Full opinion text
GARWOOD, Circuit Judge: Before us are appeals and cross-appeals in personal injury and wrongful death damage suits against several manufacturers of asbestos-containing insulation products and some of their suppliers, the district court’s jurisdiction being based on diversity of citizenship and the governing substantive law being that of Texas. This is the same set of cases addressed in In re Fibreboard, 893 F.2d 706 (5th Cir.1990), but the judgments now before us result from a trial plan modified following that decision. Principally at issue on this appeal is the validity of that modified trial plan. The district court originally consolidated the some 3,031 such cases then pending in the Beaumont Division of the Eastern District of Texas for trial of certain common issues under Fed.R.Civ.P. 42(a) and also certified a class action under Fed.R.Civ.P. 23(b)(3), the class generally consisting of the insulation and construction workers, then-survivors and household members, who were plaintiffs in those pending cases. As explained in more detail below, the trial plan ultimately implemented after Fibreboard consisted of three phases, generally described as follows; Phase I comprised a complete jury trial of the entire individual eases of the ten class representatives and also a class-wide determination of issues of product defectiveness, warning, and punitive damages (including a multiplier as to each defendant). Phase II, which was to address exposure on a craft and job site basis, was dispensed with on the basis of a stipulation. In phase III, 160 different individual cases (“sample cases”), some from each of the five different allegedly asbestos-related diseases included in the entire group of underlying cases, were tried to two other juries to determine only each of those individual sample case plaintiffs’ respective actúal damages from their asbestos-related disease. Thereafter, and following a one-day bench hearing on the basis of which the district court determined that in each disease category the 160 sample cases were reliably representative of the cases involving the like disease among the remaining some 2,128 cases, the court ruled that each of these remaining 2,128 cases (the “extrapolation cases”) would be assigned by the court to one of the five disease categories and each would be entitled to judgment based on an amount of actual damages equal to the average of the verdicts rendered in those of the 160 sample cases involving the same disease category. Punitive damages in each case would be essentially based on the phase I verdict. By the time of the phase I trial, many of the defendants had settled and others had taken bankruptcy or otherwise been disposed of, so only five remained, namely appellant Pittsburgh Corning Corporation (Pittsburgh Corning), Carey Canada, Celotex, Fibre-board, and appellant Asbestos Corporation, Limited (ACL). The case against ACL was tried to the court under the Foreign Sovereign Immunities Act, 28 U.S.C. §§ 1330(a), 1603(b). By the time the amount of the extrapolation case judgments was to be calculated, all defendants except Pittsburgh Corning and ACL had passed out of the case. Judgment was entered against ACL in only two of the ten class representative cases (and in none of either the phase III sample cases or the extrapolation cases). Judgment was actually entered against Pittsburgh Corning in a total of 157 cases, consisting of 9 of the class representative phase I cases, 143 of the, phase III sample cases, and 5 of the extrapolation cases (1 from each of the 5 different diseases included in the class). In these 157 cases, Pittsburgh Corning has been cast in judgment for a total of approximately $69,000,000. Pittsburgh Corning and ACL each appeal the referenced judgments entered against them, and the plaintiffs cross-appeal as to each. The issues presented in the ACL appeal and cross-appeal are few and narrow, and we address them last. Pittsburgh Coming’s appeal presents essentially two groups of contentions, summarized as follows: first, those challenging the implemented Cimino trial plan as a whole, particularly its asserted failure to properly try and determine individual causation and, in the five extrapolation cases, damages also, as to any plaintiffs other than the class representatives, assertedly contrary to our decision in Fibreboard and Texas substantive law and in derogation of Pittsburgh Coming’s Seventh Amendment and Due Process rights; and second, various other issues of a more particular and traditional sort. Plaintiffs’ cross-appeal presents issues of only the latter variety. We now turn to consider Pittsburgh Coming’s appeal, addressing first its attacks on the trial plan. I. PITTSBURGH CORNING APPEAL A. Trial Plan Attack 1. Trial Plan Initial Plan The Cimino trial plan initially adopted by the district court, which we subsequently set aside in Fibreboard, also called for three phases. In phase I, the jury would decide which, if any, of each defendant’s products were defective as marketed and unreasonably dangerous, when each defendant knew or should have known workers or their household members were at risk, whether each defendant was guilty of gross negligence in marketing its offending product and, as to each defendant so guilty, a punitive damages multiplier. In phase II, the same jury would' decide the percentage of plaintiffs in the class exposed to each defendant’s products, the percentage of claims barred by limitations and other defenses, and would determine a lump sum amount of actual damages for each disease category for all plaintiffs in the class. The jury in this phase would also make a full determination of liability and damages with respect to each of the eleven class representatives individually. And the jury in phase II would also hear such evidence as the parties desired to present from up to thirty other illustrative plaintiffs, fifteen chosen by the defense and fifteen by plaintiffs, as well as expert testimony regarding the " total actual damages of the class, such expert testimony to be based, among other things, on questionnaires filled out by all class members and other discovery, including forty-five-minute oral depositions of class members taken by defendants. In phase III, to be non-jury, the court would distribute the awarded damages among the individual class members. Fibreboard In Fibreboard, we found “no impediment to the trial of Phase I,” id. at 712, but held the balance of the plan invalid, stating: “It infringes upon the dictates of Erie that we remain faithful to the law of Texas, and upon the separation of powers between the judicial and legislative branches. “Texas has made its policy choices in defining the duty owed by manufacturers and suppliers of products to consumers. These choices are reflected in the requirement that a plaintiff prove both causation and damage. In Texas, it is a ‘fundamental principle of traditional products liability law ... that the plaintiffs must prove that the defendant supplied the product which caused the injury.’ These elements focus upon individuals, not groups. The same may be said, and with even greater confidence, of wage losses, pain and suffering, and other elements of compensation. These requirements: of proof define the duty of the manufacturers. ... The inescapable fact is that the individual claims of 2,990 persons will not be presented. Rather, the claim of a unit of 2,990 persons will be presented. ... That procedure cannot focus upon such issues as individual 'causation, but ultimately must accept general causation as sufficient, contrary to Texas law. It is evident that these statistical estimates deal only with general causation, for ‘popula.tion-based probability estimates do not speak to a probability of causation in any one case; the estimate of relatiye risk is a property of the studied population, not of an individual’s case.’ This type of procedure does not allow proof that a particular defendant’s asbestos ‘really’ caused a particular plaintiff’s disease; the only ‘fact’ that can be proved is that in most cases the defendant’s' asbestos would have been the cause.” Id. at 711-712 (footnotes omitted; emphasis added except in interior quotation and in last clause). ! Present Plan Following this Court’s decision in Fibre-board, the district court initially determined that “[t]his case will now proceed under the procedures set out in Jenkins v. Ray-mark” — i.e. phase I to be followed by a series of mini-trials for all plaintiffs on their individual causation and damage issues (see note 8, supra) — and set its previously adopted phase I (which we had declined to block) for trial. The court observed that its “task appears to be insurmountable,” but stated that it would nonetheless “take[] its place behind the old mule and start down that long row.” Some months later, however, the court changed its mind and adopted the trial plan now before us (except that a stipulation was ultimately utilized instead of phase II), observing: “Phase One will leave unresolved the questions of exposure, comparative causation, and damages. These remaining questions could easily be resolved by.the procedure established in Jenkins if the numbers were manageable. The numbers are not manageable. Jenkins envisioned groupings of ten plaintiffs submitted to a succession of juries. If we could try one group a week, the process would take years. Additional judicial power and the utilization of multiple courtrooms could shorten the time to resolve all these cases, but it would not decrease total court time or attorney time. Transaction costs to the parties under the Jenkins procedure is unacceptable.” Instead of utilizing the Jenkins procedure, the court determined to employ new phases II and III: “asking the jury in Phase Two to make findings on exposure that are specific to job site, craft and time; and then by submitting to a jury in Phase Three individual damage cases of a statistically significant, randomly selected sample from each of the five disease categories.” For purposes of phase II, twenty-two different worksites— principally refineries, shipyards, and chemical plants, and also including other industrial-type facilities and a power plant — in Beaumont, Port Arthur, Orange, and Port Neches, Texas, and including two sites in Lake Charles, Louisiana, would be considered. The district court contemplated that the phase II jury (the same jury as in phase I) would: “hear evidence concerning: (a) the presence of. the Defendants’ products at the worksites; (b) the presence of asbestos dust at the worksites; and (c) the nature of the various crafts at the worksites and the relationship between these crafts and the presence of asbestos dust at these facilities. Specifically, the jury will hear evidence concerning the working conditions of machinists, pipefitters, insulators, carpenters, etc. and the relationship between these workers and the Defendants’ asbestos products. The jury will make a determination as to which crafts at the work-sites were exposed to which Defendants’ asbestos products (if any) for a sufficient period of time to cause injury, harm, or disease. The Court will make a non-jury determination as to which Plaintiffs or Plaintiffs’ decedents worked for a sufficient period of time at each worksite so as to be a proper member of that worksite’s group and which Plaintiffs were proper members of each of the crafts at these worksites.... The Court will submit the issue of exposure to the jury pursuant to ten-year intervals. So, for example, the jury will be asked whether the product(s) of Defendant X were present at Worksite Y during the 1940’s, the 50’s, 60’s, etc. And, for example, the jury will be asked whether the carpenters at Worksite Y were exposed to Defendant X’s product(s) during the 1940’s, the 50’s, 60’s, etc. During Phase Two, the jury will apportion responsibility among settling and non-settling Defendants for the Plaintiffs’ exposure (if any).” In Phase III, two other juries would determine for 160 sample cases only “two damage issues,” namely “(e) whether the Plaintiffs suffered from an asbestos-related injury or disease and, if so, (b) what damages the Plaintiffs incurred.” The court ultimately determined, based on information from plaintiffs, that the entire class of 2,298 eases could be broken down into the 5 disease categories, and the court then randomly selected 160 sample eases, some from each disease category, as follows: Disease Number of Number of Sample Cases Cases in Class Mesothelioma 15 32 Lung Cancer 25 186 Other Cancers 20 58 Asbestosis 50 1,050 Pleural Disease 50 972 Total 160 2,298 Individual judgment would be entered in each of the 160 sample cases based on the phase III verdict in that particular sámple case. After phase III, the district court would assign each of the remaining 2,298 cases to one of the 5 disease categories, and in each case make an award of actual damages equal to the average of the awards in the phase III cases involving the same disease. Phase I The phase I trial lasted approximately eight weeks. The defendants then remaining were Carey. Canada, Celotex, Fibreboard; and Pittsburgh Corning. The jury found in answer to the first four questions when the defendants knew or should have known that their “asbestos-containing insulation products” posed a risk of asbestos-related disease to “insulators” (question 1), to their household members, to other “crafts working with or near insulation products,” and to their household members. Pittsburgh Corning knew or should have known this since 1962 (when it first entered the business; it left it in 1972) as to both insulators and other crafts; the other three defendants since 1935 as to insulators and since 1955 as to other crafts; all four defendants as to both sets of household members since 1965. In answer to question 5, the jury found that, since 1962 as to Pittsburgh Corning and since 1935 as to the other defendants, the defendants’ listed insulation products “were defective and unreasonably dangerous as a result of not having an adequate warning.” The district court ultimately disregarded the answers to questions 2, 3, and 4, which addressed knowledge concerning other crafts and household members, and ordered judgment rendered on the basis of question 1, knowledge concerning insulators, and question 5, failure to warn. In question 7, the jury found each defendant guilty of gross negligence warranting punitive damages and • assigned a punitive damages multiplier of $3.00 per $1.00 of actual damages to Pittsburgh Corning, $2.00 to Celotex, and $1.50 each to Fibreboard and Carey Canada. Questions 8 through 17 separately addressed the individual case of each of the 10 class representatives. In each respective question, the jury was asked to find for the particular plaintiff or the defendants, and if for the plaintiff to find separate dollar amounts of past and of future damages for that plaintiff, and to “apportion causation” (in percentages. .totaling one hundred percent) among that plaintiff, some or all of the then current defendants, and some or all of the dismissed former defendants. In three of the cases, the plaintiffs causation was not submitted (in one of these the verdict was for the defendants, and a new trial was granted), in another three such causation was submitted but not found, and in four cases plaintiff causation was found (15%, 17%, 20%, and 50%). In each of the 9 cases in which the jury found for the plaintiff, Pittsburgh Coming’s causation was fixed at 20%; Fibreboard and Celotex were each assessed 15% in 8 of these cases, and in one case Celotex was assessed 30% and Fibreboard none; in the only 4 of these cases in which Carey Canada’s causation was submitted, it was found to be 15%. In each of these 9 cases, the causation of each of some 10 former defendants was submitted, separately for each, and it was found in each case in amounts ranging from as little as a total of 10% for all of them to as much as 50% for all. The jury’s phase I actual damage findings totaled some $3.5 million. Phase III Following completion of the phase I trial (and a continuance), the district court proceeded directly into phase III, without any phase II trial. It was not until approximately seven weeks into the phase III trials that the stipulation — which ultimately replaced phase II — was entered into. It was clear from the beginning of, and throughout, the phase III trials that the two juries were not to, and did not, determine whether exposure to any of defendants’ products was a cause of the sample plaintiffs’ complained-of condition. In phase III the court instructed the jury that they were to assume exposure was sufficient to be a producing cause of all the disease categories. As plaintiffs admit in their brief here, in the phase III trial “the juries were told to assume that the claimants had sufficient exposure.” Indeed, for the most part evidence of exposure and its likely or possible results was not allowed. Simply stated, whether there was exposure to Pittsburgh Coming’s — or any other defendant’s— asbestos, and, if so, whether that exposure was a cause of any of the 160 sample plaintiffs’ illness, disease, or damages, was neither litigated nor determined in any of the phase III trials. Nor were any matters concerning any individual sample plaintiffs past connection with any particular worksite or craft either litigated or determined in phase III (although some miscellaneous information in this regard was not infrequently incidentally reflected in general background or work history testimony). Following the phase III jury verdicts (including 12 zero verdicts) in the 160 sample cases, the district court ordered remittiturs in 35 of these cases (“34 of the pulmonary and pleural cases and in one mesothelioma case”), and calculated the average actual damage award, after remittitur (and considering the zero verdicts), in each disease category to be the following: mesothelioma, $1,224,333; lung cancer, $545,200; other cancer, $917,785; asbestosis, $543, 783; pleural disease, $558,900. These were the figures to be applied to the extrapolation cases. Phase II stipulation We now turn to the written stipulation— entered into after some seven weeks of the phase III trials had taken place — which replaced phase II. It was executed by all the plaintiffs and by Pittsburgh Corning, Fibre-board, and Celotex, who constituted all the then-remaining defendants (except ACL, whose case was non-jury), and was approved “so ordered” by the district court. Attached to the stipulation as an exhibit was a special verdict form that would consist of separate interrogatories, each with a part (a) and a part (b), one each for each of the twenty-two worksites at issue. For example, question 1(a) would ask “For Worksite No. 1, do you find that the following crafts had sufficient exposure to asbestos during the specified time periods to be a producing cause of the disease of asbestosis.” The jury would answer yes or no separately as to each of over fifty listed crafts for each of four specified decades, namely 1942-52, 1952-62, 1962-72, and 1972-82. Question 1(b) would state, “For the crafts and the time periods which were answered ‘yes’ bi question 1(a), causation is apportioned as follows.” This question would be answered by stating separately for each listed craft a percentage applicable to each of the current defendants and each of the former defendants who had settled as to each of the same four decades (as to each decade the percentages were to total one hundred percent). This process would be repeated, with questions 2(a) and 2(b), 3(a) and 3(b), and so forth, separately as to each of the remaining worksites. The stipulation provides in part that: “(3) It is stipulated that some individuals working in the listed crafts ... at the 22 Phase Two worksites during each decade from 1942 to 1982 were exposed to asbestos during the course of their employment. The exposure of some members of each of the crafts ... at the 22 worksites was of sufficient length and intensity to cause pulmonary asbestosis of varying degrees. Asbestos-containing products of predecessors to the Celotex Corporation and Fibreboard Corporation were present during each decade in the specified worksites. An asbestos-containing product of Pittsburgh Corning Corporation was present during the decades 1962-1982 at the specified worksites. The defendants do not stipulate that any members of the various crafts at the various worksites had the same exposure to any products or that any such individuals had the same susceptibility to asbestos-related diseases in the various crafts and worksites.” (Emphasis added). The stipulation further provides that, although “[i]f the Court were to proceed with ‘Phase Two’ ... [i]t is stipulated for purpose of appellate review that the [phase II] jury’s verdicts would assign different [causation] percentages to each” of the defendants Pittsburgh Corning, Fibreboard, and Celotex, and “would assign different percentages with respect to each Phase Two worksite ... craft ... and decade combinations” submitted, nevertheless “[defendants] stipulate it shall be deemed that the Phase Two jury” assigned in all instances the following comparative causation shares, viz: Pittsburgh Corning, ten percent; Fibreboard, ten percent; Celotex, ten percent; and Manville Personal Injury Settlement Trust, thirteen percent. The court would use these stated percentages to fashion judgments in the 160 phase III sample cases and in the extrapolation cases. Before setting out these percentages, however, the stipulation had made clear that defendants were not thereby agreeing that the trial plan — either the originally planned phase II or the contemplated extrapolation procedure — was a permissible way to adjudicate their liability and damages. Thus, it stated: “This stipulation relates to the percentage findings to be supplied through the Court’s special verdict form which the Court intends to apply to individuals pursuant to the Cimino trial management plan, to which these defendants object. If the reviewing courts reject determination of individual legal causation issues by resort to general Phase Two worksite/craft findings, or reject the use of Rule 23 class trials for asbestos injury cases, the Phase Two share percentage findings specified below are void.” (Emphasis added). Defendants’ reservations of their objections in this respect are also reflected in later passages of the stipulation. In paragraph 5 it is stated that “Defendants continue to object to these extrapolation procedures,” and paragraph 8 states: “Defendants reserve all rights to object to all past and future aspects of the Cimino trial plan and to assign as error all prior, present, and future rulings of the Court, except only that Defendants shall not assert that the evidence is or would be insufficient to support a 10% finding (as compared, e.g., to a 5% finding, etc.) with respect to any particular Phase Two job-site and craft combination.” And, the stipulation recites that defendants specifically reserved, and would be afforded, the right to contend on appeal the following (among other things): “that it is impermissible to determine medical or other causal responsibility on a jobsite or craft-wide basis; that it is impermissible to establish a single period of time sufficient to cause asbestos related disease, injury or harm except in connection with evidence presented in regard to an individual and as applied to that individual; that it is impermissible to use decades of exposure to asbestos, worksite or employment status to assess individual exposure or medical causation issues; and that it would be impermissible under governing law to assign a single percentage of ‘causation’ or ‘responsibility’ to a particular craft or job classification.” Paragraph 12 of the stipulation confirms its limited nature, viz: “(12) Without limitation, Defendants do not stipulate that: entry of any judgment based on actual or stipulated Phase Two ■findings is legally or factually sound; any Defendant in fact has legal responsibility to any individual plaintiff; any individual plaintiff was in fact exposed to injurious quantities of asbestos from the products of any Defendant; the products of any of the Defendants were in fact legal causes of injury to any individual plaintiff; or that any issue framed by the Cimino pleadings can be adjudicated on a jobsite or craft-uñde basis. Defendants have not stipulated or agreed that evidence to be received under the Cimino trial management plan is or coidd be sufficient to establish in these cases that any class member plaintiff suffers from an asbestos-related disease (except as previously stipulated on the record in particular cases), or that the asbestos-containing product or products of any defendant caused or contributed to any such disease, nor that a finding of responsibility or causation in any percentage with respect to a defendant and any class member is or could be sustained by evidence limited to asbestos-related disease among, or exposure to asbestos of, members of specified crafts at specified work-sites over ten-year periods of tim.e in the absence of evidence sufficient to show that each plaintiff class member to whom a defendant is held liable in any percent himself or herself has an asbestos-related disease and that such class member was exposed to the defendant’s asbestos product or products in quantities and for times sufficient to cause such disease. Further, defendants have not stipulated to the sufficiency of any evidence which would permit any finding by the Court or jury that any class member plaintiff has been damaged in any sum or amount by reference or resort to damages suffered by any other plaintiff, or groups of plaintiffs, in the absence of evidence specifically showing damage suffered by such plaintiff class member himself or herself individually.” (Emphasis added). Finally, the stipulation reflects that the court, by its approval thereof, had ruled, and “would have adhered to such ruling throughout the trial” and “will adhere to this ruling in reviewing offers of proof’ mentioned in the stipulation, that, with presently immaterial exceptions, “... it would not submit to the jury for a verdict (or receive individual evidence for individual adjudication) as to each plaintiff class member except where it has done so in proceedings to date, several issues, including: whether he or she was exposed to an asbestos-containing product; whether that exposure was sufficient to cause injury; the identity of those who manufactured the products to which such each plaintiff was exposed; and the individual damages suffered by such person as a result of exposure.” After the stipulation, the phase III trials continued for approximately five more weeks, conducted in all material respects on the same basis and in the same manner as they had been during the some seven weeks before the stipulation was entered into. Extrapolation The final phase was that of extrapolation. About a month after completion of the phase III trials, a one-day non-jury hearing was held in which the district court heard evidence concerning the degree to which the 160 sample cases were representative, in their respective disease categories, of the cases in the same disease category among the 2,128 extrapolation cases. Essentially the only evidence at this hearing was the testimony of three expert witnesses called by the plaintiffs, namely Dr. John Dement, Director, Office of Occupation Health and Technical Services, National Institute of Environmental Health Sciences; Professor Ronald Frankew-itz of the University of Houston, a Ph.D. in Evaluation, Measurement, and Statistics; and University of Texas Law School Professor of Trial Practice Patrick Hazel, an experienced personal injury trial lawyer. The district court’s opinion dealing with extrapolation does not refer, either generically or specifically, to any evidence other than Professor Frankewitz’s testimony. He stated that he was furnished by someone in the offices of plaintiffs’ counsel computerized written data reflecting, as to each of the 160 sample eases and each of the 2,128 extrapolation cases, whether the case was a sample case or an extrapolation case, which of the 5 disease categories the case involved, and an answer to each of 12 specific variables pertaining to the particular plaintiff or plaintiffs decedent alleged injury to whom formed the basis of the suit. The 12 variable were gender, race, whether living, whether ever smoked, whether was a wage earner (when not specified), age, first year of exposure, last year of exposure, total years of exposure, latency, pack years smoked, trade and pre-' dominant craft. Professor Frankewitz testified that the sample cases in each of the five disease categories were representative of the extrapolation cases in the same disease category “in terms of the variables that I’ve analyzed,” so that, for example, if one were to randomly select another 50 asbestosis cases from the 2,128 extrapolation cases, 99 out of 100 times (98 out of 100 in two minor respects) those 50 eases would have “the same mix of variables” as the 50 asbestosis cases which were a part of the 160 sample phase III cases. Dr. Frankewitz did not select the variables, nor did he determine what those variables were in any of the cases; rather‘he was simply furnished that information by plaintiffs’ counsel’s office. Similarly, he made no independent judgment as to which disease category any case fit in, but simply was furnished that conclusion by the office of plaintiffs’ attorneys. And Dr. Frankewitz was even not sure just what some of the variables meant. When asked what the variable “total years of exposure” meant, he replied “As far as I’m concerned, I believe it’s ... I’d be guessing. I would say it’s the number of years that an individual was exposed to asbestos in a particular setting, particular situation” (earlier in his testimony he had indicated that it was “a function of’ first and last years of exposure). He did not calculate “total years of exposure” and when asked who did, said “My belief would be it would be a clerk under the supervision or direction of one of the plaintiffs’ attorneys.” The district court concluded “that the distribution of variables between the samples and their respective subclasses is comparable.” 2. Analysis As noted, Pittsburgh Corning attacks the Cimino trial plan, as it did at all times below, principally on the basis that it fails to properly try and determine individual causation, and in the extrapolation cases also fails to properly try and determine individual damages, as to any plaintiffs other than the ten class representatives whose individual cases were fully tried in phase I. Pittsburgh Corning asserts in this connection, among other things, that these aspects of the trial plan are contrary to Fibreboard, impose liability and damages where they would not be imposed under Texas substantive law, and invade its Seventh Amendment and due process rights. Although we do not separately address the due process contention as such, we conclude that the Cimino trial plan is invalid in these respects, necessitating reversal of all the phase III sample case judgments as well as the five extrapolation case judgments before us. We begin by stating some very basic propositions. These personal injury tort actions for monetary damages are “a prototypical example of an action at law, to which the Seventh Amendment applies.” Wooddell v. Intern,. Broth. of Elec. Workers, 502 U.S. 93, 112 S.Ct. 494, 498, 116 L.Ed.2d 419 (1991). The Seventh Amendment applies notwithstanding that these are diversity cases. Simler v. Conner, 372 U.S. 221, 83 S.Ct. 609, 9 L.Ed.2d 691 (1963). See also Gasperini v. Center for Humanities, Inc., 518 U.S. 415, 116 S.Ct. 2211, 135 L.Ed.2d 659 (1996). But because these are diversity eases, the Rules of Decision Act, 28 U.S.C. § 1652, and Erie R. Co. v. Tompkins, 304 U.S. 64, 58 S.Ct. 817, 822-23, 82 L.Ed. 1188 (1938), with its seeming constitutional underpinning, mandate that the substantive law applied be that of the relevant state, here Texas. Substantive law includes not only the factual elements which must be found to impose liability and fix damages, but also the burdens of going forward with evidence and of persuasion thereon. Palmer v. Hoffman, 318 U.S. 109, 63 S.Ct. 477, 482, 87 L.Ed. 645 (1943); Cities Service Oil Co. v. Dunlap, 308 U.S. 208, 60 S.Ct. 201, 84 L.Ed. 196 (1939). None of the foregoing is or can be altered by the utilization of Fed.R.Civ.P. 23(b)(3) or Fed.R.Civ.P. 42(a). As to the Seventh Amendment, the Court in Ross v. Bernhard, 396 U.S. 531, 90 S.Ct. 733, 24 L.Ed.2d 729 (1970), held that in a stockholders’ derivative action seeking monetary relief-now provided for in Fed.R.Civ.P. 23.1— although the right of the stockholders to sue on behalf of the corporation was an equitable matter determinable by the court, the monetary claims of the corporation against the defendants were legal claims to which the Seventh Amendment applied. The Court observed that “The Seventh Amendment question depends on the nature of the issue to be tried rather than the character of the overall action,” 'id. at 738, and “nothing turns now upon the form of the action or the procedural devices by which the parties happen to come before the court.” Id. at 739. It also noted that it was “inclined to agree with the description” of derivative suits “as one kind of ‘true’ class action,” and that “it now seems settled in the lower federal courts that class action plaintiffs may obtain a jury trial on any legal issues they present.” Id. A leading text gives the following commentary on Ross: “The language just quoted, that nothing turns on ‘the procedural devices by which the parties happen to come before the court,’ makes the Ross ease controlling not only for derivative actions but also for the other procedural devices that the Civil Rules borrowed from equity. In all of these it will be for the judge to decide whether the device may be used, but once he or she does so there will be a right to jury trial on any of the underlying issues that- are legal in nature. Indeed, the Ross decision itself relied in part on lower court decisions reaching this result with regard to class actions under Rule 23. The Court said that ‘it now seems settled in the lower federal courts that class action plaintiffs may obtain a jury trial on any legal issues they present,’ and indicated its agreement with the view that derivative suits are one kind of ‘true’ class action.” 9 Wright & Miller, Federal Practice and Procedure, § 2307 at 79 (footnotes omitted). And, this Court has long held that the applicability of the Seventh Amendment is not altered simply because the case is Rule 23(b)(3) class action. State of Alabama v. Blue Bird Body Co., Inc., 573 F.2d 309, 318 (5th Cir.1978). Similarly, use of Rule 23(b)(3) or 42(a) does not alter the required elements which must be found to impose liability and fix damages (or the burden of proof thereon) or the identity of the substantive law — here that of Texas — -which determines such elements. We squarely so held in Fibreboard. And the rules enabling act, 28 U.S.C. § 2072 likewise mandates that conclusion. As we said in Blue Bird Body Co.: “This Circuit has also explained that the meaning of liability for antitrust purposes does not change simply because a trial is bifurcated under Fed.R.Civ.P. 42(b). In Response of Carolina, Inc. v. Leasco Response, Inc., 537 F.2d 1307 (5th Cir.1976), this court stated that there was ‘no basis in law or logic to give liability different meanings depending upon the trial procedure used.’ Id. at 1321. The Leasco opinion explained that bifurcation in no way diminishes the requirement that a plaintiff show some evidence that a violation caused him injury before a defendant is found liable. Just as the meaning of liability does not vary because a trial is bifurcated, the requisite proof also in no way hinges upon whether or not the action is brought on behalf of a class under Rule 28. It is axiomatic that a procedural rule cannot ‘abridge, enlarge, or modify any substantive right.’ [citing 28 U.S.C. § 2072] Consequently, this court has no power to define differently the substantive right of individual plaintiffs as compared to class plaintiffs.” Id. at 317-318 (footnote omitted; emphasis added). Nor is deviation from these settled principles authorized because these are asbestos cases whose vast numbers swamp the courts. Fibreboard clearly so holds. So, also, in Jackson v. Johns-Manville Sales Corp., 750 F.2d 1314 (5th Cir.1985), cert. denied, 478 U.S. 1022, 106 S.Ct. 3339, 92 L.Ed.2d 743 (1986), a diversity asbestos case arising in Mississippi, we declined to adopt a federal common law rule for asbestos cases (or to certify to the United States Supreme Court whether to do so), stating: “... [U]nder our federal system Congress is generally the body responsible for balancing competing interests and setting national policy. There is no doubt that a desperate need exists for federal legislation in the field of asbestos litigation. Congress’ silence on the matter, however, hardly authorizes the federal judiciary to assume for itself the responsibility for formulating what essentially are legislative solutions. Displacement of state law is primarily a decision for Congress, and Congress has yet to act....” Id. at 1327. When, after Fibreboard, the district court adopted the present trial plan, it initially justified doing so on the basis of its conclusion that “the Texas Supreme Court, if faced with the facts of this case, would apply a collective liability theoiy, such as the Court’s plan, to an asbestos consolidated action.” The court based this conclusion on a passage in Gaulding v. Celotex Corp., 772 S.W.2d 66, 71 (Tex.1989), stating “We are not to be construed as approving or disapproving alternative liability, concert of action, enterprise liability, or market share liability in an appropriate case.” We are compelled to reject the district court’s conclusion for each of several independently sufficient reasons. To begin with, it is contrary to Fibreboard, which plainly holds that under Texas substantive law causation of plaintiffs injury by defendant’s product and plaintiffs resultant damages must be determined as to “individuals, not groups.” Fibreboard’s determination of Texas law is precedent which binds this panel. See, e.g., F.D.I.C v. Abraham, 137 F.3d 264, 268-69 (5th Cir.1998); Broussard v. Southern Pacific Transportation Company, 66.5 F.2d 1387, 1389 (5th Cir.1982) (en banc). Gaulding furnishes no basis to depart form Fibreboard because it was quoted and relied on therein. Fibreboard at 711, n. 4. No Texas appellate decision or statute subsequent to Fibreboard casts doubt on the correctness of its reading of Texas law. In the second place, even were we not bound by Fibreboard we would reach the same conclusion it did, namely that under Texas personal injury products liability law causation and damages are determined respecting plaintiffs as “individuals, not groups.” We know of no Texas appellate decision which in that or a similar context has even approved of in dicta, much less adopted, the theories of'“alternative liability, concert of action, enterprise liability, or market share liability” which Gaulding states it was not “approving or disapproving.” Id. at 71. “We have long followed the principle that we will not create ‘innovative theories of recovery or defense’ under local law, but will rather merely apply it ‘as it currently exists.’ ” Johnson v. Sawyer, 47 F.3d 716, 726 (5th Cir.1995) (en banc) (citations omitted). Consistent with that principle, we have on more than one occasion expressly refused to hold that Louisiana would apply a market share liability theory to asbestos personal injury claims, where no Louisiana appellate decision had either done so or declined to do so. Thompson v. Johns-Manville Sales Corp., 714 F.2d 581, 583 (5th Cir.1983) (refusing to hold that Louisiana would adopt either “enterprise” or “market share” liability; noting “[b]oth theories represent radical departures from traditional theories of tort liability” and “[sjuch departures are for the Louisiana courts, not for us”); Bateman v. Johns-Manville Sales Corp., 781 F.2d 1132, 1133 (5th Cir.1986) (market share liability). See also Jefferson v. Lead Industries Ass’n, Inc., 106 F.3d 1245 (5th Cir.1997) (declining to adopt market share liability in Louisiana diversity suit for lead paint poisoning); Rhynes v. Branick Mfg. Corp., 629 F.2d 409 (5th Cir.1980) (declining to adopt “product line” liability theory in Texas diversity case). We apply Texas law as it currently exits, which is correctly stated in Fibreboard. Finally, it is clear that this case was neither tried nor determined on any of “the collective liability theories” mentioned in Gaulding. See id. at 71. Thus, the question becomes: did the implemented trial plan include a litigated determination, consistent with the Seventh Amendment, of the Texas-law mandated issues of whether, as to each individual plaintiff, Pittsburgh Coming’s product was a cause of his complained-of condition and, if so, the damages that plaintiff suffered as a result. We turn first to the phase III plaintiffs. In these eases, the trial plan was adequately individualized and preserved Seventh Amendment rights with respect to each individual’s actual damages from an asbestos-related disease. However, it was not designed or intended to, and did not, provide any trial or any determination of whether a Pittsburgh Corning product was a cause of that disease. It was strictly a damages trial as to those individual plaintiffs. The stipulation — not entered into until midway through phase III — established merely that “some” individuals working in each of the listed crafts, “during” each of the four decades 1942-1982 and at each of the twenty-two worksites, “were exposed to asbestos” with “sufficient length and intensity to cause pulmonary asbestosis of varying degrees” and that “an asbestos-containing product of Pittsburgh Corning Corporation was present during the decades 1962-1982 at the specified worksites.” It was expmssly not stipulated “that any members of the various crafts at the various worksites had the same exposure to any products,” or “that any such individuals had the same susceptibility to asbestos-related diseases in the various crafts and worksites,” or that “any individual plaintiff was in fact exposed to injurious quantities of asbestos from the products of any defendant.” Phase III did not litigate or determine whether or to what extent any of the one hundred sixty individual plaintiffs was exposed to Pittsburgh Coming’s — or any other defendant’s — asbestos, or was exposed to asbestos at any of the twenty-two work-sites, or whether any such exposure was in fact a cause of that plaintiffs illness or disease. Nor did phase III litigate or determine either any individual plaintiffs past connection with any particular worksite o'r craft, or whether or to what extent such individual was exposed to asbestos 'otherwise than at any of the specified worksites. Indeed, for the most part exposure evidence was not allowed and the jury was instructed to assume sufficient exposure. Nor did phase III either litigate or determine whether or to what extent asbestos exposure, either generally or to the product of any particular defendant, was uniform or similar for members of any given craft at any one or more of the specified worksites! ■We note that at least two of the twenty-two sites actually each involved two plants, and ahother involved “the facilities” of a company, “including” its powerhouse. Further, Pittsburgh Corning tendered evidence that a typical refinery covers several -square miles and indicating that at ^refineries, shipyards, and other installations asbestos exposure levels were not uniform at the site or throughout a craft or within a decade or between decades, and that most individuals employed at the twenty-two worksites did not have sufficient exposure to cause asbestosis. Also so tendered was evidence indicating that exposure to asbestos below some level would not produce asbestosis and even above that level risks remain very low until a multiple of five or ten or twenty times the threshold level is reached; that not all those exposed to asbestos in substantial quantities and for protracted periods of time develop asbestosis; that asbestosis develops in “a relatively small percentage of patients with significant asbestos exposure”; and, that although there is a dose response relationship — the more exposure the more risk, the less, the less risk — respecting asbestosis, nevertheless the effect of the same exposure is not the same as between different individuals and “two similarly exposed asbestos workers with exactly the same asbestos historical exposure can go on to have in one case asbestosis and the other case no lung problems.” Moreover, we have held, in a Texas law diversity case, that “the appropriate test for a [plaintiffs] minimum showing of producing cause in asbestos cases” is that stated in Lohrm-ann v. Pittsburgh Corning Corp., 782 F.2d 1156 (4th Cir.1986), namely the “ ‘frequency: regularity-proximity’ test” under which “a motion for summary judgment cannot be defeated merely by alleging work at a shipyard in which defendants’ asbestos products had somewhere been present. Rather, there must be proof of frequent and regular work in an area of the shipyard in proximity to some specific item of defendants’ asbestos containing product.” Slaughter v. Southern Talc. Co., 949 F.2d 167, 171 (5th Cir.1991) (emphasis addedj. It is important to note that this is merely a minimum showing; Slaughter makes clear that making such a showing merely gets a plaintiff to the jury, it does not entitled him to judgment as a matter of law. See id. at 173. Further, it is obvious that for these purposes a shipyard is not considered as a single, undifferentiated, and uniform mass. We have noted that the district court, in the order in which it initially adopted the present plan, stated that for purposes of the then-contemplated phase II trial it would “make a non-jury determination as to which Plaintiffs or Plaintiffs’ decedents worked for a sufficient period of time at each worksite so as to be a proper member of that worksite’s group and which Plaintiffs were proper members of each of the crafts at these work-sites .... ” As previously observed, after phase I the case proceeded directly into phase III without any phase II, and the stipulation was not entered into until phase III was half complete. It is not clear that the district court ever determined that any (or, if so, which) of the tried one hundred sixty phase III plaintiffs, or that any (or if so, which) of the unsevered extrapolation plaintiffs, actually did work at the worksites “for a sufficient period of time” to be “proper members of each of the crafts at these work-sites.” And, if such determinations were made, it is not clear what criteria were employed and what source or sources of information were utilized either in selecting or in applying the criteria. In any event, it is clear not only that any such determination was made non-jury, but further that it was made without either any evidentiary (or other) hearing or any summary judgment procedure (or Fed.R.Civ.P. 50 motion). Accordingly, no such ■ determination can serve to justify or sustain the trial plan as implemented. With one exception, noted below, we are aware of no appellate decision approving such a group, rather than individual, determination of cause in a damage suit for personal injuries to individuals at widely different times and places. For example, in a personal injury suit by individuals living in the neighborhood of a landfill allegedly contaminated by defendant, the Sixth Circuit remarked; “Thus, the court, as is appropriate in this type of mass tort class action litigation, divided its causation analysis into two parts. It was first established that Velsi-col was responsible for the contamination and that the particular contaminants were capable of producing injuries of the types allegedly suffered by the plaintiffs. Up to this point in the proceeding, the five representative plaintiffs were acting primarily in their representative capacity to the class as a whole. This enabled the court the determine a kind of generic causation— whether the combination of the chemical contaminants and the plaintiffs’ exposure to them had the capacity to cause the harm alleged. This still left the matter of individual proximate cause to be determined. Although such generic and individual causation may appear to be inextricably intertwined, the procedural device of the class action permitted the court initially to assess the defendant’s potential liability for its conduct without regard to the individual components of each plaintiffs injuries. However, from this point forward, it became the responsibility of each individual plaintiff to show that his or her specific injuries or damages were proximately caused by ingestion or otherwise using the contaminated water.” Sterling v. Velsicol Chemical Corp., 855 F.2d 1188, 1200 (6th Cir.1988). See also In Re Agent Orange Product Liability Litigation, 818 F.2d 145 (2d Cir.1987), cert. denied, 484 U.S. 1004, 108 S.Ct. 695, 98 L.Ed.2d 648 (1988) (in appeal from settlement in Rule 23(b)(3) class action for agent orange exposure, in which general liability issues, including the military contractor defense, were to be tried class-wide and individual issues, such as each individual’s damages caused by exposure, “were to be left to individual trials,” id. at 150, 164, the court holds certification proper only because of “the centrality of the military contractor defense” and that certification “would have been error” in an action by civilians for exposure during civilian affairs, noting “[t]he relevant question ... is not whether Agent Orange has the capacity to cause harm, the generic causation issue, but whether it did cause harm and to whom. That determination is highly individualistic, and depends upon the characteristics of individual plaintiffs (e.g., state of health, lifestyle) and the nature of their exposure to Agent Orange ...,” id. at 165-166). The district court also justified its trial plan by reliance on Pettway v. American Cast Iron Pipe Co., 494 F.2d 211, 258-63 (5th Cir.1974), where, in a Title VII Rule 23(b)(2) class action, we stated that back pay could be awarded on a class-wide basis, using average rates of pay and approximations, and did not require an individual plaintiff by individual plaintiff approach. However, Pettway is inapplicable here, for each of several reasons. In the first place, Title VII actions are entirely equitable actions and back pay awards therein are strictly equitable remedies, as we recognized in Pettway (“the award of back pay” is “one element of the equitable remedy,” id. at 1125), and as we have held in other decisions both before and after Pettway. Johnson v. Georgia Highway Express, Inc., 417 F.2d 1122, 1125 (5th Cir.1969) (no entitlement to a jury in Title VII action seeking back pay as that is “an integral part of the statutory equitable remedy, to be determined through the exercise of the court’s discretion, and not by a jury”); Wilson v. Belmont Homes, 970 F.2d 53, 54-56 (5th Cir.1992). See also Johnson v. Chapel Hill ISD, 853 F.2d 375, 383 (5th Cir.1988) (front pay). Thus, in Pettway there was no Seventh Amendment right to jury trial. Johnson; Wilson. Here, by contrast, we have personal injury damage suits, the pro-typical Seventh Amendment case. In the second place, Pettway involved only federal law, and hence this Court was not constrained by the Rules of Decision Act and Eñe, as it is here. Relatedly, Pettway involved what Johnson had characterized as an “equitable remedy, to be determined through the exercise of the court’s discretion,” while here the elements of liability and recoverable damages are fixed by state substantive law. Nor do we consider that In Re Chevron U.S.A., Inc., 109 F.3d 1016 (5th Cir.1997), justifies the instant trial plan. That action involved claims by approximately 3,000 neighboring property owners for personal injury and property damage allegedly caused contamination from Chevron’s former crude oil storage waste pit. Apparently no form of class action was involved, although some eases were consolidated. The district court directed that thirty individual plaintiffs be chosen, fifteen by the plaintiffs and fifteen by the defendants, and that there be “a unitary trial on the issues of ‘general liability or causation’ on behalf of the remaining plaintiffs, as well as the individual causation and damage issues of the [thirty] selected plaintiffs.” Id. at 1017. Apparently, the individual causation and damage issues of the remaining un selected plaintiffs would be determined subsequently in individual trials (if the unitary trial established “liability on the part of Chevron for the pollutants that, allegedly, give rise to all of the plaintiffs’ claims,” id. at 1019). Chevron sought mandamus, contending “that the goal of the ‘unitary’ trial was to determine its liability, or lack thereof, in a single trial and to establish bellwether verdicts to which the remaining claims could be matched for settlement purposes.” Id. at 1017. We stated that the thirty selected plaintiffs were not shown or chosen so as to be representative of the other plaintiffs, and observed that “[a] bellwether trial designed to achieve its value ascertainment function for settlement purposes or to answer troubling causation or liability issues common to the universe of claimants has as a core element representativeness .... ” Id. at 1019 (emphasis added). We granted mandamus prohibiting “utilization of the results obtained from the trial of the thirty (30) selected cases for any purpose affecting issues or claims of, or defenses to, the remaining untried cases.” While the majority opinion (one judge specially concurred) contains language generally looking with favor on the use of bellwether verdicts when shown to be statistically representative, this language is plainly dicta, certainly insofar as it might suggest that representative bellwether verdicts could properly be used to determine individual causation and damages for other plaintiffs. Cf. Sterling, 855 F.2d at 1200 (difference between generic and individual causation). To begin with, no such question was before this Court, as the trial plan contemplated that individual causation and damages issues would not be controlled by the thirty individual bellwether verdicts, which would be used to encourage settlement. Moreover, what we did — our holding — was to prevent any preclusive use of the unitary trial results (whether for general causation or individual causation or otherwise) in eases other than those of the thirty selected plaintiffs. And, we concluded that if the district court carried out another, different trial plan, that would present “matters for another panel to consider in the event those decisions are subject to appellate review.” Id. at 1021. Finally, the majority opinion in In Re Chevron U.S.A. does not even cite Fibreboard, or the Seventh Amendment (or discuss the right to jury trial), and does not refer to the Texas substantive law elements of liability and damages in the matter before it. Clearly, In Re Chevron U.S.A. does not control the result here, and this panel is not bound by its dicta. In Hilao v. Estate of Marcos, 103 F.3d 767 (9th Cir.1996), a divided panel of the Ninth Circuit in a rule 23(b)(3) class action permitted recoverable tort damages to be determined in a lump sum for the entire class. Hilao was a suit under the Alien Tort Claims Act, and the Court essentially applied substantive principles of federal or international “common law.” See id. at 776-778. The majority distinguished Fibreboard on the basis that there “the proposed procedure worked a change in the parties’ substantive rights under Texas law that was barred by the Eñe doctrine.” Id. at 785 (footnote omitted). By the same token, Hilao is distinguishable here; it did not operate under the constraints of the Rules of Decision Act or Eñe; the present ease, by contrast, does operate under those constraints. If Hilao is not thus distinguishable it is simply contrary to Fibreboard, which binds us and which in our opinion is in any event correct. Further, Hilao did not address — and there was apparently not presented to it any contention concerning — the Seventh Amendment. Finally, we find ourselves in agreement with the thrust of the dissenting opinion there. Id. at 788 (“Even in the context of a class action, individual causation and individual damages must still be proved individually”). In sum, as Fibreboard held, under Texas law causation must be determined as to “individuals, not groups.” And, the Seventh Amendment gives the right to a jury trial to make that determination. There was no such trial determination made, and no jury determined, that exposure to Pittsburgh Coming’s products was a cause of the asbestos disease of any of the one hundred sixty phase III plaintiffs. Nor does the stipulation determine or establish that. Accordingly, the judgments in all the one hundred forty-three phase III cases before us must be reversed and remanded. We turn now to the extrapolation eases. As to the matter of individual causation, it is obvious that the conclusion we have reached in respect to the phase III cases applies a fortioñ to the extrapolation cases. In the extrapolation cases there was no trial and no jury determination that any individual plaintiff suffered an asbestos-related disease. Indeed, in the extrapolation cases there was no tidal at all — by jury or otherwise — and there was no evidence presented. So, our 'holding as to the phase III cases necessarily requires reversal of the judgments in the five extrapolation cases before us. As to the matter of actual damages, the extrapolation cases are likewise fatally defective. Unlike the phase III cases, in the extrapolation cases there was neither any sort of trial determination, let alone a jury determination, nor even any evidence, of damages. The district court considered that these deficiencies were adequately compensated for by awarding each extrapolation case plaintiff who alleged an asbestos-related disease an amount of actual damages equal to the average of the awards made in the phase III cases for plaintiffs claiming the same category of disease. This plainly contravenes Fibreboard’s holding that under the substantive law of Texas recoverable damages are the “wage losses, pain and suffering, and other elements of compensation” suffered by each of the several particular plaintiffs as “individuals, not groups.” We also observe in this connection that none of the experts at the extrapolation hearing purported to say that the damages suffered by the phase III plaintiffs in a given disease category (whether as disclosed by the phase III evidence or as found by the jury) were to any extent representative of the damages suffered by the extrapolation plaintiffs in the same disease category. The procedure also violates Pittsburgh Coming’s Seventh Amendment right to have the amount of the legally recoverable damages fixed and determined by a juiy. The only juries that spoke to actual damages, the phase I and III juries, received evidence only of the damages to the particular plaintiffs before them, were called on to determine only, and only determined, each of those some one hun