Early Settlement Based on Computer Data: A Case Study

In the Northern District of Ohio, Judge Thomas D. Lambros and Special Masters Eric Green and Francis McGovern have designed a program, called the Ohio Asbestos Litigation (OAL), that focuses on early settlement of cases before full litigation expenses, especially deposition costs, have been incurred.¹⁴⁰ Cases are grouped in clus­ters of five each, generally organized according to occupation, worksite, and disease. At least two groups of cases proceed through the OAL process simultaneously. This report devotes considerable attention to these innovative procedures. The intent is to provide an assessment of the procedures for the benefit of courts that are using or have used similar procedures.

The OAL program is premised on addressing both informational and communications needs of the parties. To meet the former, pre-trial exchange of information via interrogatories and document production is carefully structured. A special standardized question­naire-protocol was also developed that includes the amount and types of information that the lawyers have traditionally considered necessary for settlement purposes. Trial-oriented information, such as depositions of experts, is restricted to the postsettlement confer­ence stage. The OAL also addresses information needs by use of special masters and computers to generate data helpful for estimat­ing case values. Communication needs are addressed by two rou­tinely scheduled settlement conferences, 120 and 360 days after filing of the complaint. In those conferences the court uses its of-

1. See D. Waterman & M. Peterson, supra note 108, at 8.
2. For a description and evaluation of the program by one of the special mas­
   ters, see McGovern, supra note 37, at 478-91.

flees to break down any barriers to early discussion of settlement and participates directly in the evaluation of cases.

In the OAL, the second settlement status conference (SSC II) is the main event. In preparation for the conference, the parties submit the information protocol to the court on a confidential basis. The protocol covers more than 300 variables that the special masters determined to be related to the value of a case. The judge’s law clerk collects the protocols and serves as a clearinghouse for information among the participants and as manager of the details of the program.

Clerks feed these data into a computer, using rules of thumb for disputed facts (e.g., entering the facts alleged by the party with the burden of proof and flagging the item as “disputed”). These same data from over three hundred prior trials and settlements are stored in a comparative data base in the computer’s memory.¹⁴¹ Prior to the settlement conference, the court generates a computer printout that lists the names and final values of three closed cases that match most closely the case being examined. For each of the three cases, the computer printout lists the items (variables) that match and those that do not. One of the special masters reviews the printouts and talks with the judge’s law clerk and lawyers for the parties about special characteristics of the cases that may not be captured by the computer analysis.

At the conference,¹⁴² with the judge presiding in a courtroom, plaintiffs attorney presents a summary of the salient features of a case, comments on the evidence, and presents a settlement demand. Defendant’s counsel counter with their view of the evi­dence and their defenses, ending with a counterproposal. A special master then projects the computer values for the three cases on a screen, discusses the special characteristics of the case at hand, and recommends a settlement range. Because of the special characteris­tics, the range recommended by the special master generally dif­fers from the range generated by the computer for the three com­parison cases. Counsel for the parties then bargain to resolution or impasse on each individual case. After considering all of the indi­vidual cases in a group, if differences remain—and they always do—the parties bargain about aggregate settlement amounts. If an agreement is reached, the total amount is allocated among the plaintiffs by plaintiffs’ counsel and reported to the court for review and entry into the computer’s data base.

1. Id. at 487-88.
2. For a sample of the agenda at a conference, see Continuation of OhioAsbes­
   tos Litigation Case Clusters Groups I and II, Settlement/Status Conference II, Apr.
   9, 10, 1984 (on file with theFederalJudicialCenter).

Chapter VI

When the OAL was formulated in 1983, the court clustered the cases in two groups of five for each settlement conference. Cases were selected to include a mix of high-risk cases, such as mesotheli-oma victims with weak product identification evidence, and rou­tine, low-risk cases. The object was to give each some leeway for agreeing on a total amount for the group of cases despite differ­ences in individual evaluations of individual cases. As experience with the system increased, so did the size of the groupings: In the final settlement conference of Phase I of the OAL, fifty-six separate claims were settled.¹⁴³ In Phase II, a dramatically truncated sched­ule has been imposed, and a total of 150 claims will be prepared for a final settlement conference on April 20, 1987, and, if unresolved, a trial on June 1, 1987,¹⁴⁴

Overall, OAL Phase I led to the settlement of all 112 asbestos cases within twenty-seven months of the implementation of the plan.¹⁴⁸ Each of these settlements was within the range recom­mended by the special master. Evaluation of the plan by the spe­cial masters and the court is currently in process. A preliminary evaluation by one of the special masters contains “mixed reviews” on criteria of cost and fairness to the parties.¹*⁶

The clearest success of the OAL is that all cases did settle and that the parties did address the settlement of cases before the eve of trial, based on a truncated discovery process that was limited to information designed for evaluation, not trial. Traditional methods of case management do not produce such early settlements. The ab­breviated discovery schedule probably reduced the transactions costs to the parties more than the expenses of the OAL increased such costs. The benefit to the court is less clear. A primary benefit may be the ability of the trial judge to plan a trial calendar with­out the dislocations that last-minute settlements cause. Savings in judge or magistrate time in ruling on discovery disputes or pretrial motions might have been considerable in the early days of asbestos

143. In re Ohio Asbestos Litig., OAL Order No. 41 (N.D. Ohio Oct. 31, 1985). Some of these cases had apparently been discussed without settlement at earlier SSC Us. , 144. In re Ohio Asbestos Litig., OAL Order No. 56 & exhibit A (N.D. Ohio Jan. 16, 1987). The proposed format for the trial is to schedule ten simultaneous trials of groups of fifteen cases for each district judge. Id. at exhibit A, p.l. The time for full operation of the OAL system from designation of clusters to SSC II is 180 days (down from 360 in OAL I) and to trial, 260 days (down from 480). In re Ohio Asbes­tos Litig., OAL Order No. 48 (N.D. Ohio Aug. 8, 1986).

1. McGovern, supra note 78, at 489. The first groups of cases had been on a 480-
   day cycle from filing to trial. The final settlement was achieved at the SSC II, which
   was held at the 360-day point. Settlement of later clusters was accelerated in the
   last two SSC Us. See In re Ohio Asbestos Litig., OAL Order No. 41 (N.D. Ohio Oct.
   31, 1985).
   1. McGovern, supra note 3*7, at 489-93.

litigation, before these rulings became standardized. Further bene­fits may be derived from satisfaction of the court’s interests in pro­moting fair and just settlements.

A final assessment of the OAL cannot be made at this time. As Professor (and Special Master) McGovern indicates, the program should properly be seen as a “capital investment”¹⁴⁷ in the devel­opment of computer-based systems for settlement of mass tort cases. Recent caseload increases in the Northern District of Ohio will test the capacity of the system to operate efficiently and effec­tively in its home district.

Whether the program is transferable remains a serious question. Features of the plan, especially the computer-based evaluation of cases, have been exported to Eastern Texas,¹⁴⁸ Massachusetts,¹⁴⁹ and the claims facility to be created in the Manville bankruptcy. More sophisticated models remain to be tested.¹⁵⁰ Judge Lambros and the special masters have created a well-documented record of efforts to promote early settlements through the use of computers and other case management interventions. The Ohio results, how­ever, seem to be primarily the product of the dynamic force of the personality of the judge who initiated the system, operating syner-gistically with innovative and knowledgeable special masters.¹⁵¹

Viewed as a short-term, isolated venture, the OAL does not seem justified from the perspective of judicial case management, empha­sizing the use of court resources. All cases settled, but as table 6 indicates, trials in all of the study districts have been rare during

1. Id. at 489.
2. In Eastern Texas, a computer profile of cases was presented to the jury as
   jart of the class action proceedings. Had it proved to be necessary, the computer
   lata might have lent credibility to the plan to allow a single jury to assess punitive
   lamages for the class before compensatory damages were assessed for the class. The
   :hallenge was to devise a procedure that met the state-law requirement that puni-
   ;ive damages bear a reasonable relationship to compensatory damages for the class.
   Computer data would enable the jury to make an informed decision about how rep-
   -esentative those plaintiffs were and thereby to estimate compensatory damages for
   ;he class, the foundation for a punitive award. The Fifth Circuit approved this ap­
   proach as one of two acceptable alternatives. Jenkins, 782 P.2d at 468 (5th Cir. 1986).

An additional use of the computer data was in calculating the total settlement amounts. Lawyers reported that the computer data were useful in that respect.

1. Professor Eric Green submitted a report to the court and the Administrative
   Office of the U.S. Courts describing the utility of a computer system. E. Green, Mas­
   sachusetts Asbestos Litigation Feasibility Study (Aug. 22, 1986) (unpublished manu­
   script). He concluded that a computer-based case management and evaluation proc­
   ess is both feasible and desirable for the Massachusetts Asbestos Litigation. As of
   February 1987, the court had not obtained money or equipment to implement the
   report.
2. Brazil, Special Master in Complex Cases: Extending the Judiciary or Reshap­
   ing Adjudication?, 53 U. Chi. L. Rev. 394, 400-01 (1986).
3. Cf. Schuck, The Role of Judges in Settling Complex Cases: The Agent Orange
   Example, 53 U. Chi. L. Rev. 337, 359-85 (1986).

Chapter VI

1985 and 1986, As the data in table 9 indicate, the OAL did not serve to reduce the time from filing to termination.

TABLE 9

Time from Filing to Termination of Asbestos Cases in Ten Federal District Courts

These data include the start-up time for a plan such as OAL as well as the time accumulated prior to consolidation of the cases in Judge Lambros’s court. The first OAL order was issued in June 1983. The order concluding OAL Phase I was entered on October 31, 1985, approximately twenty-eight months (840 days) after the original consolidation. The data indicate that start-up time plus the 480-day trial track of the OAL surpasses the mean and median times for all asbestos cases in the ten districts. As modified, how­ever, the OAL should reduce the times dramatically.

As a capital investment, the OAL shows more promise. Among the lawyers interviewed inNorthern Ohio, several critics of the OAL still found merit in the use of computers to generate a range of prior settlements in similar cases. This support was not unquali­fied, and computer data were not seen as one of the main factors driving settlement. Indeed, one of the lawyers who found the com­puter data useful also was certain that lawyers could evaluate cases in their traditional way. The system seems acceptable, even valuable, to lawyers because it assembles information, using a sys­tematic mode that parallels traditional legal models of evaluating and settling cases. Lawyers look at precedents and then at facts that might distinguish those precedents from each other. In addi­tion, the system is attractive to lawyers because there is a built-in opportunity for a special master to adjust the range to account for less tangible variables that might affect the outcome, such as the

degree of impairment and suffering evident in plaintiffs appear­ance.¹⁵²

One might reasonably ask why a court should collect and orga­nize the data. A short answer is that only through the court is the information likely to become available to all parties. Absent a pro­gram like the OAL, plaintiffs’ attorneys are the only source of com­plete settlement information in a given case, but they do not neces­sarily share this information with each other. Likewise, defendants did not, prior to Wellington, generally share settlement informa­tion with each other. Jury reporter services are sporadic in cover­age of trials; settlement information is not systematically col­lected.¹⁵³ In that context, judicial initiatives are the only path to complete and relatively trustworthy information. In the end, the issue of whether courts should collect and distribute settlement in­formation depends on value judgments about the proper role of courts in society.

Computer data reportedly facilitated the settlement process by giving the parties a reasonable starting point for discussions,¹⁵⁴ in contrast to the “blue sky bargaining” that often characterizes the early stages of settlement discussions.¹⁵⁵ On the other hand, law­yers in other jurisdictions and in many cases are able to bargain through those barriers without the aid of data from the court. In this study, several private firms had created or were in the process of creating data bases for settlement. Wellington reportedly will create a national data base. Such experiences suggest that the fed­eral courts may serve a meaningful role in demonstrating to the parties a system that they may choose to adopt to improve their private dispute resolution capacities.¹⁵⁶

1. Query: Should the role of the special master—or the court, in other con­
   texts—be to predict likely jury outcomes or to employ a model of fair adjudication
   based solely on legal factors and excluding factors that should not, by themselves,
   influence the jury, such as the appearance or attractiveness of a plaintiff or the cor­
   porate status of a defendant?
2. Daniels, Civil Juries, Jury Verdict Reporters, and the Going Rate 5-9 (1986)
   (unpublished manuscript prepared for delivery at the 1986 Meeting of the Law &
   Soc’y Ass’n, May 29-June 1, 1986, Chicago, 111.; copy on file at the Federal Judicial
   Center).

Specialized litigation services, such as Mealey’s Litigation Reports: Asbestos and the Asbestos Litigation Reports, regularly report settlements and verdicts across the nation, but their reports include only a fraction of the universe of settlements.

1. Brazil, supra note 150, at 401 (quoting McGovern: ‘”the lawyers could not be
   too far apart and still be realistic and credible'”).
2. The phrase is from P. Gulliver, Disputes and Negotiations: A Cross-Cultural
   Perspective 137-38 (1979). The Arusha in Tanzania describe this stage of the negoti­
   ation process as “talking to the mountain.” Another commentator calls it “oratori­
   cal fireworks.” C. Karrass, The Negotiating Game 134-38 (1970).
   1. See generally D. M. Provine, supra note 130.

Chapter VI

In sum, the case for long-term judicial collection of the data has to rest on improving the quality of settlements and on the value of improving public access to data relating to litigation, both of which are difficult to test empirically. Justification for judicial involve­ment may hinge on whether public collection of, and access to, set­tlement information is superior to private collection, and retention, of that data.

Another factor in the evaluation of the computer system and the other features of the OAL is the question of cost. Professor McGov-ern states that the total cost of the OAL during the three years of Phase I was “over $250,000” for “special masters, experts, com­puter runs, and other expenses of designing and implementing the OAL plan.”¹⁸⁷ The cost per case would thus be over $2,200, distrib­uted among more than ten parties. None of the respondents initi­ated complaints about the cost and no appeal of the judge’s assess­ment of these costs was entered, despite some attorneys’ misgivings about the judge’s authority to impose these substantial costs on the parties. Parties might reasonably question the propriety of taxing them for capital outlays that primarily benefit litigants in other districts. Apparently, however, the cost savings per case through reduced transaction costs more than offset any party’s assessment of OAL development costs. Most defendants are also active litigants in the other districts in which the program will be used. Plaintiffs are likely to use the Manville facility, which probably will incorpo­rate some OAL features.

The price of the OAL is high and the parties pay much of the direct costs (e.g., fees of the special masters and other experts). Sav­ings in other litigation costs may offset these expenditures. In addi­tion to the direct costs of the OAL, there are indirect costs. Court personnel provide the support (such as data entry) and structure (such as the settlement conference and judicial planning) for the system through the judge and his clerks. Are these public expendi­tures justified by improvements in the process, such as freeing re­sources for other cases awaiting trial?

Evaluation of the OAL must also address the question of whether the process is fair to all parties. Professor McGovern reported that defendants and their attorneys “found the original OAL plan overly complex, the constant revisions disconcerting, and the court’s interventionist posture constraining their ability to repre­sent their clients as they saw fit.”¹⁵* This report’s findings are

1. McGovern, supra note 78, at 489. See also In re Ohio Asbestos Litig., OAL
   Orders Nos. 5, 7, 42, 45 (N.D. Ohio).
   1. McGovern, supra note 37, at 490.

similar. Some defendants’ attorneys have felt that their opportuni­ties for trial were abridged by the application of the OAL plan, es­pecially by what they perceived as shifting formats for trials, with settlement conferences scheduled during the time needed for final trial preparation. One attorney on the defense side termed the set­tlements “coerced” and felt that the procedures, especially the set­tlement conferences, were biased toward the plaintiffs’ interests and that the values of the cases were inflated.

Nor are all defendants entirely satisfied with the computer eval­uation process. Some see the “rule of thumb” for noting disputes (giving the benefit of the doubt to the party with the burden of proof) as being inadequate to cope with major disputes about diag­noses of diseases. Some also see the computer evaluation and the masters’ recommendations as giving the process a bias toward higher awards.

Plaintiffs and their counsel have adapted to the process more readily than defendants. Plaintiffs attend the final settlement con­ference and reportedly are totally satisfied that they have had their “day in court.” Conversely, defendants are currently resisting efforts to require their personal attendance at the conferences. Indeed, the process may provide maximum satisfaction to the plaintiff. Participation (through counsel) and a sense of control of the decision-making process is provided by mandating relatively public negotiations.¹⁸⁹ Use of computer precedents limits the risk of catastrophic loss to all parties by providing a range and a proc­ess for negotiating away extreme risks. Plaintiffs (and defendants, although perhaps more reluctantly) benefit from avoidance of the wide variations that are possible in the formal legal system.¹⁶⁰

1. See generally Thibault & Walker, Procedural Justice (1975). In their seminal
   study of adversarial versus inquisitorial systems, Thibault and Walker concluded
   that litigants preferred an adversarial system. One of two key factors contributing
   to increased satisfaction with adversary procedures was the “high degree of regu­
   lated contentiousness,” exemplified by the “separation of presentations, the designa­
   tion of opposing representatives, and the partisan identification of the attorney with
   . . , [the] client.” Id. at 119. The other major factor, which they considered to be of
   dominant importance, was “the maintenance of a high degree of control over its
   process by the disputants.” Id. See also Tyler, The Role of Perceived Injustice in De­
   fendants ‘ Evaluations of Their Courtroom Experience, 18 Law & Soc’y Rev. 51 (1984).
2. See, e.g., A. Chin & M. Peterson, Deep Pockets, Empty Pockets 11985), in
   which the authors report wide variations in reported jury verdicts in Cook County,
   111. This report, however, should be read in context. Awards in work injury cases
   involving corporate defendants were double the amount for similar injuries in
   injury-on-property (street hazard) cases. The authors indicate, however, that diffi­
   culty controlling for the variable of lost income seriously limits the explanatory
   value of their general conclusion that the jury system is the primary source of the
   variability. Failure to examine outcomes for bench trials results in a lack of com­
   parative data. Nevertheless, their data support the general impression, articulated
   primarily by counsel for corporate defendants, that “the tort system has become

Chapter VI

The diminution of the trial option, however, has fundamental ramifications for all parties, not just asbestos litigants. If the com­plaints that the trial options are effectively foreclosed by the system are, indeed, well founded, justification for computer evalua­tions and intensive settlement conferences would have to be enor­mous because the result would amount to a suspension of jury trial guarantees.

The crucial issue is how the OAL system compares to available alternatives. One commentator has questioned whether special masters may be subject to subtle role pressures to design new sys­tems and perhaps “overlook available resources and downplay the utility of adapting or refining established ways of solving prob­lems.”¹⁶¹ Are lawyers incapable of evaluating asbestos cases singly or in small groups? Evidence from other jurisdictions with moder­ate caseloads, such as Maryland, South Carolina, Eastern Tennes­see, and Western Pennsylvania, suggests, in fact, that lawyers re­solve asbestos claims promptly when they are called for trial. The Jenkins class action,¹⁶² with a settlement of approximately 755 cases during trial, suggests that an upper numerical limit to the ability of lawyers to evaluate and settle asbestos cases on the trial docket has yet to be reached.

Justification for the information generated by the computer de­pends on its utility in dealing with massive caseloads such as those experienced in the Districts of Massachusetts and Southern Missis­sippior in the multitudinous claims against A. H. Robins in its Chapter 11 proceedings,¹⁶³ At those levels of caseload, the value of

more nearly a crapshoot than an evenhanded compensation system.” Feinberg, The Toxic Tort Litigation Crisis: Conceptual Problems and Proposed Solutions, 24 Hous. L. Rev. 155, 163 (1987).

Such broadside critiques assume that the system has changed dramatically in recent decades and that other systems would produce more systematic, consistent results while remaining attentive to individual differences among litigants. Exami­nation of alternative systems, however, does not necessarily support that conclusion. In their classic study of judicial decision making in criminal sentencing, Partridge and Eldridge found considerable disparity among judges in an experimental context in which all respondents reviewed the same presentence report before making their (hypothetical) decisions. T. Partridge & W. Eldridge, The Second Circuit Sentencing Study (Federal Judicial Center 1974). See also Hensler, supra note 1, at 117-19 (alternative administrative systems are not encouraging).

1. Brazil, supra note 150, at 402.
2. Jenkins v. Raymark, No. M-84-193-CA (E.D. Tex, 1986). See also Texas Class
   Action Settles for More Than $100 Million, Mealey’s Litig. Reps.: Asbestos, Apr. 11,
   1986, at 4,089.
3. More than 327,000 claims have reportedly been filed with the bankruptcy
   court. Cooper, Robins Buy-Out Offer Withdrawn, Nat’l L.J., Feb. 23, 1987, at 3, 30.
   Professor Francis McGovern has also been appointed as a special master in these
   proceedings.