110 F.3d 954

65 USLW 2668, 37 Fed.R.Serv.3d 600

In re FORD MOTOR COMPANY, Petitioner,
Susan I. KELLY, Administratrix and Personal Representative
of the Estate of Gerald A. Kelly, deceased, on Behalf of
Said Decedent's Heirs-At-Law and Next-Of-Kin and on Her Own
Behalf, Respondent/Appellee
v.
FORD MOTOR COMPANY, Appellant.

Nos. 96-2092, 96-2133.

United States Court of Appeals,Third Circuit.

Argued Jan. 28, 1997.
Decided April 9, 1997.
As Amended May 2, 1997.

Joseph V. Pinto, (argued), Evan S. Eisner, Robert Toland, II, White and Williams, Philadelphia, PA, John M. Thomas, Ford Motor Company, Office of General Counsel, Dearborn, MI, for Petitioner in No. 96-2092. Appellant in No. 96-2133, Ford Motor Company.

Robert C. Daniels (argued), Larry Bendesky, Daniels, Saltz, Mongeluzzi & Barrett, Ltd., Philadelphia, PA, for Respondent in No. 96-2092. Appellee in No. 96-2133, Susan I. Kelly.

Before: BECKER, ROTH, Circuit Judges, and ORLOFSKY, District Judge.*

OPINION OF THE COURT

BECKER, Circuit Judge.

1

By this appeal and companion petition for a writ of mandamus in one of the Bronco II product liability cases, the defendant Ford Motor Company, invoking the attorney-client privilege and/or the work product doctrine, challenges a district court order denying it protection from disclosure in discovery of certain documents requested by the plaintiff, Susan Kelly. The question of the appropriate jurisdictional vehicle is precedentially important, for our ability to review such disputes is frequently called into question. Therefore, as a threshold matter, we address the question whether the challenged order is appealable, see 28 U.S.C. § 1291, or reviewable by mandamus, see 28 U.S.C. § 1651.

2

We conclude that, because the district court's order finally resolved an important issue separate from the merits that would be effectively unreviewable after final judgment, we have appellate jurisdiction under the collateral order doctrine. In reaching this conclusion, we consider in some detail the anatomy of the "importance" facet of that doctrine, and we necessarily resolve certain tensions that exist in our recent jurisprudence in the area. Because we have appellate jurisdiction, we do not review the challenged order by way of mandamus, even though our case law would require us to do so if we lacked appellate jurisdiction.

3

In contrast, the merits issues are quite straightforward. We have examined each of the documents in question in camera. They fall into two groups--minutes of a meeting attended by top-level executives of Ford Motor Company regarding the Bronco II, and agendas for a discussion of the technical characteristics of the Bronco II. We conclude that the minutes of the meeting reflect that the recorded communications were for the purpose of obtaining legal advice and hence are protected by the attorney-client privilege. With respect to the agendas and the handwritten notes referring to them, we determine that they were produced by an agent of an attorney in preparation for litigation and hence are protected by the work product doctrine; the other requirements for work product doctrine are not at issue. We will, therefore, reverse the challenged portions of the district court's order and remand with directions to issue an appropriate order protecting the documents from discovery.

I. FACTS AND PROCEDURAL HISTORY

4

In the underlying lawsuit, Kelly claims that Ford defectively designed the Bronco II, a four-wheel drive utility vehicle with a relatively high center of gravity, by rendering it too susceptible to rollover.1 That defective design, Kelly submits, caused the death of her husband, Gerald Kelly, who was killed when the Bronco II he was driving rolled over. Kelly sought to discover Ford documents related to the development, marketing, and safety of the Bronco II. Ford responded, in part, by asserting that the attorney-client privilege and the work product doctrine shielded certain documents from discovery. Ford sought from the district court a protective order that would have preserved the confidentiality of those documents. By letter ruling of October 4, 1996, the district court held that, for the vast majority of documents for which Ford sought protection, the attorney-client privilege and/or the work product doctrine applied. However, the court found that two sets of documents--those at issue here--were discoverable.

5

The first set of documents is the final draft of the minutes of a November 18, 1982 meeting of Ford's Policy and Strategy Committee. The Policy and Strategy Committee is made up of top executives at Ford, and acts as an advisory body to Ford's chief executive officer. At the meeting in question, Ford's general counsel, Henry R. Nolte, Jr., briefed the committee about a report he had drafted regarding the Bronco II. According to the minutes, the committee engaged in an extensive discussion of the report and ultimately adopted the recommendations contained therein.

6

The second set of documents is composed of a series of agendas, one with handwritten notations, for a meeting in 1988, and one document pertaining to a 1989 meeting on which handwritten notes refer to the 1988 agendas. By 1988, numerous lawsuits similar to that brought by Kelly were pending, alleging faulty design of the Bronco II. As part of its defense strategy, Ford retained an outside technical consultant, Failure Analysis Associates (FAA), to assist in the defense of those lawsuits. FAA, in turn, relied in part on the help of in-house technical assistants to Ford. Ernest Grush, one of these technical assistants, prepared the agendas for the 1988 meeting. The meeting was called to explain the technical aspects of the Bronco II litigation defense strategy, and Ford attorneys were present. Grush has declared that the handwritten notes on the document pertaining to the 1989 meeting are his, and that they refer to the 1988 meeting.

7

On October 4, 1996, the district court made a letter ruling denying protection for the documents here at issue. Ford requested that the court reconsider its decision. On November 13, 1996, the court denied Ford's request and, by a subsequent letter ruling of November 27, 1996, ordered the production of the documents by December 18, 1996. The mandamus petition followed. On December 18, the court again ordered the production of the documents. Ford sought and we granted a motion for a stay of the December 18 order. Ford also filed a notice of appeal from the December 18 ruling. We consolidated the appeal and the petition for a writ of mandamus and will, therefore, consider them together.

II. APPELLATE AND MANDAMUS JURISDICTION

A. Introduction; The First Prong of Cohen

8

The question of our jurisdiction is somewhat complicated. A writ of mandamus is an extraordinary exercise of our jurisdiction; moreover, such a writ is not a substitute for appeal. See Madden v. Myers, 102 F.3d 74, 77 (3d Cir.1996). Because we will not issue a writ of mandamus if relief may be granted by way of an ordinary appeal, we must first determine whether Ford may appeal the district court's ruling. See Hahnemann Univ. Hosp. v. Edgar, 74 F.3d 456, 461 (3d Cir.1996). A fortiori, only if an appeal is unavailable will we determine whether a writ of mandamus will issue. See PAS v. Travelers Ins. Co., 7 F.3d 349, 352 (3d Cir.1993).

9

As a general rule, discovery orders are not final orders of the district court for purposes of obtaining appellate jurisdiction under 28 U.S.C. § 1291. See Hahnemann Univ., 74 F.3d at 461. Therefore, discovery orders normally may not be appealed until after final judgment. See id. However, the collateral order doctrine, first enunciated by the Supreme Court in Cohen v. Beneficial Indus. Loan Corp., 337 U.S. 541, 69 S.Ct. 1221, 93 L.Ed. 1528 (1949), provides a narrow exception to the general rule permitting appellate review only of final orders. An appeal of a nonfinal order will lie if (1) the order from which the appellant appeals conclusively determines the disputed question; (2) the order resolves an important issue that is completely separate from the merits of the dispute; and (3) the order is effectively unreviewable on appeal from a final judgment. See Rhone-Poulenc Rorer Inc. v. Home Indem. Co., 32 F.3d 851, 860 (3d Cir.1994).

10

It is beyond cavil that the first element is satisfied here. The district court's December 18 order requiring the production of the disputed documents leaves no room for further consideration by the district court of the claim that the documents are protected.

B. The Second Prong of Cohen

1. Separability

11

The most familiar aspect of the second prong of Cohen is separability from the merits. Kelly submits that a determination of the issues of privilege and work product will in fact implicate the merits of the underlying dispute. We believe that it will not. As we understand the merits of the underlying case, Kelly seeks to show what Ford knew about the alleged rollover propensity of the Bronco II, when it knew about this alleged propensity, and what it did about the alleged propensity. The contents of the documents will certainly shed some light on these questions. However, our resolution of the privilege and work product issues has nothing to do with them. We are not concerned at this juncture about what Ford knew, when it gained this knowledge, or what it did about it. Our inquiry largely involves questions of context--e.g., who prepared the relevant documents, when were they prepared, and what was their purpose. It involves content only insofar as we must ensure that the documents were prepared in certain contexts--e.g., do the documents contain legal advice or do they disclose legal strategies? We are not required, nor will we undertake, to resolve disputed questions of Ford's knowledge of and Ford's actions with respect to the alleged rollover propensity.

12

Kelly's assertions to the contrary notwithstanding, neither Cipollone v. Liggett Group, Inc., 785 F.2d 1108 (3d Cir.1986), nor State of New York v. United States Metals Refining Co., 771 F.2d 796 (3d Cir.1985) [hereinafter USMR ], undermine this conclusion. In both Cipollone and USMR, the defendants sought to protect materials gathered for or discovered during litigation from public dissemination. Each defendant claimed that the sought-after material somehow distorted the actual facts and would, therefore, mislead the public about those facts.2 That claim, we held in both cases, would require us to examine the merits of the underlying dispute because we would need to make some determination of the actual facts presented by the case so as to compare them to the allegedly distorting or misleading material. See Cipollone, 785 F.2d at 1117; USMR, 771 F.2d at 799-800. No such determination need be made here. We can resolve the privilege and work product issues without delving into the disputed facts about Ford's knowledge and actions.

2. Importance

13

The parties have not suggested that the "importance" criterion is not satisfied. However, because of our independent responsibility to examine our own jurisdiction sua sponte, and because the jurisprudence surrounding the importance criterion is somewhat murky, we will undertake a close analysis of this aspect of the collateral order doctrine. Although "[m]ost courts have paid little attention to the 'importance' requirement," John C. Nagel, Replacing the Crazy Quilt of Interlocutory Appeals Jurisprudence With Discretionary Review, 44 Duke L.J. 200, 207 (1994), the Supreme Court has recently made patent that confusion over the criterion cannot lead to the conclusion that " 'importance' is itself unimportant." Digital Equip. Corp. v. Desktop Direct, Inc., 511 U.S. 863, 878, 114 S.Ct. 1992, 2001, 128 L.Ed.2d 842 (1994).3 Rather, application of the Cohen collateral order doctrine is incomplete without an analysis of the importance of the issue sought to be reviewed.

14

Importance has a particular meaning in this context. It does not only refer to general jurisprudential importance. Rather, the overarching principle governing "importance" is that, for the purposes of the Cohen test, an issue is important if the interests that would potentially go unprotected without immediate appellate review of that issue are significant relative to the efficiency interests sought to be advanced by adherence to the final judgment rule.4 In Johnson v. Jones, 515 U.S. 304, 115 S.Ct. 2151, 132 L.Ed.2d 238 (1995), for example, the Supreme Court noted that any analysis of the Cohen test required an examination of the competing considerations that underlie finality, i.e., the costs of piecemeal review on the one hand against the costs of delay on the other. See id. at ----, 115 S.Ct. at 2157. The Court in Digital Equipment stated this in a slightly different manner, noting that the Cohen test requires a "judgment about the value of the interests that would be lost through rigorous application of a final judgment requirement," Digital Equip., 511 U.S. at 878-79, 114 S.Ct. at 2001, and that " 'important' in Cohen 's sense [means] being weightier than the societal interests advanced by the ordinary operation of final judgment principles," id. at 879, 114 S.Ct. at 2002. As a final example, Justice Scalia, in a concurrence, stated that a right is important for Cohen purposes only if it "overcome[s] the policies militating against interlocutory appeals." See Lauro Lines s.r.l. v. Chasser, 490 U.S. 495, 503, 109 S.Ct. 1976, 1980, 104 L.Ed.2d 548 (1989) (Scalia, J., concurring).

15

Although one might assume that collateral finality would be determined by a bright-line rule, the importance determination under the Supreme Court's jurisprudence is rather a function of a balancing process. See, e.g., Behrens v. Pelletier, --- U.S. ----, ----, 116 S.Ct. 834, 844, 133 L.Ed.2d 773 (1996) (Breyer, J., dissenting) (canvassing recent collateral order jurisprudence and noting that the importance analysis is a balancing of interests); Johnson, 515 U.S. at ----, 115 S.Ct. at 2157 (stating that in determining appealability a court must look to the competing considerations that underlie questions of finality, namely "the inconvenience and costs of piecemeal review on the one hand and the danger of denying justice by delay on the other" (citations omitted)). When engaging in this balancing, the Court has relied on a number of factors. We mention here only a few contained in recent cases.

16

In Quackenbush v. Allstate Ins. Co., --- U.S. ----, 116 S.Ct. 1712, 135 L.Ed.2d 1 (1996), the Court allowed the immediate appeal of an abstention-based remand in part because the interests implicated by the appeal--namely, the scope of federal jurisdiction and the desire for comity between the federal and the individual state judicial systems--were sufficiently important. See id. at ---- - ----, 116 S.Ct. at 1719-20. In Digital Equipment, the Court reasoned that a right contained in a private settlement agreement was not sufficiently important in part because that right did not "originat[e] in the Constitution or statutes." Digital Equip., 511 U.S. at 879, 114 S.Ct. at 2001. In Puerto Rico Aqueduct and Sewer Auth. v. Metcalf & Eddy, Inc., 506 U.S. 139, 113 S.Ct. 684, 121 L.Ed.2d 605 (1993), the Court allowed the immediate appeal of a denial of Eleventh Amendment immunity in part because the right at issue "involves a claim to a fundamental constitutional protection." Id. at 145, 113 S.Ct. at 688. And, in Mitchell v. Forsyth, 472 U.S. 511, 105 S.Ct. 2806, 86 L.Ed.2d 411 (1985), the Court allowed the immediate appeal of a claim to qualified immunity in part because such immunity was intended to reduce " 'the general costs of subjecting officials to the risks of trial--distraction of officials from their governmental duties, inhibition of discretionary action, and deterrence of able people from public service.' " Id. at 526, 105 S.Ct. at 2815 (quoting Harlow v. Fitzgerald, 457 U.S. 800, 816, 102 S.Ct. 2727, 2737, 73 L.Ed.2d 396 (1982)).

17

In all of these cases, the Court has compared the apple of the desire to avoid piecemeal litigation to the orange of, for example, federalism.5 In terms of analytic purity, the results of such comparisons are, of course, debatable. What is important for present purposes is that, in a number of the just-cited cases, the Court felt that, because of the imperative of preventing impairment of some institutionally significant status or relationship, the danger of denying justice by reason of delay in appellate adjudication outweighed the inefficiencies flowing from interlocutory appeal. By the same calipers, we are convinced that in the present case the orange of the interests protected by the attorney-client privilege (which would be eviscerated by forced disclosure of privileged material) is sufficiently significant relative to the apple of the interests protected by the final judgment rule to satisfy the importance criterion of the second Cohen prong.

18

In the few cases in which our court has addressed the importance criterion, we have been less than pellucid in our discussion. For example, we stated in Nemours Found. v. Manganaro Corp., New England, 878 F.2d 98 (3d Cir.1989), that the issue on appeal must be important "in a jurisprudential sense," see id. at 100, without explaining what is meant by "jurisprudential." And, in examining whether the relevant issue was important, we have from time to time (though not consistently) raised the question whether the issue presents "a serious and unsettled question." See, e.g., Federal Ins. Co. v. Richard I. Rubin & Co., 12 F.3d 1270, 1282 (3d Cir.1993); Praxis Properties v. Colonial Sav. Bank, 947 F.2d 49, 56 (3d Cir.1991). "A serious and unsettled question" is a factor mentioned in Cohen, see Cohen, 337 U.S. at 547, 69 S.Ct. at 1226, but, for the most part, it has been ignored by the courts, see Robert J. Martineau, Defining Finality and Appealability by Court Rule: Right Problem, Wrong Solution, 54 U. Pitt.L.Rev. 717, 740 (1993).

19

We believe that presenting a serious and unsettled question is merely one means by which an issue may be important under Cohen. It is clear that if a question presents a serious and unsettled question of law, resolution of that issue in an interlocutory appeal protects an interest that is significant relative to the interests protected by deferring review until final judgment. Resolution of a serious and unsettled question has an impact beyond the parties before the court; it not only ensures the proper adjudication of the case before the court, but also may prevent erroneous adjudications in other cases and head off unnecessary appeals in those other cases. These incidental effects promote some of the same goals the final judgment rule promotes. Therefore, though it is not a sine qua non, the presence of a serious and unsettled question is sufficient to satisfy the importance criterion of the Cohen test.6

20

Given our analysis of importance for Cohen purposes, we believe that the attorney-client privilege question before us also satisfies the importance criterion because the interests protected by the privilege are significant relative to the interests advanced by adherence to the final judgment rule. It is often stated that the attorney-client privilege is at the heart of the adversary system; its purpose is to support that system by promoting loyalty and trust between an attorney and a client. See Recent Case, 108 Harv.L.Rev. 775, 779 n. 39 (1995). The privilege is thereby intended to advance the "broad[ ] public interests in the observance of law and administration of justice," Upjohn Co. v. United States, 449 U.S. 383, 389, 101 S.Ct. 677, 682, 66 L.Ed.2d 584 (1981), by encouraging the full and frank communication between attorney and client necessary for vigorous and effective advocacy. Rightly or wrongly, our system assumes that the competition between vigorous and effective advocates, when pitted against each other in an adversary setting, will help to produce the best legal result in any given litigation. In short, the attorney-client privilege is one of the pillars that supports the edifice that is our adversary system. As such, it is "deeply rooted in public policy." Digital Equip., 511 U.S. at 884, 114 S.Ct. at 2004.

21

Privilege doctrine assumes that protecting that loyalty and trust and thereby advancing these broader interests can only be accomplished if privileged material is never disclosed, for only then will clients be encouraged to make full disclosure to their attorneys. By fostering confidentiality, the attorney-client privilege, when vindicated, undermines some of the goals the final judgment rule seeks to realize. Without the benefit of the material protected by the attorney-client privilege, trial courts face a more difficult fact-finding task. Ferreting out the facts of a case becomes more costly, even if only marginally. Often, the privilege will keep trial courts, juries, and appellate courts from considering certain facts, thereby forcing them to decide cases based on less than complete records.

22

In all, the privilege introduces certain inefficiencies into the judicial system, the same inefficiencies with which the final judgment rule is concerned. See supra note 4. Yet, every jurisdiction in this nation recognizes the attorney-client privilege. For the reasons set forth supra, the attorney-client privilege is thus important in the Cohen sense; the status or relationship, deeply embedded in our legal culture, is of sufficient importance that the danger of denying justice by delay in appellate adjudication (which would result in irremediable disclosure of privileged material) outweighs the inefficiencies introduced by immediate appeal. Accordingly, prong two of Cohen is satisfied as to the attorney-client privilege question.

23

For similar reasons, the work product doctrine, at least at its "core," satisfies the importance criterion.7 Like the attorney-client privilege, the work product doctrine seeks to promote the adversary system. It does so "by protecting the confidentiality of papers prepared by or on behalf of attorneys in anticipation of litigation." Westinghouse Elec. Corp. v. Republic of the Philippines, 951 F.2d 1414, 1428 (3d Cir.1991). Absent such protection, attorneys would "fear that their work product will be used against their clients," id., and may become overly circumspect in preparing for litigation thereby reducing their effectiveness as advocates. Such circumspection frustrates the assumptions on which the adversary system is based. "Core" work product thus reflects an institutionally important status or relationship in the law.

24

As with the attorney-client privilege, the work product doctrine rests on the non-disclosure of information. Some of this information is potentially relevant to the disposition of the litigation; keeping it confidential might therefore impede the efficient functioning of the judicial system. Yet, the work product doctrine, or a form of it, is widely recognized. Thus, for the same reasons put forth in our treatment of the attorney-client privilege, core work product, such as at issue here, meets the importance criterion and satisfies the second Cohen prong.

25

C. Effective Review: The Third Prong of Cohen

26

The only remaining issue is the third element of the Cohen test, whether Ford can seek effective review of the privilege and work product issues on appeal after final judgment. The Supreme Court has stated that review after final judgment is ineffective if the right sought to be protected would be, for all practical and legal purposes, destroyed if it were not vindicated prior to final judgment. See, e.g., Lauro Lines, 490 U.S. at 497-99, 109 S.Ct. at 1978. In the context of mandamus jurisdiction, we have repeatedly held that appealing privilege and work product issues after final judgment is ineffective. See Rhone-Poulenc, 32 F.3d at 861 (discussing "privilege or other interests of confidentiality"); Haines v. Liggett Group, Inc., 975 F.2d 81, 89 (3d Cir.1992) (discussing both attorney-client privilege and work product doctrine protections); Westinghouse, 951 F.2d at 1422 (same); Sporck v. Peil, 759 F.2d 312, 314-15 (3d Cir.1985) (discussing work product doctrine protections); Bogosian v. Gulf Oil Corp., 738 F.2d 587, 591 (3d Cir.1984) (same); see also Hahnemann Univ., 74 F.3d at 461 (discussing possible mandamus jurisdiction to review claim that documents were protected by, inter alia, a state law psychotherapist-patient privilege); Glenmede Trust Co. v. Thompson, 56 F.3d 476, 483-84 (3d Cir.1995) (discussing mandamus jurisdiction over review of the terms of a protective order); Smith v. BIC Corp., 869 F.2d 194, 198-99 (3d Cir.1989) (discussing the collateral order doctrine in the context of reviewing a claim that disputed documents contained trade secrets requiring protection); Cipollone v. Liggett Group, Inc., 822 F.2d 335, 340 (3d Cir.1987) (discussing mandamus jurisdiction over review of a protective order).

27

Undergirding those previous holdings is the notion that, once putatively protected material is disclosed, the very "right sought to be protected" has been destroyed. Bogosian, 738 F.2d at 591. That is so because, as we noted previously, underlying the attorney-client privilege is the policy of encouraging full and frank communications between an attorney and client, without the fear of disclosure, so as to aid in the administration of justice. See, e.g., Westinghouse, 951 F.2d at 1423. Concomitantly, the work product doctrine is designed to promote the adversarial process by maintaining the confidentiality of documents prepared by or for attorneys in anticipation of litigation. See, e.g., id. at 1428. Appeal after final judgment cannot remedy the breach in confidentiality occasioned by erroneous disclosure of protected materials. At best, on appeal after final judgment, an appellate court could send the case back for re-trial without use of the protected materials. At that point, however, the cat is already out of the bag.

28

As the Second Circuit aptly stated with respect to the attorney-client privilege, the limited assurance that the protected material will not be disclosed at trial "will not suffice to ensure free and full communication by clients who do not rate highly a privilege that is operative only at the time of trial." Chase Manhattan Bank, N.A. v. Turner & Newall, PLC, 964 F.2d 159, 165 (2d Cir.1992). With respect to material otherwise protected by the work product doctrine, the party will be similarly irremediably disadvantaged by erroneous disclosure. "[A]ttorneys cannot unlearn what has been disclosed to them in discovery"; they are likely to use such material for evidentiary leads, strategy decisions, or the like. Id. More colorfully, there is no way to unscramble the egg scrambled by the disclosure; the baby has been thrown out with the bath water.

29

Our conclusions with respect to privilege and work product issues are buttressed by Supreme Court decisions allowing immediate appeal of official, qualified, and Eleventh Amendment immunities; of double jeopardy challenges; and of speech or debate challenges. In each of those cases, the Court held that the rights asserted protected the claimant against trial, not just liability.8 Therefore, delaying review of orders implicating these asserted rights would preclude vindication of those very rights because delay would allow trial to proceed. The same is true as to privilege and work product issues. Delay in such cases would allow the very disclosure against which those rules protect.

30

In most of our previous cases in which a party sought appellate review of an order requiring the disclosure of putatively protected documents, we did not allow review under the collateral order doctrine either because it was not raised at all by the parties or because the parties did not satisfy either element (1) or element (2) of the Cohen test. In only two cases did we examine element (3) of the Cohen test in this context. In Smith, we held that a party does not have an effective means of appealing after a final judgment an order requiring the disclosure of trade secrets. As we stated there, "once trade secrets are made public, they can obviously never be 'secrets' again." Smith, 869 F.2d at 199. Therefore, the court allowed an interlocutory appeal under the Cohen test and did not reach the question whether a writ of mandamus was appropriate. See id. at 199 n. 3.

31

In the later Rhone-Poulenc case, the other case to examine element (3) of the Cohen test, the panel distinguished Smith by reasoning that any harm caused by the erroneous disclosure of material protected by the attorney-client privilege or the work product doctrine can be remedied. According to that panel, an appellate court can, after final judgment, vacate the ruling of a trial court, remand the case for a new trial, and prohibit the use of the protected material or any material derived from the protected material at the new trial. See Rhone-Poulenc, 32 F.3d at 860. We believe, however, that this part of the holding in Rhone-Poulenc is inconsistent with both Smith and the mandamus line of cases that hold that there can be no effective review after final judgment of an order requiring the disclosure of putatively protected material. See supra. In fact, Rhone-Poulenc seems to say as much when it held that mandamus jurisdiction existed because there is "no other adequate means to attain relief from the district court's order that compels the disclosure of privileged information and work product," citing the mandamus line of cases for support. Id. at 861.

32

Because they precede Rhone-Poulenc, we are bound by the holdings in Smith and the mandamus line of cases. See O. Hommel Co. v. Ferro Corp., 659 F.2d 340, 354 (3d Cir.1981) ("[A] panel of this court cannot overrule a prior panel precedent."). Therefore, we hold that there is no effective means of reviewing after a final judgment an order requiring the production of putatively protected material. Accordingly, the strictures of the collateral order doctrine have been met in this case, and we have jurisdiction over the appeal. Our review of the district court order will be plenary.

D. Mandamus

33

Because we have appellate jurisdiction, there is no need to examine whether we have original, mandamus jurisdiction. However, we also believe that if we did not have appellate jurisdiction, we would have mandamus jurisdiction to review the district court's order. See, e.g., Rhone-Poulenc, 32 F.3d at 861 (exercising mandamus jurisdiction over review of privilege and work product issues); Haines, 975 F.2d at 88-91 (exercising mandamus jurisdiction over review of work product issues); Westinghouse, 951 F.2d at 1422 (exercising mandamus jurisdiction over privilege and work product issues); Sporck, 759 F.2d at 314-15 (exercising mandamus jurisdiction over work product issues); Bogosian, 738 F.2d at 591 (same).

34

The practical difference between appellate jurisdiction and mandamus jurisdiction is the standard of review. Our standard of review under mandamus jurisdiction is exceedingly narrow, see Westinghouse, 951 F.2d at 1423; our standard of review under appellate jurisdiction varies depending on the issue that we are called on to review. Accordingly, mandamus jurisdiction affords an appellate court less opportunity to correct district court error in the case before it and less opportunity to provide guidance for future cases. Moreover, comity between the district and appellate courts is best served by resort to mandamus only in limited circumstances. Review under appellate jurisdiction is therefore preferable to review under mandamus jurisdiction. In light of this preference, the wisdom of our holding that an appeal will lie in this case is confirmed.

35

III. MINUTES OF THE 1982 MEETING; ATTORNEY-CLIENT PRIVILEGE

36

After an in camera review of the relevant documents, we conclude that the final minutes of the 1982 meeting are protected by the attorney-client privilege. Primarily at issue is whether the communications memorialized by the minutes were made for the purpose of obtaining legal advice. Federal Rule of Evidence 501 states:

37

[I]n civil actions and proceedings, with respect to an element of a claim or defense as to which State law supplies the rule of decision, the privilege of a witness, person, government, State, or political subdivision thereof shall be determined in accordance with State law.

38

Fed.R.Evid. 501. In this civil, diversity case in which state law governs, Rule 501 provides that state law will govern the issue of privilege. See Rhone-Poulenc, 32 F.3d at 861-62.

39

It is not clear whether the law of Pennsylvania, the forum state, or the law of Michigan, the state in which the communications occurred, will supply the rule as to privilege. We need not reach this potentially thorny issue, however, because the law as to attorney-client privilege in Pennsylvania does not differ in any significant way from that in Michigan. The elements of the attorney-client privilege are well-known and are not, in any material respect, disputed here. We need not, therefore, dwell on them, except to note their basic contours in Pennsylvania and Michigan.

40

In Pennsylvania, the attorney-client privilege in civil matters has been codified. The relevant statutory provision reads:

41

In a civil matter counsel shall not be competent or permitted to testify to confidential communications made to him by his client, nor shall the client be compelled to disclose the same, unless in either case this privilege is waived upon the trial of the client.

42

42 Pa.Cons.Stat.Ann. § 5928 (West 1982). The communications must be for the purpose of obtaining legal advice. See Leonard Packel & Anne Bowen Poulin, Pennsylvania Evidence § 501.1(c), at 306 & n. 22 (1987 & Supp.1995). A corporation may claim the privilege for communications between its counsel and its employees who have authority to act on its behalf. See Maleski v. Corporate Life Ins. Co., 163 Pa.Cmwlth. 36, 641 A.2d 1, 3 (1994); Packel & Poulin, supra, § 501.1(b).

43

In Michigan, the standard is stated in similar terms. The attorney-client privilege "attaches to the confidential communications made by a client to his attorney acting as a legal adviser and made for the purpose of obtaining legal advice on some right or obligation." Kubiak v. Hurr, 143 Mich.App. 465, 372 N.W.2d 341, 345 (1985). Case law in Michigan also recognizes the right of a corporation to claim the privilege to protect communications between certain of its employees and its counsel. See Hubka v. Pennfield Township, 197 Mich.App. 117, 494 N.W.2d 800, 802 (1992) (citing Mead Data Central, Inc. v. U.S. Dep't of the Air Force, 566 F.2d 242, 253 n. 24 (D.C.Cir.1977) (interpreting the federal Freedom of Information Act)), rev'd on other grounds, 443 Mich. 864, 504 N.W.2d 183 (1993).

44

Our brief review of Pennsylvania and Michigan law as to the attorney-client privilege reveals that the two states agree in the respect most relevant to our case: for a communication to be privileged, it must have been made for the purpose of securing legal advice. See, e.g., Rhone-Poulenc, 32 F.3d at 862 (setting out the traditional elements of the attorney-client privilege and including the requirement that the communication be made for the purpose of securing legal advice); Restatement of the Law Governing Lawyers §§ 118, 122 (Proposed Final Draft No. 1 1996) (same).9 We now turn to determining whether the communications contained in the relevant document satisfy this standard.10

45

Our review of the final minutes, the draft minutes, the report Nolte summarized at the meeting, and relevant affidavits, leads us to conclude that the communications in the meeting were made for the purpose of securing legal advice. Ford clearly had concerns about the Bronco II; this is not surprising given that the product was in the early stages of its development. Nolte examined the legal implications of some of those concerns and proposed a particular course of action, contained in his report to the Policy and Strategy Committee, to address them. The Policy and Strategy Committee meeting itself was called in part to discuss Nolte's proposal. The discussion at the meeting, then, was intended to secure Nolte's legal advice.

46

The district court initially ruled that the minutes "disclose only factual material, contain no legal discussion, were not created in anticipation of litigation ..., and contain no communication to counsel which was intended to be kept confidential." The court later stated that the minutes were "business records" that memorialized "essentially business and safety decisions." We disagree with the district court's conclusions as to the nature of the documents. The documents do not contain merely factual material nor do they detail mere business decisions; in that respect, the district court clearly erred in describing these documents. Certainly, the ultimate decision reached by the Policy and Strategy Committee could be characterized as a business decision, but the Committee reached that decision only after examining the legal implications of doing so. Even if the decision was driven, as the district court seemed to assume, principally by profit and loss, economics, marketing, public relations, or the like, it was also infused with legal concerns, and was reached only after securing legal advice. At all events, disclosure of the documents would reveal that legal advice. We thus hold that the minutes of the 1982 meeting are protected from discovery by the attorney-client privilege.11

IV. THE AGENDAS; THE WORK PRODUCT DOCTRINE

47

Similarly, our in camera review leads us to conclude that the agendas for the 1988 meeting and the handwritten notes on the document pertaining to the 1989 meeting are protected from discovery by the work product doctrine.12 Codified in the Federal Rules of Civil Procedure, the work product doctrine allows a party to discover material prepared in anticipation of litigation or for trial only upon a showing that the requesting party has a substantial need for the material and cannot obtain the material or its equivalent elsewhere without incurring a substantial hardship. See Fed.R.Civ.P. 26(b)(3). The rule as codified provides that "[i]n ordering discovery of such materials when the required showing has been made, the court shall protect against disclosure of the mental impressions, conclusions, opinions, or legal theories of an attorney or other representative of a party concerning the litigation." Id.

48

It is also clear that the work product doctrine protects materials prepared by an agent of the attorney, provided that material was prepared in anticipation of litigation. See 8 Charles Alan Wright, Arthur R. Miller & Richard L. Marcus, Federal Practice and Procedure § 2024, at 359 (2d ed. 1994). These elements, like those of the attorney-client privilege, are well-known and are not, in any relevant respect, disputed here. We need not, therefore, elaborate on them. Rather, the dispute over the agendas turns on whether they were prepared in anticipation of litigation, since the other elements necessary for work product protection are met.13

49

It is clear from our review of the record that the agendas disclose material prepared as part of Ford's legal strategy for defending the type of case Kelly brought here. The agendas outline the results of studies conducted as to the safety of the Bronco II and, in so doing, highlight important aspects of those studies. Those studies were found by the district court to be protected by the work product doctrine because they would be used in defending anticipated lawsuits. Ford persuasively contends that experts acting on behalf of Kelly and working backwards from the agendas could determine the methodology of the studies.14 Ford's attorneys and their agents called for the studies, and Ford credibly demonstrates that if Kelly learns the methodology of the studies, then she has effectively learned of the issues of most concern to Ford's litigation defense team. Moreover, the agendas themselves were for meetings at which the experts would, inter alia, explain the technical aspects of Ford's legal defense strategy by referring to those studies. We are satisfied, in view of the foregoing, that these agendas, core work product, were prepared in anticipation of litigation.

50

The handwritten notations that appear on the document pertaining to the 1989 meeting are similarly protected by the work product doctrine. Although not as extensive as the agendas themselves, the notations refer to the agendas. In some places, the notations employ the same language as that which appears on the agendas. In others, the notations, when read in connection with the typewritten portions of the document to which they refer, provide clear hints as to what is contained in the agendas. In all instances, the notations, like the agendas themselves, would allow Kelly to determine the methodology of the studies.

51

It is true, of course, that the agendas and the handwritten notations (and, for that matter, the studies themselves) were not prepared with this particular litigation in mind. However, that is of no import given the facts of this case. At the time the relevant material was prepared, Ford was a defendant in numerous lawsuits alleging defects in the Bronco II, and this material was prepared in anticipation of those lawsuits. The literal language of Rule 26(b)(3) requires that the material be prepared in anticipation of some litigation, not necessarily in anticipation of the particular litigation in which it is being sought. See In re Grand Jury Proceedings, 604 F.2d 798, 803 (3d Cir.1979) (holding that the work product doctrine will protect material prepared in anticipation of civil proceedings from discovery in a grand jury proceeding); see also 8 Wright, Miller & Marcus, supra, § 2024, at 350-51 (collecting cases and concluding that most courts consider the work product doctrine to protect material prepared in anticipation of previous litigation).15

52

The district court ruled that nothing in the record indicated "that the meetings [for which the agendas were prepared] involved discussion or agenda items about any particular litigation or that the meetings were in anticipation of litigation nor do the documents disclose any legal advice or opinions, or that legal advice was given." Instead, the court ruled that the "meetings were in the nature of product safety meetings, not legal department meetings." As our discussion makes clear, the agendas were prepared in anticipation of litigation. That the agendas do not necessarily include legal advice is, as a matter of law, irrelevant provided, as we note above, they were prepared in anticipation of litigation. Moreover, it is of no import, again as a matter of law, that the meetings for which the agendas were prepared were not legal department meetings. Thus, the district court clearly erred (a function in part of legal error) in concluding that the agendas were not prepared in anticipation of litigation. In this case, the context in which the agendas were discussed does not change the reasons for their preparation.

53

In sum, we conclude that the work product doctrine, as codified in Rule 26(b)(3), protects the agendas from discovery.

V. CONCLUSION

54

In view of the foregoing, the order of the district court dated December 18, 1996 will be reversed in part and the case remanded to the district court with directions to deny discovery of the documents stamped with Bates numbers 6680-82, 13882, 14236, and 21831 in their entirety, and to deny discovery of the handwritten notations on the document stamped with Bates number 14241.

*

Honorable Stephen M. Orlofsky, United States District Judge for the District of New Jersey, sitting by designation

1

Jurisdiction is based on diversity of citizenship. See 28 U.S.C. § 1332

2

In Cipollone, the defendants claimed that the material at issue, though not trade secrets, would nonetheless cause embarrassment if released. See Cipollone, 785 F.2d at 1121. In USMR, the defendant sought to keep confidential a report that the plaintiffs had prepared detailing the pollution at the defendant's plant. See USMR, 771 F.2d at 798

3

In some formulations of the collateral order doctrine, the importance criterion is contained in the second prong of the test; in others, it is considered a factor in the third prong. See Digital Equip., 511 U.S. at 877-79, 114 S.Ct. at 2001. Although the language in Cohen itself implies that it is a separate element of the collateral order test, see Cohen, 337 U.S. at 546, 69 S.Ct. at 1225-26 ("This decision appears to fall in that small class which finally determine claims of right separable from, and collateral to, rights asserted in the action, too important to be denied review and too independent of the cause itself to require that appellate consideration be deferred until the whole case is adjudicated.") (emphasis added), the most frequently cited Supreme Court statement of the test incorporates the importance criterion in the second prong, see Coopers & Lybrand v. Livesay, 437 U.S. 463, 468, 98 S.Ct. 2454, 2458, 57 L.Ed.2d 351 (1978) ("To come within the 'small class' of decisions excepted from the final-judgment rule by Cohen, the order must conclusively determine the disputed question, resolve an important issue completely separate from the merits of the action, and be effectively unreviewable on appeal from a final judgment.") (emphasis added). However, the Supreme Court recently suggested that the importance criterion is a necessary part of the third prong of the test. See Digital Equip., 511 U.S. at 878-79, 114 S.Ct. at 2001 ("[T]he third Cohen question ... simply cannot be answered without a judgment about the value of the interests that would be lost through rigorous application of the final judgment requirement."). Indeed, the ratio decidendi of this portion of the opinion, see infra, has third prong overtones. Yet, in its most recent pronouncement on the collateral order doctrine, the Court included "importance" as a separate prong. See Quackenbush v. Allstate Ins. Co., --- U.S. ----, ---- - ----, 116 S.Ct. 1712, 1718-19, 135 L.Ed.2d 1 (1996). As noted in the text, this court generally incorporates the importance criterion within the second prong. No matter where it is placed, however, it is clear that it must be examined in order to satisfy the collateral order doctrine

4

The Supreme Court has recently described the interests protected by the final judgment rule as follows:

An interlocutory appeal can make it more difficult for trial judges to do their basic job--supervising trial proceedings. It can threaten those proceedings with delay, adding costs and diminishing coherence. It also risks additional, and unnecessary, appellate court work either when it presents appellate courts with less developed records or when it brings them appeals that, had the trial simply proceeded, would have turned out to be unnecessary.

Johnson v. Jones, 515 U.S. 304, ----, 115 S.Ct. 2151, 2154, 132 L.Ed.2d 238 (1995).

5

In addition to the collateral order doctrine cases cited elsewhere in this opinion, we list in this footnote a number of Supreme Court collateral order doctrine cases and the issues that were appealed therein as illustrative of the type of balancing that might be implicated. The list is not exhaustive, nor does the Court explicitly engage in balancing in each of the cases. Those cases are: Swint v. Chambers County Comm'n, 514 U.S. 35, 115 S.Ct. 1203, 131 L.Ed.2d 60 (1995) (municipal liability); Midland Asphalt Corp. v. United States, 489 U.S. 794, 109 S.Ct. 1494, 103 L.Ed.2d 879 (1989) (public disclosure of grand jury matters); Van Cauwenberghe v. Biard, 486 U.S. 517, 108 S.Ct. 1945, 100 L.Ed.2d 517 (1988) (service of process and forum non conveniens ); Gulfstream Aerospace Corp. v. Mayacamas Corp., 485 U.S. 271, 108 S.Ct. 1133, 99 L.Ed.2d 296 (1988) (abstention-related stay); J.B. Stringfellow, Jr. v. Concerned Neighbors in Action, 480 U.S. 370, 107 S.Ct. 1177, 94 L.Ed.2d 389 (1987) (intervention); Richardson-Merrell, Inc. v. Koller, 472 U.S. 424, 105 S.Ct. 2757, 86 L.Ed.2d 340 (1985) (disqualification of counsel); Moses H. Cone Mem'l Hosp. v. Mercury Constr. Corp., 460 U.S. 1, 103 S.Ct. 927, 74 L.Ed.2d 765 (1983) (abstention-related stay); Firestone Tire & Rubber Co. v. Risjord, 449 U.S. 368, 101 S.Ct. 669, 66 L.Ed.2d 571 (1981) (disqualification of counsel); Helstoski v. Meanor, 442 U.S. 500, 99 S.Ct. 2445, 61 L.Ed.2d 30 (1979) (Speech and Debate Clause); United States v. MacDonald, 435 U.S. 850, 98 S.Ct. 1547, 56 L.Ed.2d 18 (1978) (speedy trial); Abney v. United States, 431 U.S. 651, 97 S.Ct. 2034, 52 L.Ed.2d 651 (1977) (double jeopardy)

6

Nixon v. Fitzgerald, 457 U.S. 731, 102 S.Ct. 2690, 73 L.Ed.2d 349 (1982), is a case in which the Supreme Court stated that an appeal must present a serious and unsettled question to fall within Cohen 's scope. See id. at 742, 102 S.Ct. at 2697. However, in so doing, the Court seemed to imply that a serious and unsettled question is merely one part of Cohen 's importance requirement. In determining that the appeal before it did present a serious and unsettled question, the Court relied on the fact that it had never ruled on the question; that the Court of Appeals had done so did not settle the question for Cohen purposes. See id. at 743, 102 S.Ct. at 2697-98. This is curious reasoning; following it to its logical extreme, it would categorize as serious and unsettled any issue the Supreme Court has not decided. At all events, later in Nixon the Court seems to limit this reasoning. It noted that the case before it pertained to sensitive issues related to the separation of powers between the executive and judicial branches of government. See id. The mention of these sensitive issues "might hint that the calculus of appeal includes the importance of the interests involved as well as the general importance of the question to other litigants." 15A Charles Alan Wright, Arthur R. Miller & Richard L. Marcus, Federal Practice and Procedure § 3911.5, at 431 (2d ed. 1994). As is evident in the text, our rendering of the importance prong is consistent with this discussion

7

By the "core," we mean the "mental impressions, conclusions, opinion, or legal theories of an attorney or other representative of a party concerning the litigation." Fed.R.Civ.P. 26(b)(3). Such core work product is generally afforded near absolute protection from discovery. See 8 Charles Alan Wright, Arthur R. Miller & Richard L. Marcus, Federal Practice and Procedure § 2026 (2d ed. 1994). Because, as we discuss infra, the work product at issue here is at the core of the doctrine, we have no occasion to discuss whether work product generally is important for Cohen purposes

8

See Puerto Rico Aqueduct and Sewer Auth. v. Metcalf & Eddy, Inc., 506 U.S. 139, 142-47, 113 S.Ct. 684, 687-89, 121 L.Ed.2d 605 (1993) (examining Eleventh Amendment immunity); Mitchell v. Forsyth, 472 U.S. 511, 525-27, 105 S.Ct. 2806, 2814-16, 86 L.Ed.2d 411 (1985) (examining qualified immunity); Nixon v. Fitzgerald, 457 U.S. 731, 741-43, 102 S.Ct. 2690, 2696-98, 73 L.Ed.2d 349 (1982) (examining absolute immunity); Helstoski v. Meanor, 442 U.S. 500, 506-08, 99 S.Ct. 2445, 2448-49, 61 L.Ed.2d 30 (1979) (examining the Speech and Debate Clause); Abney v. United States, 431 U.S. 651, 660-662, 97 S.Ct. 2034, 2040-42, 52 L.Ed.2d 651 (1977) (examining double jeopardy)

9

It should be noted that the law makes no distinction between communications made by a client and those made by an attorney, provided the communications are for the purpose of securing legal advice. See Restatement of the Law Governing Lawyers §§ 118, 120 (Proposed Final Draft No. 1 1996). In other words, the entire discussion between a client and an attorney undertaken to secure legal advice is privileged, no matter whether the client or the attorney is speaking

10

The parties do not dispute that Nolte was Ford's attorney at all relevant times and that the members of the Policy and Strategy Committee had the authority to act on behalf of Ford. Although Kelly does argue that Ford did not intend for the communications to be kept confidential, we find that argument to be without merit. Ford's actions with respect to these documents clearly evinced such an intent

11

Discussion in a published opinion of our conclusions based upon an in camera review is necessarily limited. We cannot reveal too much about the contents of the documents for fear of undermining the very purposes of such review. Our methodology is to reveal only as much of the content as is necessary to produce a reasoned opinion that can itself be reviewed. If further review is necessary, the en banc court or the Supreme Court can examine for itself the relevant documents in conjunction with our opinion. We recognize that the advocacy of the attorneys representing the party seeking allegedly protected documents is hampered by their inability to review those same documents. That disadvantage is one we must accept; otherwise, the very purpose of the privilege will be destroyed

The observations made in this footnote apply equally to our discussion of the documents allegedly protected by the work product doctrine.

12

Ford claims that the agendas are also protected by the attorney-client privilege. We disagree. There is no indication in the record that the relevant 1988 meeting at which the agendas were discussed involved the kind of communications the privilege protects. Ford's assertions to the contrary and the affidavits supporting them are nothing more than conclusory

Ford also claims that the handwritten notes on a document pertaining to the 1989 meeting are protected by the attorney-client privilege because they refer to legal advice provided at the 1988 meeting. (Ford does not claim that the meeting itself or the typewritten portions of the document are protected.) Because we do not see the 1988 meeting as involving confidential communications made to secure legal advice, we do not believe these handwritten notes are privileged. However, these notes do refer to the agendas from the 1988 meetings and to the studies on which the agendas were based. We will, therefore, consider these notes as being equivalent to the 1988 agendas. As we discuss in the text, the 1988 agendas are protected by the work product doctrine. These notes, then, are similarly protected.

13

The record makes it clear that the agendas were prepared by an agent of Ford's attorneys. In addition, Kelly has not made the requisite showing of substantial need to overcome the work product doctrine protections

14

Although Kelly does not dispute this contention, we suspect that it might have been difficult for Kelly to do so given that she has not seen the agendas

15

As in In re Grand Jury Proceedings, there is "an identity of subject matter" between the litigation for which the material was prepared and the present litigation. See In re Grand Jury Proceedings, 604 F.2d at 803. We therefore need not decide whether the work product doctrine protects material prepared for any previous litigation, or only previous litigation related to the present litigation