the results in light of the parties' ability to select which cases reach jury or judge trial. (210) Lawyers entertain long-standing perceptions of the jury as biased and incompetent, relative to the judge. (211) There is, however, no actual evidence that juries are relatively biased or incompetent. (212) Those perceptions nevertheless could have the consequence of a selection of cases reaching jury trial that differs from the case selection reaching judge trial. In particular, the theorizing ran that in certain categories of cases such as product liability and medical malpractice, lawyers view the jury as relatively favorable to plaintiffs. They then settle cases in a way that leaves for trial by jury or judge a residue of what they consider close cases, with juries accordingly seeing, on average, weaker cases. The perceptions then turn out to be misperceptions, as jury and judge turn out to perform similarly. (213) Thus, the jury produces fewer winners than expected, while the judge produces more winners.
Our theorizing and analysis led, after a lengthy article based on a wealth of data covering all sorts of cases, to three conclusions. First, the most plausible explanation of the data lies in small differences between judges' and juries' treatment of cases and, much more substantially, in the parties' varying the case selection that reaches judge and jury. Second, litigants' stereotypical views about juries may lead them to act unwisely in choosing between judge trials and jury trials. Third, the surprising win rates in product liability and medical malpractice cases may stem from the especially strong misperceptions litigants hold about judge and jury behavior in these cases. (214) More simply put, certain groups of plaintiffs do far better before judges, but the reason likely lies in prevailing misperceptions about juries, rather than in differences between judges and juries. Judges and juries are in fact not so different. (215)
Despite the research that rebuts stereotypes about juries, every day lawyers and policymakers act on the basis of those old stereotypes. In general, longstanding misperceptions about the legal system are not uncommon. (216) But why are such misperceptions about the legal system so resilient, rather than eventually undergoing correction as lawyers repeatedly observe the consequences of their misperceptions? On the particular subject of jury/judge performance, elitist perceptions of a biased and incompetent jury system seem to conform to the natural order of things and can even be comforting. Persuasive and accessible empirical evidence to the contrary has been slow in accumulating. Finally, many lawyers simply prefer to rely on intuition informed by personal experience and anecdote. (217) All in all, lawyers' misperceptions of jury/judge differences have understandably prevailed for a long time.
If one were to accept the new empirical evidence, however, practical lessons would emerge. Returning to the same example of product liability and medical malpractice, one could conclude that the jury is less of an advantage for plaintiffs, and the judge less of a disadvantage, than lawyers think. That realization should affect the terms of settlement. Moreover, if only one side comes to that realization, that side could manipulate the jury/judge choice to its bargaining advantage.
A. Win Rates
As already observed, a popular form of empirical study involves examining the parties' success in obtaining judgment after litigation. (218) Not only are such judgment data readily available, but also they appear to be full of meaning. An analyst might see the win-rate data as revealing some underlying factor affecting outcome generally, such as some substantive or procedural rule or some nonlegal factor favoring one side or the other in the set of all disputes. Yet this interpretive step can easily lead the analyst astray. Because win-rate data convey the system's output while hiding the variable composition of its input, win-rate data inherently entail a near-fatal ambiguity that theorists call the selection effect. (219)
More specifically, disputes and cases that clearly favor either the plaintiff or the defendant tend to settle readily, because both sides can save costs by settling in light of their knowledge of the applicable law and all other aspects of the case. Difficult cases falling close to the applicable decisional criterion tend not to settle, because the parties are more likely to disagree substantially in their predicted outcomes. These unsettled close cases fall more or less equally on either side of the criterion, regardless of the position of that criterion and regardless of the underlying distribution of disputes. Thus, even if, say, a legal criterion such as strict liability highly favors plaintiffs, one might not observe a plaintiff win rate well above 50%. Instead, case selection will leave for adjudication a residue of unsettled close cases, which consequently exhibit some nonextreme equilibrium win rate.
In other words, the parties' selection of which cases to push into and through litigation produces a biased sample from the mass of underlying disputes. This case-selection effect means that the win rate reveals something about the set of adjudged cases, a universe dominated by close cases--but reveals not much about the underlying, variegated mass of disputes and cases, and indeed little about the litigation process' treatment thereof. According to case-selection effect theory, any distinction between two streams of cases that the parties evaluate without systematic inaccuracy, say, product liability and medical malpractice, should lead to no difference in adjudicated win rates. Indeed, under simplifying assumptions, and as a limiting implication, the theory suggests a trial win rate of 50% for both streams.
Actually, however, the fully developed theory does not predict any universal win rate, or even that any two streams' rates will be the same. Reality is too complicated to produce a 50% win rate. There are three main types of factors that might lead to win rates different from 50%: differential stakes, parties' misperceptions, and influences such as case strength that survive because of imperfect case selection. That last set of influences indeed does mean that win rates may retain residual meaning, which the settlement process has not obliterated. Careful research and theorizing can often succeed in untangling the neutralizing effect of settlement. The challenge is to tease out the residual meaning in win-rate data.
For example, Professor Eisenberg, Dean Schwab, and I have done a fair amount of study on employment discrimination litigation, or so-called jobs cases. (220) This is an important category, emerging in the 1970s, and then exploding in the 1990s so that it peaked at almost 10% of the federal civil docket by the end of the decade (221)--and accounting for an even bigger percentage of federal civil trials. (222) We showed that in federal court the plaintiff win rate for jobs cases (15%) was lower than that for nonjobs cases (51%). (223) The win rate in jobs was consistently low, not only for race and sex discrimination but also for the various other subtypes such as disability and age discrimination. Over the period of fiscal years 1979-2006, plaintiffs won 28% of jobs trials, but 45% of nonjobs trials. Plaintiffs in jobs cases won 4% of judgments that came by pretrial motion, but 21% in nonjobs cases. (224)
What to make of these results? It could be that overly litigious civil rights plaintiffs start with weak cases and then present them less effectively than the defendants. But there is no evidence for this theory; in fact, these plaintiffs and their lawyers should have the same economic disincentives against pressing weak cases, or indeed suing at all, as do other claimants. Instead, our painstaking review of employment discrimination cases throughout the litigation process, including settlement and appeal, suggested the existence of a legal system biased against employment discrimination plaintiffs, making theirs a tough row to hoe.
B. Foreigner Effect
Everyone knows that foreigners fare badly in U.S. courts, right? Well, no, that is not true, according to our research. (225) In fact, foreign plaintiffs suing domestic defendants have enjoyed a higher win rate (75%) than domestic plaintiffs suing domestic defendants (59%), in federal diversity actions over the last two decades. (226) Likewise, domestic plaintiffs suing foreign defendants fare worse (50%) than domestic plaintiffs suing domestic defendants (59%). (227) This foreigner effect was not specific to certain courts or certain case categories, and did not depend on the procedural route taken to judgment, but instead prevailed across the board. (228)
Why? Our analysis rejected the implausible notion that U.S. courts have a pro-foreigner bias, as well as the more plausible explanation that foreign parties litigate better than domestic parties. (229) Instead, it appears that foreigners' fear of U.S. courts lead them to pursue only an unusually strong set of cases. (230) That is, foreigners are averse to litigating here and hence are more selective in choosing strong cases to pursue to judgment. When the foreigners do not encounter the expected level of bias, they end up winning more of their cases. So, it is case selection at work.
Accordingly, although we cannot prove that antiforeign bias is nonexistent in U.S. courts, we can say that the available data do not support the view that U.S. courts harbor xenophobic bias. The data instead suggest that foreigners would be wise to lessen their general aversion to litigation here.
As I said above, however, one must ensure that one sees most pieces of the puzzle before drawing broad conclusions. It is important to draw data from a variety of types of fora and locales and a variety of case categories, as well as from a range of years. Here, time trends are a big part of the story, just as they are elsewhere. (231) Indeed, a recent and rapid change in time trends shapes the story. (232) When we first wrote on this topic, we had data only through fiscal year 1994, so that the foreigners' edge appeared sizable. (233) We have since taken the longer view, and have thereby detected that the foreigners' aversion waxes and wanes with the times and thus creates the pattern seen in Figure 5. (234) As it shows, in the past foreigners substantially outperformed their domestic counterparts in obtaining favorable judgments, but more recently the foreigners' "advantage" has all but disappeared. Still, the point remains that case selection drives the outcomes for foreigners, so that the skewed sample of foreigner cases that reach judgment does not reveal any supposed partiality of the legal system and hence does not require any of the proffered structural or cultural explanations of xenophobia. (235)
A. Affirmance Effect
While win rates in the trial court can be high or low across case categories, affirmance rates in the appellate court are elevated for all kinds of cases. Figure 6, now extended in years and focused on my running example of employment discrimination, shows these patterns nicely. (236) It uses the federal court data on judgments for plaintiff or defendant and decisions for appellant or appellee to show rates for jobs cases and all other civil cases. The lower set of two lines comprises the jobs and nonjobs win rates in district court, each line limited to results at trial so that the win rate can be most meaningful. Note that the nonjobs win rate over time is fairly steady or maybe descending, while jobs has a much lower win rate but one that has been gently increasing over the period. The cluster of two lines near the top comprises affirmance rates for jobs cases and all other civil cases, whether tried or not. The affirmance rate for jobs is slightly higher than nonjobs of late. Jobs cases are usually unsuccessful below, and the district court's result usually meets affirmance on appeal. As already mentioned, jobs plaintiffs have a tough row to hoe.
[FIGURE 5 OMITTED]
[FIGURE 6 OMITTED]
The most striking feature about appeals is the high rate of affirmance. (237) Our work in a number of articles shows the affirmance rate for federal civil appeals to be about 80%. (238) At first glance, this affirmance effect may seem unsurprising. One may expect a high affirmance rate because of frequent appellate deference to the district court's result. One may even expect a high affirmance rate when review is de novo, because of the tendency of experts to agree at about a 75% rate. (239) Combining the two expectations based on appellate deference and expert agreement would push one's expected affirmance rate even higher toward 80%. Appellate judges should and do lean toward affirmance as the usual course. (240)
However, if the high affirmance rate is owing to those deference and expertise factors, why do the parties not take them into account and settle all but the close appeals, thereby whittling down that high affirmance rate? The usual brand of case-selection theory says that appeals should act like trials. (241) Indeed, under simplifying assumptions, and as a limiting implication, case-selection theorizing would predict a 50% affirmance rate. That is clearly wrong, as the data prove.
Thus, the persistently elevated affirmance rate suggests that settlement is not very effective at the appellate phase in weeding out clear cases. After all, if every judgment underwent appeal, one would expect about an 80% affirmance rate because of reviewer's deference and because of experts' agreement. In fact, only a fraction of judgments undergo appeal--less than a fifth of decisive judgments, with less than half of these proceeding all the way to a decisive appellate outcome (242)--and yet one nevertheless still sees an 80% affirmance rate. It seems as if the parties have chosen to appeal, by whatever selection method they employ, a set of cases that is not random but still functions, at least with regard to overall affirmance, as if it were a random sampling. In sum, case selection might have a very limited effect in systematically filtering the cases for adjudication on appeal. (243)
Why would that be? Judgment below leaves the winner feeling vindicated, and the aggrieved loser wanting justice at long last. Something telling emerges in the countless scenes on the evening news in which losers immediately proclaim on the courthouse steps their intention to appeal. After slogging through the trial court, the losing party must see the small cost and effort in appealing as insignificant when compared to the big return of reversal. Nearly one-fifth of losing parties decide that they might as well stagger to the finish line, pretty much regardless of the chances on appeal. (244) Perhaps, then, the failure to filter out clear-cut appeals is owing to appeals' not being very costly in relative terms. Simply put, an 80% affirmance rate suggests that the law should consider reform aimed at the efficiency of forcing the would-be appellant to pause. A possible reform proposal would involve shifting attorney's fees on appeal to a losing appellant, which would seem a fair condition of access to a second court for a party already found to be in the wrong. (245)
Beyond such an indirect lesson, it may be that gross appeal rates and affirmance rates do not have much to tell policymakers, for example, about the quality of first-instance justice. Appeal rates may turn mainly on the cost of appeal. Affirmance rates may mainly reflect any selection effects. One must dive much more deeply into the data to draw meaningful lessons, as I shall next illustrate.
B. Anti-Plaintiff Effect
Our research also revealed a surprising plaintiff/defendant difference in the federal courts of appeals. (246) After matching individual district court cases with their appeals, if any, we could show that defendants succeed more than plaintiffs on appeal. For example, defendants appealing their losses after completed trial obtain reversals at a 33% rate, while losing plaintiffs succeed in only 12% of their appeals from completed trial. (247) Therefore, defendants emerge from the appellate court in a much better position than when they left the trial court. Again, the effect is especially pronounced in jobs cases.
Why would this plaintiff/defendant difference exist? This question takes me from fact into speculation. I think that the plaintiffs' lower reversal rate stems from real but hitherto unappreciated differences between appellate and trial courts. Both our descriptive analyses of the results and our more formal regression models dispelled explanations based on selection of cases, and instead supported an explanation based on appellate judges' attitudes toward trial-level adjudicators. The appellate judges may be acting on their perceptions of the trial courts as being pro-plaintiff. The appellate court consequently would be more favorably disposed to the defendant than are the trial judge and the jury.
This appellate favoritism would be appropriate if the trial courts were in fact biased in favor of the plaintiff. But, as recounted in our articles, empirical evidence tends to refute trial court bias on the plaintiff/defendant axis, and so any such appellate judges' perceptions appear increasingly to be misperceptions. Or unconscious biases may be at work. Perhaps, for example, appellate judges' greater distance from the trial process creates an environment in which it is easier to discount harms to the plaintiff. In either event, the data on appellate leaning in favor of the defendant become a cause for concern. In short, I think we have unearthed an anti-plaintiff effect in federal appellate courts that is troublesome.
Nevertheless, it merits stressing that we have never claimed that the attitudinal explanation of the anti-plaintiff effect is irrefutable. What is the best counterargument? It would be that plaintiffs start with weak cases, and then present them less effectively than defendants. We are looking at output data, after all; by making appropriate assumptions about the input, one can explain any particular pattern in the output data. Thus, weak cases, weakly pushed by overly litigious plaintiffs who also appeal too readily, will mathematically result in a higher reversal rate for defendants, and so could produce the look of an anti-plaintiff effect in reversal rates, even before perfectly neutral courts. (248)
My response is that no empirical basis exists for inferring such a difference between the strength of plaintiffs' and defendants' cases. Moreover, even if plaintiffs are flooding the district courts with weak cases, those stalwart few who make it through pretrial, through settlement, and then through to trial victory should at the least have relatively strong cases; these are cases that survived the prefiling and pretrial screening, and so are nonfrivolous cases with a genuine factual issue; the settlement-litigation process should have weeded out the lopsided cases, leaving a pool of claims comprising mainly close cases. Yet these tried cases exhibit a more extreme anti-plaintiff effect on appeal than do pretrial judgments. This result is strongly inconsistent with any "weak cases produce divergent reversal rates" argument. Finally, our prior research found the anti-plaintiff effect on appeal prevails even between corporate parties. Thus, rather than yielding to the intuitive attraction of the view that plaintiffs are overly litigious, I tentatively conclude that appellate judges are acting as if it is they who accept that view. Their resulting attitude then produces at least some of the observed anti-plaintiff effect. (249)
The six stories of Litigation Realities now stand renewed. But their conclusion stays the same: data are good. Or maybe better than ever!
(1) Kevin M. Clermont & Theodore Eisenberg, Litigation Realities, 88 CORNELL L. REV. 119, 120 (2002).
(2) See DUC DE CASTRIES, L'AGONIE DE LA ROYAUTE 193 (1959). But cf. PAUL GIRAULT DE COURSAC & PIERRETTE GIRAULT DE COURSAC, ENTRETIENS SUR LOUIS 16, at 144-45 (1990) (explaining that, in actuality, this much-maligned monarch was noting only that there had been no hunt that day).
(3) See THE YALE BOOR OF QUOTATIONS 620 (Fred R. Shapiro ed., 2006) (translating "mieulx vault aucun bien que neant," from the early fifteenth century). Frederick Mosteller, as quoted in Time magazine, observed: "It is easy to lie with statistics, but easier to lie without them." Melissa August et al., Milestones, TIME, Aug. 7, 2006, at 23, 23. Nevertheless, a major lesson from the present Article is that before drawing conclusions, the researcher should seek out as many pieces of the puzzle as possible--using data from as many years as possible, and from as many settings as possible (different case categories, civil and noncivil proceedings, state and federal forums, U.S. and foreign settings). See, e.g., Dan Childs et al., Studies Gone Wild: Death by Shower Curtain?, ABC NEWS, June 12, 2008, http://abcnews.go.com/Health/story?id=50571 41&page=1 ("The group tested a total of five shower curtains, of which only one shower curtain--not one brand; one curtain--was subjected to complete testing for chemicals in its composition, as well as those it released into the air--a phenomenon known as 'off-gassing.'").
(4) See also Kevin M. Clermont & Theodore Eisenberg, Foreigners' Fate in America's Courts: Empirical Legal Research, 1 ACADEMIA SINICA L.J. 237, 239-50 (2007), available at http://empirical.law.cornell.edu/articles/sinica.pdf (discussing the impact of empirical research on the law).
(5) See, e.g., Kevin M. Clermont & Theodore Eisenberg, Exorcising the Evil of Forum-Shopping, 80 CORNELL L. REV. 1507, 1508 (1995) (noting that forum selection "has a major impact on outcome"); see also Mary Garvey Algero, In Defense of Forum Shopping: A Realistic Look at Selecting a Venue, 78 NEB. L. REV. 79 (1999) (stressing the benefits of forum shopping); Debra Lyn Bassett, The Forum Game, 84 N.C. L. REV. 333 (2006) (defending forum shopping as a legitimate tactic when the law authorizes bringing a lawsuit in more than one forum); Friedrich K. Juenger, Forum Shopping, Domestic and International, 63 TUL. L. REV. 553, 570 (1989) ("[N]ot all forum shopping merits condemnation."); Richard Maloy, Forum Shopping? What's Wrong with That?, 24 QUINNIPIAC L. REV. 25, 33-50 (2005) (distinguishing permissible from impermissible forum shopping); James E. Pfander, Forum Shopping and the Infrastructure of Federalism, 17 TEMP. POL. & CIV. RTS. L. REV. 355, 355 (2008) (explaining that forum shopping "ensure[s] that firms operating through the nation must comply with relatively more pro-consumer policies at the state level") ; J. Skelly Wright, The Federal Courts and the Nature and Quality of State Law, 13 WAYNE L. REV. 317, 333 (1967) (criticizing forum shopping as "a national legal pastime").
(6) See infra Parts III-IV.
(7) See, e.g., 28 U.S.C. [section] 1391 (2006) (governing venue within the federal system).
(8) See, e.g., id. [section] 1441(a) (authorizing removal generally).
(9) See, e.g., id. [section] 1404(a) (authorizing transfer between districts within the federal system).
(10) Clermont & Eisenberg, supra note 5, at 1509 n.3, reported that the numbers of transfer motions and trials were about the same through fiscal year 1991, but since then the number of trials has dropped precipitously. See infra note 184.
(11) ROBERT C. CASAD, JURISDICTION AND FORUM SELECTION (2d ed. 2007 & Update 2008).
(12) See generally KEVIN M. CLERMONT, PRINCIPLES OF CIVIL PROCEDURE 200-05 (2d ed. 2009) (giving background on law of removal).
(13) The most important exception to the removability of cases appears in 28 U.S.C. [section] 1441(b) (2006), whereby defendants cannot remove a diversity case if any served defendant is a citizen of the forum state.
(14) Id. [section] 1446(a)-(b).
(15) Id. [section] 1446(d).
(18) Id. [section] 1447(d).
(19) See Kevin M. Clermont & Theodore Eisenberg, Do Case Outcomes Really Reveal Anything About the Legal System? Win Rates and Removal Jurisdiction, 83 CORNELL L. REV. 581, 599 (1998).
(20) See RICHARD H. FIELD ET AL., MATERIALS FOR A BASIC COURSE IN CIVIL PROCEDURE 221-22 (9th ed. 2007).
(21) See DANIEL JOHN MEADOR, AMERICAN COURTS 31--33 (2d ed. 2000); Brian J. Ostrom et al., Examining Trial Trends in State Courts: 1976-2002, 1 J. EMPIRICAL LEGAL STUD. 755, 757 (2004).
(22) These data were gathered by the Administrative Office of the United States Courts (AO), assembled by the Federal Judicial Center, and disseminated by the Inter-university Consortium for Political and Social Research. See Theodore Eisenberg & Kevin M. Clermont, Courts in Cyberspace, 46 J. LEGAL EDUC. 94 (1996). These data convey details of all cases terminated in the federal courts since fiscal year 1970. When any civil case terminates in a federal district court or court of appeals, the court clerk transmits to the An a form containing information about the case. The forms include, inter alia, data regarding the names of the parties, the subject matter category (the form distinguishes among some ninety categories, including specific branches of contract, tort, and other areas of law) and the jurisdictional basis of the case, the case's origin in the district as original, removed, or transferred, the amount demanded, the dates of filing and termination in the district court or the court of appeals, the procedural stage of the case at termination, the procedural method of disposition, and, if the court entered judgment or reached decision, the prevailing party and the relief granted. Thus, the computerized database, compiled from these forms, contains all of the millions of federal civil cases over many years from the whole country. Clermont & Eisenberg, supra note 1, at 127-29, more fully describes this database and its strengths and weaknesses.
(23) For Figure 1, I eliminated asbestos cases from the Northern District of Ohio in calendar year 1990 to avoid the distortion of their unusually high number. I did the same for multidistrict product liability terminations from the Northern District of Alabama in 1998 and 1999; from the District of Minnesota and the Northern District of Ohio in 2004; and from the District of Minnesota, the Northern District of Ohio, and the Eastern District of Pennsylvania in 2006 (while also eliminating over sixteen thousand reopened asbestos and diet-drug cases that were dismissed in 2006 by that last district).
(24) See 28 U.S.C. [section] 1332(a) (2006). In rough terms, this statute extends jurisdiction to cases for more than $75,000 between citizens of different U.S. states or between foreigners and state citizens, but it requires the diversity to be "complete," that is, no two opposing parties can be citizens of the same state. See Strawbridge v. Curtiss, 7 U.S. (3 Cranch) 267, 267 (1806). The Constitution would permit "minimal" diversity, that is, the only requirement, absent a statutory restriction, is that a state citizen and someone of different citizenship must be on opposite sides. See State Farm Fire & Cas. Co. v. Tashire, 386 U.S. 523, 530-31 (1967) (upholding the interpleader statute on such basis).
(25) As to the remand rate, it is reliable only from fiscal year 1979, when the AO's coding practices made the necessary changes.
(26) Clermont & Eisenberg, supra note 1, at 123 fig.1.
(27) See id. at 122 & n.16.
(28) Theodore Eisenberg & Trevor W. Morrison, Overlooked in the Tort Reform Debate: The Growth of Erroneous Removal, 2 J. EMPIRICAL LEGAL STUD. 551, 565 fig.2, 566 fig.4 (2005). Their article triggered an interesting debate on TortsProf Blog, Abusive Removals (updated), http://lawprofessors.typepad.com/tortsprof/2006/10/abusive_removal.html (Oct. 10, 2006). Ted Frank, Director of the American Enterprise Institute Legal Center, there says: "I've been very disappointed in empirical legal work. [It] usually consist[s] of a study that performs technically accurate counting of statistics, and then wild jumps to conclusions that coincidentally correspond to the authors' biases without acknowledgement of the limits of the data." Id. (Oct. 10, 2006, 17:10 EST). But the only point such critics made is that "erroneous removal" means no more than removals that the system has determined were in error and so require remand, an obvious point on which Eisenberg and Morrison had been perfectly explicit.
(29) Eisenberg & Morrison, supra note 28, at 564.
(31) See id. at 561, 576. But cf. Martin v. Franklin Capital Corp., 546 U.S. 132, 141 (2005) (reading narrowly the authority in 28 U.S.C. [section] 1447(c) to award "just costs" for improper removal relatively).
(32) The upswing in the removal rate could be attributable in part to the statutory amendments in 1988 that changed removal from a relatively burdensome petition-and-bond process to a simple notice-of-removal scheme that facilitates defendants' forum shopping. See Christopher Terranova, Erroneous Removal as a Tool for Silent Tort Reform: An Empirical Analysis of Fee Awards and Fraudulent Joinder, 44 WILLAMETTE L. REV. 799, 805-09 (2008) (observing the recent dive in remand rate, but not knowing how to explain it even after extensive analysis; also making the point that the observed pattern does not result from using termination data, even though the quickly terminated remand cases do not perfectly align year-by-year with the more slowly terminated nonremand cases). However, the transfer rate for federal cases has shown a similar increase over recent decades, thus implying a more generic cause for increased forum shopping. See Clermont & Eisenberg, supra note 5, at 1526-29.
(33) The new trend seems pervasive. The downward trend prevails, most often in a statistically significant way, in every circuit since 2001, except the First Circuit where it did not appear until 2003. It prevails, most often in a statistically significant way, in each case grouping (contract, real property, personal injury torts, product liability, personal property torts, and statutory actions).
(34) For present purposes, I narrow the AO definition of judgments to include only those cases where the data indicate a win by plaintiff or defendant, not by both or by an unknown party. Note, however, that these judgments comprise much more than trial outcomes: for AO purposes, judgments might be the result of adjudication, consent, or default, although they normally do not include voluntary dismissals or dismissals for lack of prosecution.
(35) See Clermont & Eisenberg, supra note 19, 592-607.
(36) See Clermont & Eisenberg, supra note 5, at 1511-30; see also Kevin M. Clermont & Theodore Eisenberg, Simplifying the Choice of Forum: A Reply, 75 WASH. U. L.Q. 1551 (1997) (using an empirical study to examine the effects of transfer of venue on case outcomes).
(37) Clermont & Eisenberg, supra note 19, at 593.
(38) Id. at 606.
(39) Id. at 606-07.
(40) Clermont & Eisenberg, supra note 5, at 1512.
(41) Id. at 1514.
(42) Cf. Clermont & Eisenberg, supra note 19, at 603 n.67 (showing the transfer effect, which reduced 50% odds to 38% for diversity cases).
(43) Pub. L. No. 109-2, 119 Stat. 4 (2005) (codified in scattered sections of 28 U.S.C.). See generally GEORGENE M. VAIRO, CLASS ACTION FAIRNESS ACT OF 2005 (2005) (giving background on CAFA). For a different but even more astounding example of policymakers' interest in the effect of forum on outcome, see Elizabeth G. Thornburg, Judicial Hellholes, Lawsuit Climates, and Bad Social Science: Lessons from West Virginia, 110 W. VA. L. REV. 1097, 1134-37 (2008) (calling for more empirical research to rebut tort reform propaganda).
(44) See 7A CHARLES ALAN WRIGHT ET AL., FEDERAL PRACTICE AND PROCEDURE [section] 1756.2 (3d ed. 2005 & Supp. 2008) (describing Congress' goal of preventing litigation abuse through CAFA).
(45) Joel Roberts, Lawsuit Limits Become Law, CBS NEWS, Feb. 18, 2005, http://www.cbsnews.com/stories/2005/02/11/politics/main673159.shtml (quoting President George W. Bush); see also Edward A. Purcell, Jr., The Class Action Fairness Act in Perspective: The Old and the New in Federal Jurisdictional Reform, 156 U. PA. L. REV. 1823, 1861-63, 1867 (2008) (stressing partisan support for CAFA).
(46) 28 U.S.C. [section] 1332(d) (2006); see also supra note 24 (describing the difference between complete and minimal diversity).
(47) 28 U.S.C. [section] 1453 (2006); see also supra notes 13, 18 and accompanying text (discussing the non-CAFA requirement of having all defendants agree on removal and the nonappealability of remand orders).
(48) Kevin M. Clermont & Theodore Eisenberg, CAFA Judicata: A Tale of Waste and Politics, 156 U. PA. L. REV. 1553 (2008). On the dangers of limiting study to such "published" opinions, see id. at 1559-60, 1562-63.
(49) Id. at 1560.
(50) Id. at 1565.
(51) Id. at 1579-84.
(52) As to this surprising result, see id. at 1584-91.
(53) See CLERMONT, supra note 12, at 37-54.
(54) See Edward Cavanagh, Pleading Rules in Antitrust Cases: A Return to Fact Pleading?, 21 REV. LITIG. 1, 1-2 (2002).
(55) See Mark D. Robins, The Resurgence and Limits of the Demurrer, 27 SUFFOLK U. L. REV. 637, 642 n.24 (1993).
(56) See Douglas A. Blaze, Presumed Frivolous: Application of Stringent Pleading Requirements in Civil Rights Litigation, 31 WM. & MARY L. REV. 935, 942 (1990).
(57) See, e.g., Conley v. Gibson, 355 U.S. 41, 48 (1957) ("[W]e have no doubt that petitioners' complaint adequately set forth a claim and gave the respondents fair notice of its basis."), abrogated by Bell Atl. Corp. v. Twombly, 550 U.S. 544 (2007).
(58) Disagreement prevails even within families, as my wife, Emily Sherwin, can attest. Compare Emily Sherwin, The Story of Conley : Precedent by Accident, in CIVIL PROCEDURE STORIES 295, 317-20 (Kevin M. Clermont ed., 2d ed. 2008) (criticizing lenient pleading rules), and Emily Sherwin, The Jurisprudence of Pleading: Rights, Rules, and Conley v. Gibson, 52 How. L.J. 73 (2008) (same), with CLERMONT, supra note 12, at 37-41 (favoring a move toward even purer notice pleading).
(59) See Christopher M. Fairman, Heightened Pleading, 81 TEX. L. REV. 551, 574-82 (2002); Christopher M. Fairman, The Myth of Notice Pleading, 45 ARIZ. L. REV. 987, 1002-09 (2003) ; Richard L. Marcus, The Puzzling Persistence of Pleading Practice, 76 TEX. L. REV. 1749, 1759-61, 1774-75 (1998); Richard L. Marcus, The Revival of Fact Pleading Under the Federal Rules of Civil Procedure, 86 COLUM. L. REV. 433, 444-51 (1986).
(60) See Swierkiewicz v. Sorema N.A., 534 U.S. 506, 513-15 (2002) (involving a Title VII employment discrimination claim); Leatherman v. Tarrant County Narcotics Intelligence & Coordination Unit, 507 U.S. 163, 168 (1993) (involving a civil rights claim against a municipality).
(61) See Michael Chiorazzi et al., Empirical Studies in Civil Ptvcedure: A Selected Annotated Bibliography, Law & CONTEMP. PROBS., Summer 1988, at 87, 120 ("[T]here are no studies on pleading per se, perhaps because of the advent of modern notice pleading."); Civil Rules Advisory Comm., Judicial Conf. of the U.S., Minutes Oct. 27-28, 2005, at 29-35, available at http://www.uscourts.gov/rules/Minutes/CV11-2005-rain.pdf; Civil Rules Advisory Comm., Judicial Conf. of the U.S., Minutes May 22-23, 2006, at 37-38, available at http://www.uscourts.gov/rules/Minutes/CV05-2006-min.pdf.
(62) 550 U.S. 544 (2007); see also Ashcroft v. Iqbal, 129 S.Ct. 1937 (2009) (elaborating on the meaning of Bell Atlantic).
(63) Bell Atl. Corp., 550 U.S. at 549-51.
(64) Id. at 553.
(65) Id. at 565.
(66) Id. at 565 & n.10.
(67) Id. at 570.
(68) Id. at 566-68.
(69) Id. at 565 n.10.
(70) Id. at 570.
(71) Id. at 573 (Stevens, J., dissenting).
(73) By discombobulating a basic area of law, the case managed to generate an absolutely extraordinary 5000 case citations in its first eleven months, see Amy J. Wildermuth, What Twombly and Mead Have in Common, 102 Nw. U. L. REV. COLLOQUY 276, 276 (2008), http://www.law.northwestern.edu/lawreview/colloquy/2008/12/ LRColl2008nl2Wildermuth.pdf, and 7000 case citations after thirteen months. See Gregory P. Joseph, Federal Litigation--Where Did It Go Off Track ?, LITIGATION, Summer 2008, at 5, 62. For a judge's consequent lament, see Colleen McMahon, The Law of Unintended Consequences: Shockwaves in the Lower Courts After Bell Atlantic Corp. v. Twombly, 41 SUFFOLK U. L. REV. 851 (2008). On the implications for state courts, see Z.W. Julius Chen, Note, Following the Leader." Twombly, Pleading Standards, and Procedural Uniformity, 108 COLUM. L. REV. 1431 (2008).
(74) Cf. Robert G. Bone, Twombly, Pleading Rules, and the Regulation of Court Access, 94 IOWA L. REV. 873, 935-36 (2009) (arguing that the Court's thin plausibility standard could be justifiable, if adopted by the proper statute or rule process); Keith Bradley, Pleading Standards Should Not Change After Bell Atlantic v. Twombly, 102 Nw. U. L. REV. COLLOQUY 117, 122 (2007), http://www.law.northwestern.edu/lawreview/ colloquy/2007/31/lrcoll2007n31Bradley.pdf ("'Plausibility' is an element of a certain kind of antitrust conspiracy claim, not a standard for pleadings in general."); Stephen B. Burbank, Pleading and the Dilemmas of "General Rules, "2009 WIS. L. REV. 535 (arguing that substantive-specific federal common law could modify the transsubstantive Federal Rules on pleading); Charles B. Campbell, A "Plausible" Showing After Bell Atlantic Corp. v. Twombly, 9 NEV. L.J. 1, 2 (2008) (translating "plausible" to mean that a complaint must "'contain either direct or inferential allegations respecting all the material elements necessary to sustain a recovery under some viable legal theory'" (quoting In re Plywood Antitrust Litig., 655 F.2d 627, 641 (5th Cir. 1981))); Edward D. Cavanagh, Twombly, The Federal Rules of Civil Procedure and the Courts, 82 ST. JOHN'S L. REV. 877, 879 (2008) (concluding that "(1) the Court's assertion that judges cannot effectively control litigation costs because the parties--not the courts--control claims and defenses as well as the nature and amount of discovery in any given case is contrary to fact; and (2) certain classes of cases may well warrant particularized pleading but that decision should be made by the rulemakers through amendments to the Federal Rules of Civil Procedure and not by judges on an ad hoc basis"); Scott Dodson, Pleading Standards After Bell Atlantic Corp. v. Twombly, 93 VA. L. REV. IN BRIEF 135, 140, 142 (2007), http://www.virginialawreview.org/inbrief/2007/07/09/ dodson.pdf (offering that "the best reading of Bell Atlantic is that Rule 8 now requires notice-plus pleading for all cases," but "it' will spawn years of increased litigation"); Richard A. Epstein, Bell Atlantic v. Twombly: How Motions to Dismiss Become (Disguised) Summary Judgments, 25 WASH. U. J.L. & POL'Y 61, 99 (2007) (approving the result of Bell Atlantic, and ultimately supporting "dismissal at the close of pleadings in any case where the defendant has negated all inferences of culpability by using the same kinds of public evidence that the plaintiff has used to establish a factual underpinning to the underlying complaint"); Lonny S. Hoffman, Burn Up the Chaff with Unquenchable Fire: What Two Doctrinal Intersections Can Teach Us About Judicial Power over Pleadings, 88 B.U. L. REV. 1217, 1255-70 (2008) (criticizing Bell Atlantic from a broader perspective, but ultimately approving a limited screening of conclusory pleadings unless the pleader can show a special need for discovery); Max Huffman, The Necessity of Pleading Elements in Private Antitrust Conspiracy Claims, 10 U. PA. J. Bus. & EMP. L. 627, 640-45 (2008) (approving Bell Atlantic's notice-plus pleading as a transsubstantive rule implicit in Federal Rule 12(b) (6)'s requirement that the pleader show entitlement to relief); Keith N. Hylton, When Should a Case Be Dismissed? The Economics of Pleading and Summary Judgment Standards, 16 SUP. CT. ECON. REV. 39, 42-54 (2008) (using economic analysis to conclude that pleading standards should vary with the case's evidentiary demands and the social costs of litigation); Allan Ides, Bell Atlantic and the Principle of Substantive Sufficiency Under Federal Rule of Civil Procedure 8(a)(2): Toward a Structured Approach to Federal Pleading Practice, 243 F.R.D. 604, 631-36, 639 (2007) (approving, with reservations, Bell Atlantic's application of pleading's substantive-sufficiency test to this antitrust case); Randal C. Picker, Twombly, Leegin and the Reshaping of Antitrust, 2007 SUP. CT. REV. 161, 176-77 (suggesting that the Supreme Court should have proceeded instead by limiting discovery);J. Douglas Richards, Three Limitations of Twombly : Antitrust Conspiracy Inferences in a Context of Historical Monopoly, 82 ST. JOHN'S L. REV. 849 (2008) (taking a very narrow view of Bell Atlantic); Douglas G. Smith, The Twombly Revolution?, 36 PEPP. L. REV. (forthcoming 2009) (manuscript at 17-24), available at http://ssrn.com/abstract=1147862 (downplaying and defending Bell Atlantic); A. Benjamin Spencer, Plausibility Pleading, 49 B.C. L. REV. 431, 460-89 (2008) (strongly criticizing the new regime of plausibility pleading); Paul J. Stancil, Balancing the Pleading Equation (Ill. Law & Econ. Research Paper Series, Research Paper No. LE08-018, 2008), available at http://ssrn.com/abstract=1266323 (calling for a return to fact pleading, but for only certain classes of cases) ; Richard M. Steuer, Plausible Pleading: Bell Atlantic Corp. v. Twombly, 82 ST. JOHN'S L. REV. 861, 875 (2008) (saying that Bell Atlantic "increases the burden by replacing the 'no set of facts test' with a 'show me the facts' test"); Suja A. Thomas, Why the Motion to Dismiss Is Now Unconstitutional, 92 MINN. L. REV. 1851 (2008) (contending that plausibility testing violates the Seventh Amendment); Ettie Ward, The After-Shocks of Twombly: Will We "Notice" Pleading Changes?, 82 ST. JOHN'S L. REV. 893, 905-18 (2008) (stressing that much still remains unclear as to the meaning of Bell Atlantic); Ryan C. Gist, Note, Transactional Pleading: A Proportional Approach to Rule Eight in the Wake of Bell Atlantic v. Twombly, 2008 WIS. L. REV. 1013, 1035-46 (proposing a rule amendment that would require plaintiffs to plead a varying degree of factual particularity proportional to the dangers of both over-restriction and abuse in any given situation); Elizabeth M. Schneider, The Changing Shape of Federal Civil Pretrial Practice: The Disparate Impact on Civil Rights and Employment Discrimination Cases 7-8 (2009) (unpublished manuscript) (lamenting Bell Atlantic's substantive impact). Professor Dodson's above-cited article is a good place to start, and Professor Spencer's article most helped me to understand the narrow question of what the decision actually meant. See also Scott Dodson, Comparative Convergence in Pleading Standards, 158 U. PA. L. REV. (forthcoming 2009), available at http://ssrn.com/abstract=1351994 (comparing American pleading to pleading in the rest of the world); A. Benjamin Spencer, Understanding Pleading Doctrine, 108 MICH. L. REV. (forthcoming 2009) (manuscript at 13-18), available at http://ssrn.com/abstract=1349951 (arguing that pleading today centrally requires a complaint to describe events about which there is a presumption of impropriety).
(75) Kendall W. Hannon, Note, Much Ado About Twombly ? A Study on the Impact of Bell Atlantic Corp. v. Twombly on 12(b)(6) Motions, 83 NOTRE DAME L. REV. 1811 (2008).
(76) See id. at 1835-38; cf. Joseph A. Seiner, The Trouble with Twombly: A Proposed Pleading Standard for Employment Discrimination Cases, 2009 U. ILL. L. REV. (forthcoming), available at http://ssrn.com/abstract=1273713 (finding a small increase in the rate of dismissal among a small example of Westlaw cases, but the methodology of searching for the permissive Conley in early cases and the restrictive Twombly in the later cases would bias the sample in favor of increasing dismissal).
(77) Hannon, supra note 75, at 1837 (reporting a civil rights dismissal rate of 41.7% under the pre-Twombly standard and 52.9% under Twombly).
(78) See generally Stephen B. Burbank, Ignorance and Procedural Law Reform: A Call for a Moratorium, 59 BROOK. L. REV. 841, 845-46 (1993) (noting that there was "little relevant empirical evidence" when the Advisory Committee adopted the mandatory disclosure rules) ; Carl Tobias, Civil Justice Delay and Empirical Data: A Response to Professor Heise, 51 CASE W. RES. L. REV. 235, 237-38, 244-46 (2000) (describing the events leading up to the adoption of the amendments).
(79) See FED. R. CIV. P. 26(a) (1993) (amended 2000).
(80) See id. 26(a)(1), (2), (3).
(81) Id. 26(a)(1).
(82) See DONNA STIENSTRA, FED. JUDICIAL CTR., IMPLEMENTATION OF DISCLOSURE IN UNITED STATES DISTRICT COURTS WITH SPECIFIC ATTENTION TO COURTS' RESPONSES TO SELECTED AMENDMENTS TO FEDERAL RULE OF CIVIL PROCEDURE 26, at 4 (1998).
(83) FED. R. CIV. P. 26(a)(1) advisory committee's note on 1993 amendments ("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.").
(84) See Linda S. Mullenix, Hope over Experience: Mandatory Informal Discovery and the Politics of Rulemaking, 69 N.C. L. REV. 795, 810 (1991) (observing as to the proposed disclosure rule that "there is virtually no empirical study of the current practice of such informal discovery, the efficacy of such experiences, or the results of informal discovery").
(85) Wayne D. Brazil, The Adversary Character of Civil Discovery: A Critique and Proposals for Change, 31 VAND. L. REV. 1295, 1348 (1978).
(86) William W Schwarzer, The Federal Rules, the Adversary Process, and Discovery Reform, 50 U. PITT. L. REV. 703, 721-23 (1989).
(87) See Clermont & Eisenberg, supra note 1, at 133.
(88) See Charles W. Sorenson, Jr., Disclosure Under Federal Rule of Civil Procedure 26(a)--"Much Ado About Nothing?, "46 HASTINGS L.J. 679, 687 (1995).
(89) THOMAS E. WILLGING ET ALL., FED. JUDICIAL CTR., DISCOVERY AND DISCLOSURE PRACTICE, PROBLEMS, AND PROPOSALS FOR CHANGE (1997); see also Thomas E. Willging et al., An Empirical Study of Discovery and Disclosure Practice Under the 1993 Federal Rule Amendments, 39 B.C.L. REV. 525 (1998).
(90) JAMES S. KAKALIK ET AL., RAND INST. FOR CIVIL JUSTICE, DISCOVERY MANAGEMENT (1998).
(91) WILLGING ET AL., supra note 89, at 1.
(92) Id. at 5-6.
(93) Id. at 6.
(95) KAKALIK ET AL, supra note 90, at 4.
(96) Id. at 5.
(97) Id. at 48-52.
(98) Id. at 51-52.
(99) See FED. R. CIV. P. 26(a)(1).
(100) Id. 26(a)(1)(A)(i); see also id. 26(a)(1)(A)(ii).
(101) FED. R. Cry. P. 26(a)(1) (1993) (amended 2000).
(102) Kuo-Chang Huang, Mandatory Disclosure: A Controversial Device with No Effects, 21 PACE L. REV. 203 (2000) ; see also Andrew T. Hayashi, The Effects of Mandatory Disclosure (July 13, 2008), available at http://ssrn.com/abstract=1212822 (evaluating the effect of the Rule 26(a) early disclosure requirement on litigation outcomes by using empirical analysis).
(103) Huang, supra note 102, at 242-43.
(104) Id. at 243-44.
(105) Id. at 263.
(107) See also Jeffrey W. Stempel, Ulysses Tied to the Generic Whipping Post: The Continuing Odyssey of Discovery "Reform, " LAW & CONTEMP. PROBS., Spring/Summer 2001, at 197, 225-28 (noting that mandatory disclosure has not had "much salutary impact" and calling for further study).
(108) For all the rule amendments, with their explanations by advisory committee notes, see Kevin M. Clermont, History of Amendments to the Federal Rules of Civil Procedure (Apr. 2007), http://legall.cit.cornell.edu/kevin/statsupps/ articles/article.htm.
(109) See Judith A. McKenna & Elizabeth C. Wiggins, Empirical Research on Civil Discovery, 39 B.C. L. REV. 785, 787 n.11 (1998) (identifying the two major studies as WILLIAM A. GLASER, PRETRIAL DISCOVERY AND THE ADVERSARY SYSTEM (1968), and PAUL R. CONNOLLY ET AL., FED. JUDICIAL CTR., JUDICIAL CONTROLS AND THE CIVIL LITIGATIVE PROCESS (1978)); cf. Daniel P. Kessler & Daniel L. Rubinfeld, Empirical Study of the Civil Justice System, in 1 HANDBOOK OF LAW AND ECONOMICS 343, 379-80 (A. Mitchell Polinsky & Steven Shavell eds., 2007) (flagging also George B. Shepherd, An Empirical Study of the Economics of Pretrial Discovery, 19 INT'L REV. L. & ECON. 245 (1999)).
(110) Kuo-Chang Huang, Does Discovery Promote Settlement? An Empirical Answer, 6 J. EMPIRICAL LEGAL STUD. (forthcoming 2009), available at http://ssrn.com/abstract= 995349.
(111) For background, see TAIWAN CODE OF CIVIL PROCEDURE (Kuo-Chang Huang trans., Richard Lawton Thurston ed., 2006); CHANG-FA Lo, THE LEGAL CULTURE AND SYSTEM OF TAIWAN (2006); Hungdah Chiu & Jyh-pin Fa, The Legal System of the Republic of China in Taiwan, in 2A MODERN LEGAL SYSTEMS CYCLOPEDIA, at ch. 12 (1989); Liane Newton & Wang Jong, A Research Guide to Taiwan (ROC) Law, 3 J. CHINESE L. 257 (1989); Tay-sheng Wang, The Legal Development of Taiwan in the 20th Century: Toward a Liberal and Democratic Country, 11 PAC. RIM L. & POL'Y J. 531 (2002).
(112) See Huang, supra note 110 (manuscript at 12-14).
(113) Id. (manuscript at 19).
(114) Id. (manuscript at 33-34).
(115) See FIELD ET AL., supra note 20, at 1266-89; Chiorazzi et al., supra note 61, at 137-38.
(116) Tobias, supra note 78, at 244; see Thomas E. Willging, Past and Potential Uses of Empirical Research in Civil Rulemaking, 77 NOTRE DAME L. REV. 1121 (2002). But cf. Robert G. Bone, The Empirical Turn in Procedural Rule, Making: Comment on Walker, 23 J. LEGAL STUD. 595, 597-98, 613 (1994) (stressing the desirability of limits on a proposal to require such research before rulemaking); Bryant G. Garth, Observations on an Uncomfortable Relationship: Civil Procedure and Empirical Research, 49 ALA. L. REV. 103, 106-13 (1997) (cataloguing the difficulties of such reform-oriented research); Richard Marcus, Confessions of a Federal "Bureaucrat": The Possibilities of Perfecting Procedural Reform, 35 W. ST. U. L. REV. 103, 113-16 (2007) (arguing that empiricists have played and will play only a small role in rulemaking). See generally Symposium, Empirical Studies of Civil Procedure, LAW & CONTEMP. PROBS., Summer 1988, at 1 (discussing how empirical analysis can further the study of civil procedure and including articles involving original empirical analysis).
(117) See Tobias, supra note 78, at 242 & n.36, 245 & n.46.
(118) Id. at 249.
(119) See Michael Heise, The Future of Civil Justice Reform and Empirical Legal Scholarship: A Reply, 51 CASE W. RES. L. REV. 251, 251-54 (2000).
(120) See Chiorazzi et al., supra note 61, at 128-31 (listing no studies of dispositive motions, other than summary judgment, as of 1988); cf. Hoffman, supra note 74, at 1223 n.28 (citing a few strands of early evidence on motions to dismiss). A more recent study of motions to dismiss, giving numbers somewhat lower than earlier estimates, found in 1988 that the percentage of federal cases involving one or more Federal Rule 12(b) (6) motions was 13% of the sample; the court decided such a motion in 10%, and granted it in 6%, of all cases in the sample; and grant of the motion resulted in termination of 3% of the sample. See THOMAS E. WILLGING, FED. JUDICIAL CTR., USE OF RULE 12(B)(6) IN TWO FEDERAL DISTRICT COURTS 8-9 (1989); cf. INST. FOR THE ADVANCEMENT OF THE AM. LEGAL SYSTEM, CIVIL CASE PROCESSING IN THE FEDERAL DISTRICT COURTS 47-49 (2009) (finding that 15% of federal cases in 2005 involved one or more motions to dismiss of any kind, of which 54% were granted in whole or part). I expect that Bell Atlantic will trigger more studies of motions to dismiss, as defendants begin to use these motions to feel out the plaintiffs' proof and as courts struggle to divine the new standard of decision. See, e.g., sources cited supra notes 75-76.
(121) See CLERMONT, supra note 12, at 79-83.
(122) FED. R. CIV. P. 56(c). For a recent discussion of a state analogy, see Robert A. Sedler, The Michigan Supreme Court Diminishes the Right to Trial by Jury in Civil Cases 7-23 (Wayne St. U. L. Sch. Legal Studies Research Paper Series, No. 08-13, 2008), available at http://ssrn.com/abstract=1118807.
(123) See Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 247-52 (1986).
(124) See id. at 255 ("Credibility determinations, the weighing of evidence, and the drawing of legitimate inferences from the facts are jury functions, not those of the judge, whether he is ruling on a motion for summary judgment or for a directed verdict.").
(125) See Chiorazzi et al., supra note 61, at 128-31 (listing only two studies of summary judgment as of 1988: an early FJC study reported in JOE S. CECIL & C.R. DOUGLAS, FED. JUDICIAL CTR., SUMMARY JUDGMENT PRACTICE IN THREE DISTRICT COURTS (1987), and the flawed study by William P. McLauchlan, An Empirical Study of the Federal Summary Judgment Rule, 6J. LEGAL STUD. 427 (1977), with the latter criticized at length by Stephen B. Burbank, Vanishing Trials and Summary Judgment in Federal Civil Cases: Drifting Toward Bethlehem or Gommorrah? 1 J. EMPIRICAL LEGAL STUD. 591, 607-11 (2004)); David L. Shapiro, The Story of Celotex: The Role of Summary Judgment in the Administration of Civil Justice, in CIVIL PROCEDURE STORIES, supra note 58, at 359, 381-83 (critiquing more recent studies of summary judgment).
(126) Compare Shapiro, supra note 125, at 386 & n.89 (citing authorities "who favor judicial efficiency in the face of what they see as increasing strains on the system and who may be willing to give second place to the significance of the 'day-in-court' tradition"), with id. at 386 & n.90 (citing authorities "who see summary judgment as a potentially costly device, as one that is in a sense 'elitist' because it tends to favor defendants, and as one that threatens both to divorce results from contextual consideration of all the evidence as presented by live witnesses in open court and to undermine the role of the jury").
(127) See id. at 386 n.90.
(128) See id. at 386 n.89.
(129) 550 U.S. 372 (2007); cf. Buckley v. Haddock, 292 F. App'x 791 (11th Cir. 2008) (granting summary judgment despite a videotape, which is available at http://www. youtube.com/watch?v=SWC7iSGCk-s).
(130) Scott, 550 U.S. at 374.
(131) Id. at 375.
(132) Id. at 381.
(133) Id. at 383-84.
(134) Id. at 380.
(135) Id. at 389 (Stevens, J., dissenting).
(136) Id. at 390.
(137) See id. at 391-92; David Kessler, Justices in the Jury Box: Video Evidence and Summary Judgment in Scott v. Harris, 127 S. Ct. 1769 (2007), 31 HARV. J.L. & PUB. POL'Y 423, 429-30 (2008); see also Dan M. Kahan et al., Whose Eyes Are You Going to Believe? Scott v. Harris and the Perils of Cognitive Illiberalism, 122 HARV. L. REV. 837, 866 (2009) (showing that about a quarter of tested people who viewed the Scott videotape, which is available at http://www.youtube.com/watch?v=DBY2y2YsmN0, disagreed with the view that deadly force had been justified); Howard M. Wasserman, Orwell's Vision: Video and the Future of Civil Rights Enforcement, 68 MD. L. REV. (forthcoming 2009), available at http://ssrn.com/abstract=1219162 (advocating caution in granting summary judgment based on video evidence in civil rights cases); Howard M. Wasserman, Video Evidence and Summary Judgment: The Procedure of Scott v. Harris, 91 JUDICATURE 180, 182-84 (2008) (challenging more generally the objectivity of video evidence).
(138) Scott, 550 U.S. at 397 (Stevens, J., dissenting).
(139) Joe S. Cecil et al., A Quarter-Century of Summary Judgment Practice in Six Federal District Courts, 4 J. EMPIRICAL LEGAL STUD. 861 (2007); see also INST. FOR THE ADVANCEMENT OF THE AM. LEGAL SYSTEM, supra note 120, at 49-52 (finding consistently that 17% of federal cases in 2005 involved one or more summary judgment motions, of which 54% were granted in whole or in part).
(140) Cecil et al., supra note 139, at 874-76.
(141) See also Theodore Eisenberg & Charlotte Lanvers, Summary Judgment Rates over Time, Across Case Categories, and Across Districts: An Empirical Study of Three Large Federal Districts (Cornell L. Sch. Research Paper, No. 08-022, 2008), available at http://ssrn. com/abstract=1138373 (reinforcing this point, while pointing out some limitations of the FJC study). For example, the sizable case category of employment discrimination (AO code 442) is often empirically distinctive. See Kevin M. Clermont & Stewart J. Schwab, Employment Discrimination Plaintiffs in Federal Court: From Bad to Worse?, 3 HARV. L. & POL'Y REV. 103 (2009) [hereinafter Clermont & Schwab, Employment Discrimination Update]; Kevin M. Clermont & Stewart J. Schwab, How Employment Discrimination Plaintiffs Fare in Federal Court, 1 J. EMPIRICAL LEGAL STUD. 429 (2004) [hereinafter Clermont & Schwab, Employment Discrimination]; Kevin M. Clermont et al., How Employment-Discrimination Plaintiffs Fare in the Federal Courts of Appeals, 7 EMP. RTS. & EMP. POL'Y J. 547 (2004). A fine study of this particular case category, Vivian Berger et al., Summary Judgment Benchmarks for Settling Employment Discrimination Lawsuits, 23 HOFSTRA LAB. & EMP. L.J. 45 (2005), looked at a sample of employment discrimination cases in two districts during the period around 2000 and found that the court decided summary judgment motions by defendants in 23% of the cases, with the defendants experiencing a 64% success rate on those motions (with a much higher rate against pro se plaintiffs). Id. at 53-57. Thus, summary judgment is a common means of disposing of this category of cases. See Laura Beth Nielsen et al., Uncertain Justice: Litigating Claims of Employment Discrimination in the Contemporary United States 16-18 (Am. Bar Ass'n Found. Research Paper Series, No. 08-04, 2008), available at http:// ssrn.com/abstract=1093313; Schneider, supra note 74, at 16-17. Moreover, a sampling of online judicial opinions on defendants' summary judgment motions in Title VII employment discrimination cases showed a statistically significant effect of the political party of the President who had appointed the trial judge. John Friedl & Andre Honoree, Is Justice Blind? Examining the Relationship Between Presidential Appointments of Judges and Outcomes in Employment Discrimination Cases, 38 CUMB. L. REV. 89 (2007).
(142) Cecil et al., supra note 139, at 882.
(143) Id. at 883.
(145) Common knowledge was that the granting of summary judgment became rampant only after the Supreme Court had explicated and blessed the summary judgment device in a trilogy of cases in 1986: Celotex Corp. v. Catrett, 477 U.S. 317 (1986); Anderson v. Liberty Lobby, Inc., 477 U.S. 242 (1986); and Matsushita Electric Industrial Co. v. Zenith Radio Corp., 475 U.S. 574 (1986). See Shapiro, supra note 125, at 379-81 (citing sources expressing that view). The FJC study proves this common knowledge wrong, in that the increase in granting summary judgment predates the trilogy. See Cecil et al., supra note 139, at 902. On the possible link between the long-term increase in summary judgment and the long-term decline in civil trial, see infra note 186.
(146) Cecil et al., supra note 139, at 886-89.
(147) The proposal, published for public comment in August 2008, appears at Report of the Civil Rules Advisory Committee 21-56, available at http://www.uscourts. gov/rules/Reports/CV_Report.pdf; see Civil Rules Advisory Comm., Judicial Conf. of the U.S., Minutes Nov. 8-9, 2007, at 19-30, available at http://www.uscourts.gov/ rules/Minutes/CV11-2007-min.pdf; Civil Rules Advisory Comm., Judicial Conference of the U.S., Minutes Apr. 19-20, 2007, at 3-22, available, at http://www.uscourts.gov/ rules/Minutes/CV04-2007-min.pdf; Civil Rules Advisory Comm., Judicial Conference of the U.S., Minutes Sept. 7-8, 2006, at 24-30, available at http://www.uscourts.gov/ rules/Minutes/CV09-2006-min.pdf; see also Adam N. Steinman, An Ounce of Prevention: Solving Some Unforeseen Problems with the Proposed Amendments to Rule 56 and the Federal Summary Judgment Process, 103 Nw. U. L. REV. COLLOQUY 230, 230-32 (2008), http:// www.law.northwestern.edu/lawreview/colloquy/2008/45/LRColl2008n45Steinman.pdf (discussing the rationale behind the Committee's proposal).
(148) See Report of the Civil Rules Advisory Committee, supra note 147, at 21-26.
(149) See, e.g., Memorandum from Joe Cecil & George Cort, Fed. Judicial Ctr., to Judge Michael Baylson (Nov. 2, 2007), available, at http://www.fjc.gov/public/pdf. nsf/lookup/insumjre.pdf/$file/insumjre.pdf (reporting that local rules requiring statements of uncontested fact seem to have little impact).
(150) HANS ZEISEE ET AL., DELAY IN THE COURT, at xxii (2d ed. 1978); see also INST. FOR THE ADVANCEMENT OF THE AM. LEGAL SYSTEM, supra note 120, at 1 (continuing the search for procedures to defeat the serious problem of delay); THE LAW'S DELAY (C.H. van Rhee ed., 2004) (featuring comparative and historical studies lamenting delay).
(151) See RICHARD A. POSNER, ECONOMIC ANALYSIS OF LAW [section] 21.14 (7th ed. 2007); Richard A. Posner, An Economic Approach to Legal Procedure and Judicial Administration, 2 J. LEGAL STUD. 399, 445-48 (1973).
(152) George L. Priest, Private Litigants and the Court Congestion Problem, 69 B.U. L. REV. 527 (1989).
(153) Id. at 557; see also Tracey E. George & Chris Guthrie, Induced Litigation, 98 Nw. U. L. REV. 545, 563 (2004) (arguing that the increased availability of court resources may fail to reduce congestion because of the corresponding increase in litigation demand); John Leubsdorf, The Myth of Civil Procedure Reform, in CIVIL JUSTICE CRISIS 53 (Adrian A.S. Zuckerman ed., 1999) (questioning more generally the efficacy of procedural reform).
(154) Priest, supra note 152, at 535-39.
(155) Michael Heise, Justice Delayed?: An Empirical Analysis of Civil Case Disposition Time, 50 CASE W. RES. L. REV. 813, 848 (2000); see also Priest, supra note 152, at 535, 537 (forwarding the "congestion equilibrium hypothesis" while suggesting "that there is likely to be some equilibrium level of delay within any jurisdiction," but recognizing that procedure can be made more or less just at any given equilibrium level of delay and also acknowledging that certain reforms such as increasing court costs or altering the local legal culture could lower the equilibrium somewhat).
(156) Theodore Eisenberg & Kevin M. Clermont, Trial by Jury or Judge: Which Is Speedier?, 79 JUDICATURE 176, 176-78 (1996) (observing that means tell the same story as medians).
(157) See, e.g., POSNER, supra note 151, [section] 21.15, at 628 ("Court queues are longest for parties seeking civil jury trials."); Leon Sarpy, Civil Juries, Their Decline and Eventual Fall, 11 LOY. L. REV. 243, 255-56 (1963) (similar implication); see also GORDON BERMANT ET AL., FED. JUDICIAL CTR., PROTRACTED CML TRIALS 43-45 (1981) (giving survey results).
(158) Eisenberg & Clermont, supra note 156, at 199.
(159) See Heise, supra note 155, at 815-16.
(160) See Thomas H. Cohen & Steven K. Smith, Civil Trial Cases and Verdicts in Large Counties, 2001, BULLETIN (Bureau of Just. Stat., Wash. D.C.), Apr. 2004, at 1, 3, 8 (noting that the sample also included a small number of real property cases); see also Theodore Eisenberg et al., Litigation Outcomes in State and Federal Courts: A Statistical Portrait, 19 SEATTLE U. L. REV. 433, 449 (1996) (explaining that these state data on juries and judges contrast the median time from filing to verdict with the median time from filing to judgment, respectively). The pattern persisted in a more recent study. See Lynn Langton & Thomas H. Cohen, Civil Bench and Jury Trials in State Courts, 2005, SPEC. REP. (Bureau of just. Stat., Wash. D.C.), Oct. 2008, at 1, 8. A separate but significant point is that in terms of time on the docket, state jury cases take longer than federal jury cases. See Thomas H. Cohen, General Civil Jury Trial Litigation in State and Federal Courts: A Statistical Portrait, 5 J. EMPIRICAL LEGAL STUD. 593, 606-08 (2008).
(161) Cohen & Smith, supra note 160, at 1.
(162) The studies' difference in time periods is not the explanation either. Our study found that judge-tried cases took longer than jury-tried cases every year from fiscal year 1947 through fiscal year 1994, while it performed its specific computations and analyses on fiscal years 1979-1994. See Eisenberg & Clermont, supra note 156, at 178-79. The state data include 11,675 cases (76% jury-tried) from only calendar year 2001. Nat'l Ctr. for State Courts, Civil Justice Survey of State Courts, 2001, http:// www.icpsr.umich.edu/cocoon/NACJD/STUDY/03957.xml. Accordingly, I did a new study of the federal data for fiscal years 1995-2006, a period bracketing the state year and including 16,709 cases (73% percent jury-tried). I dropped 276 extraordinarily delayed jury-tried personal injury cases all terminated in the District of Puerto Rico on December 4, 2000. The federal situation has not changed since our earlier study: judge-tried cases still take longer than jury-tried cases every year, with the median judge-tried case now spending 641 days on the district court docket, compared to the median jury-tried case terminating in 581 days.
(163) The state data, see Nat'l Ctr. for State Courts, supra note 162, indicate on average that state jury trials start almost six months after state bench trials, while our study suggested that federal judges do not delay jury trials by much if at all. See Eisenberg & Clermont, supra note 156, at 180 & n.21.
(164) The state data, see Nat'l Ctr. for State Courts, supra note 162, indicate on average that state judges delay only about twenty-three days after the trial ends, while our study suggested that federal judges delay more than three months. See Eisenberg & Clermont, supra note 156, at 180, 199.
(165) See POSNER, supra note 151, [section] 21.15, at 628 ('Jury trials are more costly than bench trials.... Parties are therefore 'charged' more for jury trials by being made to wait in line longer."); Posner, supra note 151, at 447 (recommending "a substantial fee" to discourage demands for jury trial). Policies to discourage jury trial are not unthinkable, as courts have long discouraged criminal jury trials by imposing harsher sentences on those defendants who pursued a jury trial rather than a bench trial. See Thomas M. Uhlman & N. Darlene Walker, "He Takes Some of My Time; I Take Some of His": An Analysis of Judicial Sentencing Patterns injury Cases, 14 LAW & SOC'Y REV. 323 (1980).
(166) Taking the state data for 2001 and the federal data for 1995-2006, see supra note 162, I computed, for each state county and for each federal district, the difference between the median days for a jury-tried case and the median days for a judge-tried case. The standard deviation is about 158 for those county figures and about 114 for those district figures, meaning that states show more variability in their relative speeds of handling trials than do the federal courts. Some of the federal variability may indeed stem from the effect of the local state attitude; the federal and state data do show some correlation, so that as the county's tendency to delay jury trials increases, the local federal district's tendency to speed jury trials decreases. See also infra note 191 (noting that the number of state jury trials has fallen more precipitously than state judge trials perhaps due to state procedures that discourage jury trials).
(167) See generally FIELD ET AL., supra note 20, at 125-36 (giving background on settlement process); William L.F. Felstiner et al., The Emergence and Transformation of Disputes: Naming, Blaming, Claiming ..., 15 LAW & SOC'Y REV. 631, 636 (1981) ("[O]nly a small fraction of injurious experiences ever mature into disputes."); Marc Galanter, Reading the Landscape of Disputes: What We Know and Don't Know (and Think We Know) About Our Allegedly Contentious and Litigious Society, 31 UCLA L. REV. 4, 11-36 (1983) (exploring "current American disputing patterns").
(168) See Galanter, supra note 167, at 11-36.
(169) See, e.g., Carrie Menkel-Meadow, Whose Dispute Is It Anyway?: A Philosophical and Democratic Defense of Settlement (in Some Cases), 83 GEO. L.J. 2663, 2671-91 (1995) (arguing that settlements are not inherently inferior to adjudicated outcomes). In reality, however, reformers would not find it easy to raise, or for that matter lower, the settlement rate. See supra text accompanying note 153.
(170) See Robert H. Mnookin & Lewis Kornhauser, Bargaining in the Shadow of the Law: The Case of Divorce, 88 YALE L.J. 950, 997 (1979).
(171) See Richard A. Posner, The Summary Jury Trial and Other Methods of Alternative Dispute Resolution: Some Cautionary Observations, 53 U. Chi. L. Rev. 366, 388 (1986).
(172) See id. at 372, 388 (suggesting that there is an optimal settlement rate); see also Owen M. Fiss, Against Settlement, 93 YALE L.J. 1073 (1984) (discussing the problem of relying too heavily on settlement).
(173) See Samuel R. Gross & Kent D. Syverud, Don't Try: Civil Jury Verdicts in a System Geared to Settlement, 44 UCLA L. REV. 1, 1-4 (1996).
(174) See Marc Galanter & Mia Cahill, "Most Cases Settle": Judicial Promotion and Regulation of Settlements, 46 STAN. L. REV. 1339, 1354-59, 1387-88 (1994); Stephen McG. Bundy, The Policy in Favor of Settlement in an Adversary System, 44 HASTINGS L.J. 1, 73-78 (1992).
(175) See Chiorazzi et al., supra note 61, at 131-37.
(176) Richard E. Miller & Austin Sarat, Grievances, Claims, and Disputes: Assessing the Adversary Culture, 15 LAW & SOC'Y REV. 525, 534-37 (1981); see also David M. Trubek et al., The Costs of Ordinary Litigation, 31 UCLA L. REV. 72, 86-87 (1983). For a comparison to numbers from abroad, see Masayuki Murayama, Japanese Disputing Behavior Reconsidered 7-9 (June 2008) (unpublished manuscript), available at http://www. iias.sinica.edu.tw/resource/970621/970621_1_10.pdf (suggesting that percentages in other countries--Japan and England--are similar at early steps of the dispute pyramid, before persons get involved in the legal system itself).
(177) See Miller & Sarat, supra note 176, at 544 (showing 50 total court filings out of 1000 total grievances).
(178) We showed the fate of cases over the years from 1979 to 2000. The coding of disposition method became consistent enough to use only in fiscal year 1979, and data were then available only through fiscal year 2000.
(179) Gillian K. Hadfield, Where Have All the Trials Gone? Settlements, Nontrial Adjudications, and Statistical Artifacts in the Changing Disposition of Federal Civil Cases, 1 J. EMPIRICAL LEGAL STUD. 705, 723-28 & tbls. 4, 5, 6 (2004) (faulting especially AO codes 6, 14, and 17, although their errors are offsetting). Her conclusions included that the AO data, without impractical unscrambling that involves at least auditing the cases' files, tended to overstate the decline of the civil trial, although this decline certainly existed. Id. at 728-33. Furthermore, she saw an increase (over time) in pretrial disposition as offsetting that decline in civil trials and a slight decline in settlement, although this conclusion rested heavily on a comparison of recent data to incommensurable data from 1970. Id.
Finally, she argued that the AO classifies many nonfinal dispositions as settlements, so that she found only a 51.4% settlement rate--3.5% abandonment, 5.4% default judgment, 2.0% consent judgment, and 40.5% "settled"--in a few hundred cases sampled from fiscal year 2000. See id. at 730 tbl. 7 (omitting prisoner cases and government recovery of overpayments and student loans). The AO data, analyzed by my method without her unscrambling, would show a 66.2% settlement rate, not 51.4%, for the same sort of cases from the same fiscal year. However, her 32.1% rate for nonfinal dispositions derived in part from her broad definition of nonfinal disposition, including not only transfer and the like, but also voluntary dismissal recorded without any indicator of settlement and even dismissal pending consummation of settlement. Many of these so-called nonfinal dispositions, then, are actually agreed settlements or are effectively settlements; most of the remaining dispositions are in the nature of temporary dismissals that will result in eventual settlement, if not immediate abandonment, because settlement is how most cases end anyway. Therefore, her settlement rate is understated. An alternative, but still crude, comparison of our results would be to change the denominators, by omitting her nonfinal dispositions and my so-called other dispositions, the latter being 13.3%. Then, her settlement rate and mine are both 76%.
(180) I have refined the division of AO codes along the following lines. First, tried cases are now those with a method-of-disposition value of 7 to 9. Second, cases adjudicated without trial are those with a method-of-disposition value of 6, 15, 17, 19, or 20. Third, settled cases are those with a method-of-disposition value of 2 (lack of prosecution), 4 (default judgment), 5 (consent judgment), 12 to 14 (dismissals: voluntary, settled, or other), or 18 (statistical closing). Code 3 switched in usage around 1991 from voluntary dismissal to dismissal for lack of jurisdiction, so I grouped its earlier usage with settlement, but its usage in 1991 and later with nontrial adjudication. Fourth, other dispositions are all remaining method-of-disposition values, predominantly remand or transfer to another court. See Clermont & Schwab, Employment Discrimination, supra note 141, at 440 n.14.
The definition of settlement rate is critical. See Theodore Eisenberg & Charlotte Lanvers, What Is the Settlement Rate and Why Should We Care?, 6 J. EMPIRICAL LEGAL STUD. 111, 112-15 (2009). The different settlement rates that others sometimes invoke, e.g., Hadfield, supra note 179, at 730, usually stem from different definitions. My interest here is in the grievance pyramid, and hence in the difference between cases that the system has to adjudicate and those that exit the sides of the pyramid. To make this distinction from contested judgments, I am defining settlement rate in the district courts to include the plaintiffs abandonment or the defendant's concession, as well as compromise by private negotiations or through ADR. I could alter my definition in various ways. First, if I were interested in disposition without any judicial input, I could add the requirement that the procedural progress code shows no court action, which would drop the 67.7% settlement rate for fiscal year 2005 all the way to 17.5%. Second, if I were more interested in compromise by the parties, I could instead excise method-of-disposition codes 2 (lack of prosecution) and 4 (default judgment) from the realm of settlement, which would drop that 67.7% settlement rate to 61.5%. Third, if I were interested only in dispositions of a more final sort, I could omit my so-called other dispositions from the denominator, which would raise that 67.7% rate to 75.5%. Fourth, no matter what my interest, I could try to get inside the dismissals coded 12, 13, 14, or 18 to determine which entries represented compromise and which represented adjudication; but in deference to the limits of time, I take comfort in Professor Hadfield's findings that the adjudicated dismissals in those particular codes are offset by the settlements erroneously included within the codes for adjudicated dispositions. See id. at 723-28.
(181) See LEONIDAS RALPH MECHAM, ADMIN. OFFICE OF THE U.S. COURTS, JUDICIAL BUSINESS OF THE UNITED STATES COURTS 2005, at 155 tbl.C-1 (2005), available at http://www.uscourts.gov/judbus2005/appendices/c1-pdf. Settlement practice varies considerably with locale and case type, as well as over time. See Stewart J. Schwab & Theodore Eisenberg, Explaining Constitutional Tort Litigation: The Influence of the Attorney Fees Statute and the Government as Defendant, 73 CORNELL L. REV. 719, 733 (1988). On the determinants of settlement more generally, see Kessler & Rubinfeld, supra note 109, at 381-83.
(182) For these percentages, I am categorizing the AO data that underlie the table cited supra note 181 by using the codes specified supra note 180.
(183) For the percentage of litigated judgments, I used 20.7% + 1.3% + .22(10.3%) = 24.3%. The comparable percentage for 1980, at the time of the telephone survey, was 23.8%. For the percentage on appeal, see infra note 242 and accompanying text.
(184) The trend of the vanishing civil trial is apparent from the hard copy of the Annual Report of the Director of the Administrative Office of the United States Courts. Over the years its Table C-4, prepared with the procedural progress codes for cases terminated during or after trial, shows a steady decrease from almost 12% of civil terminations having reached trial in the 1960s to the current levels approaching 1%. During that period, the growing number of federal judges managed to increase the absolute number of civil trials as the caseload grew, until reaching a peak of 12,570 trials in fiscal year 1985 according to the AO's measure. See Shari Seidman Diamond & Jessica Bina, Puzzles About Supply-Side Explanations for Vanishing Trials: A New Look at Fundamentals, 1 J. EMPIRICAL LEGAL STUD. 637, 639-45, 649 (2004) (emphasizing the inadequacy of increase in number of federal judges, as well as showing the increase in weighted filings per judge over time). But civil trials per year have since dropped, so that in fiscal year 2006 there were many fewer civil trials (3555) than in fiscal year 1961 (5553 trials). The AO reports an uptick for fiscal year 2007.
(185) See Ad Hoc Comm. on the Future of the Civil Trial, Am. College of Trial Lawyers, The "Vanishing Trial": The College, the Profession, the Civil Justice System, 226 F.R.D. 414, 414 (2005) ("'For every complex problem there is an answer that is clear, simple, and wrong.'" (quoting H.L. Mencken)); Clermont & Eisenberg, supra note 1, at 142-44; Symposium, The Vanishing Trial, 1 J. EMPERICAL LEGAL STUD. 459 (2004); Symposium, Vanishing Trial, 2006 J. Disc. RESOL. 1.
(186) See, e.g., Sara Sun Beale, Federalizing Crime: Assessing the Impact on the Federal Courts, 543 ANNALS AM. ACAD. POL. & SOC. SCI. 39, 46-48 (1996) (emphasizing growth of the competing criminal trial docket); Lisa Blomgren Bingham, When We Hold No Truths to Be Self-Evident: Truth, Belief, Trust, and the Decline in Trials, 2006 J. DISP. RESOL. 131, 154 (emphasizing loss of public belief in truth) ; Dennis J. Drasco, Public Access to Information in Civil Litigation vs. Litigant's Demand for Privacy: Is the "Vanishing Trial" an Avoidable Consequence?, 2006 J. DISP. RESOL. 155, 155-57 (emphasizing litigants' desire for confidentiality); Gross & Syverud, supra note 173, at 1-3 (emphasizing systemic pressure to settle); Patrick E. Higginbotham, Judge Robert A. Ainsworth, Jr. Memorial Lecture, Loyola University School of Law: So Why Do We Call Them Trial Courts?, 55 SMU L. REV. 1405, 1420 (2002) (emphasizing increased litigation costs); Patrick E. Longan, The Shot Clock Comes to Trial: Time Limits for Federal Civil Trials, 35 ARIZ. L. REV. 663, 669-72, 677-78 (1993) (emphasizing, in part, the growth in number and complexity of civil cases); Judith Resnik, Failing Faith: Adjudicatory Procedure in Decline, 53 U. CHI. L. REV. 494, 526-39 (1986) (emphasizing judicial and cultural assumptions); Hope Viner Samborn, The Vanishing Trial, A.B.A. J., Oct. 2002, at 24, 26-27 (emphasizing the push toward ADR); Stephen C. Yeazell, The Misunderstood Consequences of Modern Civil Process, 1994 WIS. L. REV. 631, 632-39 (emphasizing the increase in judges' pretrial tasks).
One particularly suggestive study links, albeit tentatively and partially, the decline in the civil trial to the increase in summary judgment grants. See Burbank supra note 125, at 617-18; see also supra text accompanying notes 139-147 (discussing trends in summary judgment rates). A number of other articles intuited the same link. See, e.g., Arthur R. Miller, The Pretrial Rush to Judgment: Are the "Litigation Explosion," "Liability Crisis," and Efficiency Cliches Eroding Our Day in Court and Jury Trial Commitments ?, 78 N.Y.U.L. REV. 982, 1048-57 (2003) (observing the increased use of summary judgment following the Supreme Court's 1986 trilogy); Martin H. Redish, Summary Judgment and the Vanishing Trial: Implications of the Litigation Matrix, 57 STAN. L. REV. 1329, 1333 (2005) ("Whatever influence these factors have actually had in the reduction in the number of trials, however, it is not unreasonable to suspect that one of the primary contributors to this result, at least at the federal level, has been the Supreme Court's substantial modification and expansion of the modern doctrine of summary judgment."); Milton I. Shadur, Trials or Tribulations (Rule 56 Style)?, LITIGATION, Winter 2003, at 5, 5 (noting that the Supreme Court's 1986 trilogy "has worked a systemic sea change" in the granting of summary judgment motions).
(187) See Marc Galanter, The Vanishing Trial: An Examination of Trials and Related Matters in Federal and State Courts, 1 J. EMPIRICAL LEGAL STUD. 459, 492-500 (2004).
(188) See Ostrom et al., supra note 21, at 770-72.
(189) See Herbert M. Kritzer, Disappearing Trials ? A Comparative Perspective, 1 J. EMPIRICAL LEGAL STUD. 735, 752-54 (2004) (examining trial patterns in England, Wales, and the Canadian province of Ontario, and finding a reasonably clear pattern of declining rates of civil trials). But cf. Carolien Klein Haarhuis & Bert Niemeijer, Vanishing or Increasing Trials in the Netherlands?, 2006J. DISP. RESOL. 71, 76-97 (showing that various Dutch social pressures have increased the civil "trial" rate).
(190) The explanation does not reside in a single case category, such as employment discrimination where the jury right has expanded in the recent past. See Clermont & Schwab, Employment Discrimination, supra note 141, at 432-38 (showing that the shift to jury trials for employment discrimination cases has offset the sharp decline in bench trials). The mass of cases, even with the employment discrimination cases omitted, would show virtually the same drop in absolute and relative use of judge trial.
(191) See supra text accompanying notes 156-166. State courts, which generally act in ways that relatively discourage jury trials, have exhibited the opposite pattern: jury trials have fallen more precipitously than a broadly defined set of judge trials. See Ostrom et al., supra note 21, at 770, 777.
(192) Here I used the procedural progress codes of 7 and 9--termination during and after jury trial--to define jury trial usage. However, I used the disposition method code of 9--judgment on court trial--to define judge trial usage. The reason for abandoning the procedural progress code for judge trials was that the AO unfortunately defines "trial" for procedural progress purposes to include all contested proceedings in which evidence is introduced, thus distortingly including a good number of motion hearings as judge trials. See Higginbotham, supra note 186, at 1405-06. However, the disposition method code did not become consistent until fiscal year 1979.
(193) See Stephen B. Burbank, Keeping Our Ambition Under Control: The Limits of Data and Inference in Searching for the Causes and Consequences of Vanishing Trials in Federal Court, 1 J. EMPIRICAL LEGAL STUD. 571 (2004); Hadfield, supra note 179; Margo Schlanger, What We Know and What We Should Know About American Trial Trends, 2006 J. DISP. RESOL. 35.
(194) Compare Paul Butler, The Case for Trials: Considering the Intangibles, 1 J. EMPIRICAL LEGAL STUD. 627, 629-35 (2004) (emphasizing the important "intangible" value trials have in society and expressing concern for the vanishing civil trial), with Stephen C. Yeazell, Getting What We Asked For, Getting What We Paid For, and Not Liking What We Got: The Vanishing Civil Trial, 1 J. EMPIRICAL LEGAL STUD. 943, 964-71 (2004) (offering an "equally plausible" view of the vanishing trial--that it is a natural and desirable sign of a maturing procedural system).
(195) See supra note 184.
(196) See supra note 145 and accompanying text.
(197) See supra text accompanying notes 152-55.
(198) See supra note 186.
(199) See Diamond & Bina, supra note 184, at 654-57 (observing that the civil trial rate varies inversely with the civil caseload per judge, while arguing generally that the perceived lack of the judicial system's capacity rather than the lack of demand by litigants for trial has driven the civil trial's decline).
(200) HARRY KALVEN, JR. & HANS ZEISEL, THE AMERICAN JURY (2d ed. 1971); see also Harry Kalven, Jr., The Dignity of the Civil Jury, 50 VA. L. REV. 1055, 1063-68 (1964) (comparing jury verdicts and judges' desired outcomes in surveyed trials). For early reviews of Kalven and Zeisel's study, see John Kaplan, Book Review, 115 U. PA. L. REV. 475 (1967) (reviewing the first edition), and Michael H. Walsh, The American Jury: A Reassessment, 79 YALE L.J. 142 (1969) (same). See also Valerie P. Hans & Neil Vidmar, The American Jury at Twenty-Five Years, 16 LAW & SOC. INQUIRY 323 (1991) (assessing The American Jury twenty-five years after its publication).
(201) KALVEN & ZEISEL, supra note 200, at 63-64.
(202) See Michael J. Saks, Enhancing and Restraining Accuracy in Adjudication, LAW & CONTEMP. PROBS., Autumn 1988, at 243, 246-48 (reporting research by Shari S. Diamond).
(203) KALVEN & ZEISEL, supra note 200, at 59.
(204) Id. at 64.
(205) See Kevin M. Clermont & Theodore Eisenberg, Trial by Jury or Judge: Transcending Empiricism, 77 CORNELL L. REV. 1124 (1992).
(206) See id. at 1136-37.
(207) See id. app. A, at 1175.
(208) See id. app. B. On the later perceived dangers of using mean recoveries with the AO data, see Theodore Eisenberg & Margo Schlanger, The Reliability of the Administrative Office of the U.S. Courts Database: An Initial Empirical Analysis, 78 NOTRE DAME L. REV. 1455, 1489-90 (2003).
(209) See Clermont & Eisenberg, supra note 205, at 1140-43.
(210) See id. at 1148-57.
(211) See id. at 1149-51.
(212) See NEIL VIDMAR & VALERIE P. HANS, AMERICAN JURIES 147-89, 267-338 (2007); Theodore Eisenberg et al., Juries, Judges, and Punitive Damages: An Empirical Study, 87 CORNELL L. REV. 743, 779 (2002); Chris Guthrie et al., Inside the Judicial Mind, 86 CORNELL L. REV. 777, 778, 826-27 (2001); Valerie P. Hans & Stephanie Albertson, Empirical Research and Civil Jury Reform, 78 NOTRE DAME L. REV. 1497, 1506-11 (2003).
(213) See Clermont & Eisenberg, supra note 205, at 1156-57.
(214) See id. at 1174.
(215) Numerous smaller studies give fairly consistent support to our results. See, e.g., Kevin M. Clermont & Theodore Eisenberg, Appeal from Jury or Judge Trial: Defendants' Advantage, 3 AM. L. & ECON. REV. 125, 144-45 (2001) [hereinafter Clermont & Eisenberg, Defendants' Advantage] (citing sources); Clermont & Eisenberg, supra note 205, at 1151-55 (same); Eric Helland & Alexander Tabarrok, Runaway Judges? Selection Effects and the Jury, 16 J.L. ECON. & ORG. 306, 329-30 (2000) (similarly attributing most jury/judge differences to selection effect). The same is true in studies of most specific case categories. See, e.g., Clermont & Eisenberg, supra note 205, at 1154 n.78 (citing sources); Elizabeth Graddy, Juries and Unpredictability in Products Liability Damage Awards, 23 LAW & POL'Y 29 (2001) (comparing judge and jury awards in product liability cases). Although one commentator concluded that some significant jury/ judge differences do exist, see Kimberly A. Moore, Judges, Juries, and Patent Cases--An Empirical Peek Inside the Black Box, 99 MICH. L. REV. 365, 383-408 (2000); Kimberly A. Moore, Jury Demands: Who's Asking?, 17 BERKELEY TECH. L.J. 847, 870-75 (2002), that result seems attributable to the uniqueness of the patent litigation that she studied. See Kevin M. Clermont & Theodore Eisenberg, Xenophilia or Xenophobia in U.S. Courts? Before and After 9/11, 4 J. EMPIRICAL LEGAL STUD. 441, 447-51 (2007) [hereinafter Clermont & Eisenberg, Xenophilia II]; see also Philippe Signore, On the Role of Juries in Patent Litigation (pt. 1), 83 J. PAT. & TRADEMARK OFF. SOC'Y 791, 821-24 (2001) (citing statistics indicating judge/jury differences in patent litigation outcomes and suggesting explanations for such differences that are unique to patent cases); cf Clermont & Schwab, Employment Discrimination Update, supra note 141, at 130-31 (finding judges seemingly less favorable than juries in another case category).
(216) See Theodore Eisenberg & James A. Henderson, Jr., Inside the Quiet Revolution in Products Liability, 39 UCLA L. REV. 731, 784-86 (1992); Theodore Eisenberg & Stewart J. Schwab, What Shapes Perceptions of the Federal Court System?, 56 U. CHI. L. REV. 501, 501-02 (1989); Marc Galanter, An Oil Strike in Hell: Contemporary Legends About the Civil Justice System, 40 ARIZ. L. REV. 717, 717 (1998); James A. Henderson, Jr. & Theodore Eisenberg, The Quiet Revolution in Products Liability: An Empirical Study of Legal Change, 37 UCLA L. REV. 479, 541 (1990). See generally THOMAS GILOVICH, How WE KNOW WHAT ISN'T SO (1991); RICHARD NISBETT & LEE ROSS, HUMAN INFERENCE 195--296 (1980).
(217) See Marc Galanter, The Civil Jury as Regulator of the Litigation Process, 1990 U. CHI. LEGAL F. 201, 227-51.
(218) See, e.g., supra Part I.B (using win rate data to unearth the forum effect). See generally Clermont & Eisenberg, supra note 19, at 587-92 (cautioning about difficulties inherent in using win rates to generate conclusions about how an underlying factor affects outcomes generally).
(219) See, e.g., Daniel Kessler et al., Explaining Deviations from the Fifty-Percent Rule: A Multimodal Approach to the Selection of Cases for Litigation, 25J. LEGAL STUD. 233, 235-36 (1996) ; George L. Priest & Benjamin Klein, The Selection of Disputes for Litigation, 13 J. LEGAL STUD. 1 (1984).
(220) See sources cited supra note 141, with results applied supra notes 190, 215, and infra text accompanying notes 236, 246.
(221) Time trends remain key, as the jobs category has seen a startling drop as a percentage of the docket every year after fiscal year 2001, so that in fiscal year 2006 it accounted for under six percent of the federal civil docket. The category has dropped in absolute number of terminations every year after fiscal year 1999. Clermont & Schwab, Employment Discrimination Update, supra note 141, at 104.
(222) See Clermont & Schwab, Employment Discrimination, supra note 141, at 432-38.
(223) See Clermont & Schwab, Employment Discrimination Update, supra note 141, at 127.
(224) See id. at 118, 131-32.
(225) Kevin M. Clermont & Theodore Eisenberg, Xenophilia in American Courts, 109 HARV. L. REV. 1120 (1996); Clermont & Eisenberg, Xenophilia II, supra note 215; see also Christopher A. Whytock, Myth of Mess? International Choice of Law in Action, 84 N.Y.U.L. REV. (forthcoming 2009) (manuscript at 37), available at http://ssrn.com/ abstract=1257096 (presenting results of an empirical study, which strongly suggests that choice-of-law decisions are not motivated by judicial biases in favor of domestic litigants).
(226) Clermont & Eisenberg, Xenophilia II, supra note 215, at 456-57.
(229) See Clermont & Eisenberg, supra note 225, at 1132-33.
(230) See id. at 1133-35.
(231) See, e.g., Clermont & Schwab, Employment Discrimination, supra note 141, at 444 fig.9 (showing the decline in plaintiff win rate by pretrial adjudication in all civil cases); Clermont & Eisenberg, Xenophilia II, supra note 215, at 456 & fig.2, 463 fig.5 (demonstrating downward time trends in diversity plaintiff win rate and in rate of diversity cases that end in judgment); supra Part I.A (discussing the upward time trend in removal rate); supra text accompanying notes 120-49 (noting the rise in summary judgment grants); supra Part IV.A (discussing the decline in trial rate).
(232) See, e.g., supra Part I.A (recognizing a sudden reversal in the trend of the remand rate); supra Part IV.A (describing the leveling off of the ratio of jury trials to judge trials); supra note 221 and accompanying text (noting the peak in number of employment discrimination cases).
(233) Clermont & Eisenberg, supra note 225, at 1125. Only since fiscal year 1986 and only in diversity cases have the AO codes indicated whether the principal parties were American or foreign.
(234) See Clermont & Eisenberg, Xenophilia II, supra note 215, at 456. Figure 5 comes from the second Xenophilia article, and here replaces a graph showing the steadiness of damage awards in federal civil trials over the years that appeared in Clermont & Eisenberg, supra note 1, at 148.
(235) Others have looked at just a sliver of the recent years and so have drawn shaky conclusions. See, e.g., Utpal Bhattacharya et al., The Home Court Advantage in International Corporate Litigation, 50 J.L. & ECON. 625, 650-53 (2007); see also Clermont & Eisenberg, Xenophilia II, supra note 215, at 450-51, 453-55 (discussing Bhattarcharya et al., supra).
(236) Only since fiscal year 1979 do the AO codes indicate which party prevailed by judgment in the district court. In any event, jobs was an insignificant case category in earlier years. The affirmance rate, which is the complement of the reversal rate, means the percentage of appeals that reach a decisive outcome and emerge as affirmed rather than reversed. I narrowly define "affirmed" as affirmed or dismissed on the merits. I define "reversed" as reversed, remanded, or modified, in part or completely.
(237) This high number is characteristic of appellate courts with a predominantly mandatory docket, such as the federal courts of appeals. See Theodore Eisenberg & Geoffrey P. Miller, Reversal, Dissent, and Variability in State Supreme Courts: The Centrality of Jurisdictional Source 2, 8, 15-23, 37-38 (N.Y.U. Law & Legal Theory Research Paper Series, Working Paper No. 08-01, 2008), available at http://ssrn.com/abstract= 1080563 (finding, by contrast, a 52% reversal rate for state appellate courts with discretionary jurisdiction, and attributing this to the selection effect of judges picking which cases to hear). Selection effect is more prevalent, and a bigger impediment to statistical analysis, at the level of the U.S. Supreme Court than for the federal courts of appeals. See Jonathan P. Kastellec & Jeffrey R. Lax, Case Selection and the Study of Judicial Politics, 5 J. EMPIRICAL LEGAL STUD. 407 (2008) (stressing problems in studying the Supreme Court).
(238) See Clermont & Eisenberg, Defendants' Advantage, supra note 215, at 130-34; Kevin M. Clermont & Theodore Eisenberg, Plaintiphobia in the Appellate Courts: Civil Rights Really Do Differ from Negotiable Instruments, 2002 U. ILL. L. REV. 947, 968-71.
(239) See Clermont & Eisenberg, supra note 205, at 1153-54.
(240) See also Chris Guthrie & Tracey E. George, The Futility of Appeal: Disciplinary Insights into the "Affirmance Effect" on the United States Courts of Appeals, 32 FLA. ST. U. L. REV. 357 (2005) (adding political science and psychology explanations of the tendency to affirm); cf Harlon Leigh Dalton, Taking the Right to Appeal (More or Less) Seriously, 95 YALE L.J. 62, 73-86 (1985) (questioning the worth of appellate courts as an error-correction device).
(241) See Clermont & Eisenberg, Defendants' Advantage, supra note 215, at 132 nn. 11-12; supra text accompanying note 219.
(242) See Clermont & Eisenberg, Defendants' Advantage, supra note 215, at 130-31, 154 (showing an appeal rate just over 20% for a selection of litigated judgments, and indicating 11.3% went all the way to affirmance or reversal); Clermont & Eisenberg, supra note 238, at 951-52, 967 (showing an appeal rate well under 20%, and indicating that 7.4% of all AO judgments go to affirmance or reversal). Both studies used data from fiscal years 1988-1997.
(243) Other evidence seems to confirm a limited effect of case selection on appeal. See, e.g., supra note 237. Most notably, a rich literature shows that appellate judges' attitudes (or ideologies) and other factors including case strength do influence success rates. See Jeff Yates & Elizabeth Coggins, The Intersection of Judicial Attitudes and Litigant Selection Theories: Explaining U.S. Supreme Court Decision Making, 27 WASH. U. J.L. & POL'Y (forthcoming 2009) (manuscript 7-13, 19-21, 30-31), available at http://ssrn.com/abstract=1120065. The role of attitudes would be hidden if case selection were robust on appeal.
For another example, see the state data from the National Center for State Courts, http://www.icpsr.umich.edu/cocoon/NACJD/STUDY/04539.xml, which indicate that the affirmance rate when a deferential standard of review governs is considerably higher than when a nondeferential standard governs. See also FRANK B. CROSS, DECISION MAKING IN THE U.S. COURTS OF APPEALS 49-53 (2007) (showing indirectly a similar result for the federal courts of appeals, while generally finding that case strength and judicial attitudes influence affirmance rates for those courts). That the standard of review should matter is not too surprising, one might respond. But if case selection were operating, the affirmance rates under different standards of review should tend to equate. Some evidence goes the other way, however. See Kessler et al., supra note 219, at 254, 256-57 (finding some selection effects on appeal).
(244) See supra note 242.
(245) See Steven Shavell, The Appeals Process as a Means of Error Correction, 24 J. LEGAL STUD. 379, 385, 421, 424 (1995) (suggesting a need for increased court fees on appeal). But see Scott Barclay, Posner's Economic Model and the Decision to Appeal, 19 JUST. SYS. J. 77, 95-96 (1997) (suggesting that taking an appeal is not an economic decision). Such reform would have the added benefit of lessening the workload of the appellate courts, a heavy workload having all sorts of deleterious effects on the appellate function. See Paul D. Carrington, The Function of the Civil Appeal: A Late-Century View, 38 S.C.L. REV. 411, 428-29 (1987).
(246) Kevin M. Clermont & Theodore Eisenberg, Anti-Plaintiff Bias in the Federal Appellate Courts, 84 JUDICATURE 128, 130 tbl.1 (2000); Kevin M. Clermont & Theodore Eisenberg, Judge Harry Edwards: A Case in Point!, 80 WASH. U. L.Q. 1275, 1283 (2002); Clermont & Eisenberg, Defendants' Advantage, supra note 215, at 135; Clermont & Schwab, Employment Discrimination, supra note 141, at 446-56; Clermont & Eisenberg, supra note 238, at 952 tbl.1. As to this effect, the state data tell a similar story. See Theodore Eisenberg & Michael Heise, Plaintiphobia in State Courts? An Empirical Study of State Court Trials on Appeal, 38 J. LEGAL STUD. 121 (2009).
(247) Clermont & Eisenberg, supra note 238, at 952 tbl.1.
(248) See Theodore Eisenberg, Appeal Rates and Outcomes in Tried and Nontried Cases: Further Exploration of Anti-Plaintiff Appellate Outcomes, 1 J. EMPIRICAL LEGAL STUD. 659, 677-82 (2004) (strengthening the counterargument by making strong assumptions, including an assumed but unrealistic selection effect on appeal).
(249) See id. at 682-85 (finding a residual attitudinal effect in the data even for the example of employment discrimination cases with their extremely low win rate). Appellate/trial court differences in attitude surely have an effect in certain types of cases. See, e.g., Timothy Davis Fox, Right Back "In Facie Curiae"--A Statistical Analysis of Appellate Affirmance Rates in Court-Initiated Attorney-Contempt Proceedings, 38 U. MEM. L. REV. 1, 2 (2007) ("The affirmance rate for the general appellate case population is in excess of 70%. The affirmance rate of the 932 court-initiated attorney-contempt [findings in Westlaw] cases included in this study is only about 32%.").
Kevin M. Clermont, Ziff Professor of Law, Cornell University. I want to thank Ted Eisenberg for his encouragement and help in my updating our previous work together, as well as thank Nicole Waters for her tremendous help with the new state data.
This Article is a revision of a piece that appeared in Empirical Studies of Judicial Systems 2008, published by Academia Sinica in Taiwan.