553 F.2d 714
16 Fair Empl.Prac.Cas. 941, 14 Empl. Prac.
Dec. P 7524
Louise LAMPHERE et al., Plaintiffs, Appellees,
BROWN UNIVERSITY et al., Defendants, Appellants.
United States Court of Appeals,
Argued Jan. 5, 1977.
Decided April 25, 1977.
Jordan Stanzler, Providence, R. I., with whom Abedon, Stanzler, Biener, Skolnik & Lipsey, Providence, R. I., were on motion for summary disposition and memorandum in support thereof, for plaintiffs, appellees.
Laurence S. Fordham, Boston, Mass., with whom J. Harold Flannery, Stephen B. Deutsch, Scott Carson Moriearty, Foley, Hoag & Eliot, Boston, Mass., De Witt T. Kersh, Jr., Richard A. Sherman, and Tillinghast, Collins & Grahm, Providence, R. I., were on memorandum in opposition thereto, for defendants, appellants.
Before COFFIN, Chief Judge, CAMPBELL, Circuit Judge, and BOWNES,* District Judge.
COFFIN, Chief Judge.
This is an appeal by Brown University from a district court decision to certify a class of present, past, and future actual or potential women academic employees in a sex discrimination suit brought by plaintiff Louise Lamphere, an anthropology professor who was denied tenure.1 The district court declined to certify the question of the appropriateness of its action under 28 U.S.C. § 1292(b) and we earlier refused to grant mandamus. We are now pressed by Brown to plough admittedly new ground in this circuit and accept appellate jurisdiction because of allegedly egregious errors of law committed by the district court; the novelty of a broadside class action challenging university-wide tenure policy brought against a university maintaining a decentralized, department-oriented decision making structure; and the imponderable but high costs of defense. We decline the invitation, but not without some reservations.
The complaint was filed almost two years ago, in May of 1975. Plaintiff launched an aggressive program of discovery. Four sets of interrogatories were filed and answered by extensive compilations of employment decision data and narrative answers. Statistics alone not satisfying plaintiff, the court allowed names of faculty who received an adverse employment or promotion decision to be produced, but under a protective order. Depositions were taken. Files of faculty were subpoenaed. In July of 1976 the district court certified the action as a 23(b)(2) class action, the class consisting of:
All women who have been employed in faculty positions by Brown University at any time after March 24, 1972, or who have applied for but were denied employment by Brown in such positions after said date; all women who are now so employed; all women who may in the future be so employed or who may in the future apply for but be denied such employment, and which groups of women have been, are being, or may in the future be, discriminated against on the basis of their sex by defendant's practices with respect to hiring, contract renewal, promotion, and tenure.
In its opinion, the court recognized that plaintiff had the burden of meeting the criteria of Rule 23, Fed.R.Civ.P., but it also followed the authorities which have acknowledged the legitimacy of an across-the-board approach to class certification in Title VII cases.2 It recognized that the affidavits of two department chairmen and of the Provost of Brown stated that tenure decisions were in major part left to the thirty-odd departments and divisions of the university, with only limited central review. But it also took note of contrary indications in the minutes of the central Committee on Appointment and Promotion. It concluded, first, that the evidence was "mixed" and that Brown had not demonstrated that plaintiff would be unable to prove her allegation as to central decision making; and, second, that in any event the issue of centralized or decentralized decisions implicated the merits and need not be decided "at this early stage". The court came to the same conclusions as to the issue of the uniqueness of a tenure decision compared to other employment decisions; it held that the evidence as to disparateness of criteria was disputed and that it also went to the merits. After making some exclusions in defining eligible members who had in the past been employees and ruling out those who had never applied for employment, the court certified the class above described, amounting to an estimated 20,100 people.
Brown challenges the court's action for relying on the allegations of the complaint while ignoring the evidence, for shifting the burden of persuasion on the class issue to the defendants, and for adopting an across-the-board approach which it contends is singularly inappropriate for such a multisource decision apparatus as is a modern, first class, university. It has pointed out that in defending a broad class action lawsuit it must try to insulate every one of thirty or so departments from sex discrimination taint, an effort which would require the preparation by two attorneys of some hundred witnesses at ten hours each, or 2000 hours. In addition to the disruption of its educational mission, it estimates the cost of an unreviewed and erroneous class action certification at between $150,000 and $300,000. It points out that, after all the discovery that has been had, plaintiff has found only two other putative discriminatees.
In considering whether we should extend the finality concept of 28 U.S.C. § 1291 to determinations of class, we are aware of the mine field we are being asked to enter. So far, our single foray into flexible finality has been modest.3 We have not had occasion to decide whether or not to adopt the "death knell" doctrine, which would give interlocutory review to a denial of class action status that effectively terminates the litigation.4 Beyond this adaptation of the principles of Cohen v. Beneficial Industrial Loan Corp., 337 U.S. 541, 69 S.Ct. 1221, 93 L.Ed. 1528 (1949) and Gillespie v. United States Steel Corp., 379 U.S. 148, 85 S.Ct. 308, 13 L.Ed.2d 199 (1964), only the Second Circuit has ventured, and then with uncertain tread.
After the Sixth5 and Seventh6 Circuits had rejected interlocutory review of orders granting class certification, the Second Circuit, in Eisen v. Carlisle & Jacquelin, 479 F.2d 1005 (1973) (Eisen III ), vacated, 417 U.S. 156, 94 S.Ct. 2140, 40 L.Ed.2d 732 (1974), announced its tripartite test which, when met, would justify review of a class action certification when it is crucial to the conduct of the case, collateral to the merits, and productive of irreparable harm absent immediate review. It then applied this test in Herbst v. International Telephone and Telegraph Corp., 495 F.2d 1308 (2d Cir. 1974).7 Subsequent cases have seen a pulling back. In Kohn v. Royall, Keogel & Wells, 496 F.2d 1094 (2d Cir. 1974), the court noted that one rationale for allowing appeals from certification orders was equality of treatment between plaintiffs and defendants, and therefore interpreted the requirement of cruciality to mean that reversal of the order must sound the "death knell" of the action. See also In Re Master Key Antitrust Litigation, 528 F.2d 5 (2d Cir. 1975). In General Motors Corp. v. City of New York, 501 F.2d 639 (2d Cir. 1974), the court added the notion that it would not review a decision which, unlike a decision to dispense with security as in Cohen, supra, or a decision to shift notice costs as in Eisen III, supra, did not involve a question of judicial power, but rather was a "discretionary decision, the propriety of which will necessarily vary from case to case." 501 F.2d at 647. A decision to the same effect was Parkinson v. April Industries, Inc., 520 F.2d 650 (2d Cir. 1975).8 We conclude our roundup by noting that three other circuits have recently joined the ranks of non-reviewers of class certification decisions.9
A recent thoughtful comment in 44 Fordham Law Review 548, 559-60 (1975), Appealability of Class Action Determinations, points out several criticisms of the tripartite test. The Second Circuit itself recognized in Kohn, supra, 496 F.2d at 1098-99, that reliance on whether reversal of the class certification will effectively terminate the action is inconsistent with non-review of motions denying summary judgment, where reversal on appeal would also end the litigation. The requirement that the order be separable from the merits has proven difficult for litigants to meet. In our view, the strongest criticism may be levelled at the requirement of irreparable harm. The notion that the defendant is irreparably harmed if class certification significantly increases its litigation burden runs contrary to a fundamental premise of the final judgment rule: that, in general, there is no substantive right to protection from unnecessary litigation. Numerous rulings before and during trial, from the scope of discovery to the admissibility of evidence, may impose significant burdens on the parties. If the hardships of trial are routinely held to implicate a separate interest reviewable under the collateral order doctrine, the distinction between interlocutory and final orders would be seriously undermined, if not eliminated. See Mercantile Nat. Bank v. Langdeau, 371 U.S. 555, 574, 83 S.Ct. 520, 9 L.Ed.2d 523 (1963) (Harlan, J., dissenting).
The Fordham Comment concludes its survey of the Second Circuit's venture as follows:
"Thus, it is doubtful that orders certifying class actions are appealable under any circumstances. The Cohen requirement of separability from the merits is difficult to meet; discretionary orders prohibited by Cohen are often involved; and review of the order is available after final judgment. Gillespie is clearly inapplicable since no parties are denied justice by delayed review. Perhaps, however, the most compelling argument against permitting immediate appeal is the specific provision in rule 23(c)(2) for review by the district court of its certification order. Permitting an appeal each time a district court issues a new certification order could indeed lead to successive appeals on the same issue, a result clearly at odds with the purposes of the final judgment rule." (Footnotes omitted.) 44 Fordham L.Rev. at 560-61.
We leave this appraisal of the current state of the law with disquiet at the prospect of affording interlocutory review of class determinations. We would not take the step in the absence of the most compelling circumstances, which we do not find here. On the contrary, even if we were inclined to follow the Second Circuit's lead, this case does not meet that court's requirements for review.
Plaintiff Lamphere asserts, as did the individual Title VII plaintiff in Kohn, supra, that she will continue her suit whether or not it is certified as a class action. Thus, it seems that the certification order is not "crucial" since this lawsuit would not terminate even if we were to reverse the class certification. As to separability, Kohn points out that, unlike the issues of notice and manageability that Herbst and Eisen III considered, deciding whether there are common questions, or whether an individual claim is typical, may often implicate the merits of the case. 496 F.2d at 1100. Defendants' criticism of the district court's across-the-board approach turns on the assumption that the university's tenure decisions are made by autonomous departments. But acceptance of this stance would plunge us very far into the merits. Indeed, the issues of centrality of decision making, uniqueness of tenure decisions, and validity of an across-the-board approach all implicate the merits very deeply.
The district court also recognized this. While it relied not only on the pleadings but also on the conflicting evidence as to centralized control and tenure decisions, it felt that for purposes of class certification, plaintiff had made out a prima facie case that had not been destroyed by defendants' submissions. Defendants now urge that the district court's approach improperly relieved the plaintiff of her burden of proving that she was entitled to class certification, and was so clearly an abuse of its power under Rule 23 that this court should take appellate jurisdiction to correct it. An error of this sort might impel us to intervene in mandamus, rather than to accept appellate jurisdiction.10 We have already denied defendants' petition for a writ of mandamus, In re Brown University, 546 F.2d 414 (1st Cir. 1976), and are not persuaded either to reconsider that decision, or to stretch our concept of appellate jurisdiction to review the district court's action.
Defendants ask us to chart a procedure or set of principles that would require the district court to engage in extensive fact finding prior to arriving at a class determination. In some cases this may be desirable, but in others it plainly would not be. There is authority, indeed, for making the decision entirely on the basis of the pleadings. See Developments in the Law Class Actions, 89 Harv.L.Rev. 1318, 1422 n. 175 (1976). We did indicate in Yaffe, supra, 454 F.2d at 1367, that it would be appropriate to permit discovery on a broader-than-individual basis before finally denying class certification, but neither that case nor any other we have been cited to establishes that the district court must, as a general practice, hold an evidentiary hearing before granting class status.11 Given the inexhaustible variety of cases, we think the issue must in most imaginable circumstances be entrusted to the district court during the trial. Rule 23(c)(1) specifically calls for a decision on the class issue as early as practicable in the litigation, and also provides for reconsideration of the decision as the case progresses. The underlying theme is flexibility; different cases call for different approaches. In retrospect we may, with our 20-20 hindsight vision, conclude that the court committed some reversible error, but we lack the foresight to announce guidelines. Like the Second Circuit, we decline to intervene in what is essentially an exercise of discretion by the district court. See General Motors Corp. v. New York City, supra, 501 F.2d at 647.
Finally, to establish irreparable harm, the university stresses the burden it will bear by having to defend broadly, rather than by meeting one charge of discrimination within one department. We have noted its estimate of 2000 lawyer-witness preparation hours and its estimated additional cost of $150,000 to $300,000. However, we note also that unlike Eisen III, supra, the extra burden complained of here is not a cost unique to the class procedure but a "routine litigation cost". See Kohn, supra, 496 F.2d at 1100. Moreover, the university underestimates, in our opinion, the burden and cost of defending any discrimination suit. See id. We say this with no sense of satisfaction. But, as the Fifth Circuit stated in Jenkins v. United Gas Corp., 400 F.2d 28, 33 (5th Cir. 1968), a Title VII suit brought by an individual "is perforce a sort of class action". In such an action, "(a)ll evidence showing a policy of discrimination toward the class is relevant to show discrimination against the individual plaintiff." Paddison v. Fidelity Bank, 60 F.R.D. 695, 697 (E.D.Pa.1972). Once a plaintiff establishes a prima facie case of discrimination, the burden shifts to the defendant to show a non-discriminatory motive. The burden then shifts back to plaintiff to show that the rebuttal is pretextual. McDonnell Douglas Corp. v. Green, 411 U.S. 792, 804, 93 S.Ct. 1817, 36 L.Ed.2d 668 (1973). And at this point classwide proof becomes particularly pertinent. Id. at 804-05, 93 S.Ct. 1817. See generally, Comment, 20 St. Louis University Law Journal 388, 394-95 (1976), To What Extent Can a Court Remedy Classwide Discrimination in an Individual Suit Under Title VII?
Even in an individual action, injunctive relief in effect flows to the entire class affected, id. at 395-96, 93 S.Ct. 1817 and cases cited at 396 n. 71, 93 S.Ct. 1817, and some courts have ordered reinstatement or back-pay for class members. Id. at 398-99, 93 S.Ct. 1817. In the present case, the district court specifically noted that the scope of relief would be the subject of separate inquiry after the trial on liability.
To this we add several observations, which would be obvious to a defense strategy planner. The district judge, even were a class certification denied at one point in a case, may well change his mind later on. If a class is disestablished at one stage, a continuing refusal to certify might be reversed on appeal and, if the plaintiff is individually successful, a new trial might be ordered on a class-wide basis. Both possibilities would induce prudent defense counsel to defend more broadly than might be called for were the suit unalterably an individual action. Still another possible eventuality, even if one suit is confined to its individual status, is a proliferation of suits by other potential class members. In such event, defendant might find himself not only put to additional expense but shackled by collateral estoppel. See Fordham Comment supra, 44 Fordham L.Rev. at 576. In short, the "saving" in costs between defending an individual's sex discrimination suit and defending a class action may have been overestimated. In any event, we cannot find the prospect of avoidable costs so clear and so damaging if not prevented, as to persuade us to accept interlocutory review.
The very facts that so wide a discretion is vested in the district court, that review is postponed and its efficacy is to some extent eroded, and that injustice can result from errors of discretion place a heavy responsibility on the district court. These considerations, which have led the Second Circuit to take appellate jurisdiction of some class certification decisions, lead us to underscore the need for the district court to follow closely the developing evidence as to class-wide decision making and to take seriously its power under rule 23(c)(1) to alter or amend its certification order before the decision on the merits.
For private colleges and universities, not only do heavy litigation expenses impose a financial burden which may ultimately be passed on in some measure to students or be absorbed to the detriment of some educational function, but the process of involving every department in depth in fact finding and defense may well be deeply disruptive to the institution's mission. In both a material and non-material sense they are fragile, not perhaps in their capacity to endure but in their capacity to remain effective while enduring. This is not to suggest that other institutions, such as hospitals or small businesses, may not be similarly burdened. What we do suggest is that the implications of class-wide defense should be considered, along with other factors, as the district court reevaluates the class certification through various stages of trial.
If, after all pre-trial activity has been completed, conference reveals that the complexity, length, and expense of trial would be substantially increased if the case were to continue with the class defined as it presently is, and if the court by that time has come to a clear view of the proper scope of the complaint, we think it would be well advised to review the evidence, possibly with the assistance of oral argument and briefs, and redetermine whether the class certification should remain, should be altered, or should be stricken and the suit confined to plaintiff and other individuals whose intervention seems appropriate.12 Whatever the result, the parties would know that the court had done its best to fulfill its responsibilities in class action monitoring. And for purposes of review we would have a much better record of the factors influencing the court.
Appeal dismissed for lack of appellate jurisdiction.
Of the District of New Hampshire, sitting by designation
The issue was brought to a head by plaintiff's motion for summary dismissal of the appeal and Brown's opposition. Briefs were filed, arguments were heard, and our disposition is not summary but on the merits of the appeal
See, e. g., Rich v. Martin Marietta Corp., 522 F.2d 333 (10th Cir. 1975); Barnett v. W. T. Grant Co., 518 F.2d 543 (4th Cir. 1975); Wetzel v. Liberty Mutual Ins. Co., 508 F.2d 239 (3d Cir. 1975); Mosley v. General Motors Corp., 497 F.2d 1330 (8th Cir. 1974); Piva v. Xerox Corp., 11 F.E.P. Cases 1259 (N.D.Cal.1975); Leisner v. New York Telephone Co., 358 F.Supp. 359 (S.D.N.Y.1973)
In Yaffe v. Powers, 454 F.2d 1362 (1st Cir. 1972), we permitted an appeal from a denial of class action on the theory that such denial was tantamount to narrowing the scope of any possible injunctive relief and that review thus lay under 28 U.S.C. § 1292(a)(1)
The Second Circuit first articulated the doctrine in Eisen v. Carlisle & Jacquelin, 370 F.2d 119 (2d Cir. 1966) (Eisen I ). Since that time, several circuits have considered the doctrine but, without deciding whether to adopt it, held it inapplicable in cases where the plaintiff had sufficient incentive to continue the suit. See, e. g., Williams v. Mumford, 167 U.S.App.D.C. 125, 511 F.2d 363 (1975); Hartmann v. Scott, 488 F.2d 1215 (8th Cir. 1973); Falk v. Dempsey-Tegler & Co., 472 F.2d 142 (9th Cir. 1972). It has been adopted and applied by two circuits, see Miller v. Mackey Intl., Inc., 452 F.2d 424, 427 n. 3 (5th Cir. 1971); Ott v. Speedwriting Pub. Co., 518 F.2d 1143, 1149 (6th Cir. 1975); and it has been expressly rejected by two circuits, see Hackett v. General Host Corp., 455 F.2d 618 (3d Cir. 1972); King v. Kansas City S. Indus., Inc., 479 F.2d 1259 (7th Cir. 1973)
Even in this highwater mark decision, both Senior Judge Danaher, sitting by designation, and Judge Mulligan registered troubled concurrences. 495 F.2d at 1317, 1325
The court also stated, "(t)he continuing precedential validity of Herbst . . . is an open question which is not now before us." 520 F.2d at 658 n. 9
The Second Circuit's reasoning in Herbst, supra, 495 F.2d at 1313, introduces a concept of "supervisory" appellate jurisdiction that is foreign to us
Indeed the Supreme Court's assertion, in a slightly different context, that Rule 23 does not confer authority "to conduct a preliminary inquiry into the merits of a suit in order to determine whether it may be maintained as a class action", Eisen v. Carlisle & Jacquelin, 417 U.S. 156, 177, 94 S.Ct. 2140, 2152, 40 L.Ed.2d 732 (1974), could be taken to suggest that, at least where issues such as commonality or typicality implicate the merits, the allegations in the pleadings are determinative. We note also that Rule 23(a) requires common questions of law or fact, and that a representative's claim be typical of the class: how the questions will be answered, and whether any claim will be proved, are irrelevant to the class determination. See id. 417 U.S. at 178, 94 S.Ct. 2140. Issues such as those raised by the defendant in this case are perhaps more properly the basis for a motion for summary judgment than objections to class certification. Paddison v. Fidelity Bank, 60 F.R.D. 695, 696 (E.D.Pa.1973). See generally 3B Moore's Federal Practice P 23.02-2, n. 35 (Supp.1975 at 19-21)
We also note that if the plaintiffs at the close of their case have failed to make a prima facie case of class-wide discrimination, defendants could presumably move for judgment in their favor as to the class plaintiffs