Bogosian v. Gulf Oil Corp.

BECKER, Circuit Judge,

dissenting.

I agree with many of the sentiments expressed by the majority concerning the desirability of enabling attorneys to put their thoughts on paper without having to fear that there recollections will be subject to the eyes of adversaries during litigation. 1 Sharing the majority’s high regard for Judge Van Artsdalen, I also understand the delicacy of the majority in its refusal to grant a writ of mandamus despite what it apparently believes to be the “clear legal error” of the district court. See Gold v. Johns-Manville Sales Corp., 723 F.2d 1068, 1074 (3d Cir.1983) (setting forth standard for mandamus). 2 I believe, however, that observance of the formalities is the better and more certain practice. I would issue the writ in this case, though the course I would direct the district court to follow upon remand differs significantly from that suggested by the majority.

I. THE REFUSAL TO ISSUE MANDAMUS

To begin with, I believe it is a mistake for the majority to refuse to issue a writ of mandamus after having written an opinion that in essence directs the district court to reverse itself and protect the work product. 3 First, there is mischief resulting from this discrepancy between opinion and judgment. If, as I assume will happen, the district court heeds the majority’s advice and denies production of the documents the defendants had sought, the defendants will, as a practical matter, have lost this case. Because of the failure of the court to issue *597 the writ, however, the defendants may technically have won and thus be foreclosed from petitioning the Supreme Court for certiorari review of today’s decision. 4 Second, because the impact of the majority opinion is not clear, at least to me, in certain of its aspects, 5 I think the better practice is to issue a clearly worded writ lest the district court be confused as to its charter on remand.

II. THE MERITS

Although the majority does not so state in terms, as I read its opinion, there is no absolute protection against discovery of “core work product.” Rather, under the majority’s approach, core work product is subject to an extremely high degree of protection that can be overcome only by a kind of showing that has yet to be made in any reported case but that may yet be made on some extraordinary, as yet unforeseeable record. Accord In re Murphy, 560 F.2d 326 (8th Cir.1977). Because the majority has itself conducted a balancing and found the value of the documents sought after to be only “marginal,” I also read the majority as advising Judge Van Artsdalen to vacate his order (Class Action Order # 88) for production of attorney core work product, without any further proceedings.

In holding as it has, the majority has failed to consider the most important interest of the defendants in discovering this material at this point in the litigation: the need to impeach plaintiffs’ expert economist at his deposition. Although ordinarily this interest would not be sufficient to require the production of “core work product,” in this case, as Judge Van Artsdalen understood, the plaintiffs might well seek to introduce the deposition as evidence at trial, in lieu of live testimony. Under these circumstances, I believe the question we must confront is whether the defendants’ interest in having this material available for cross-examination of the expert economist “at trial” outweighs the plaintiffs’ interest in protecting this core work product.

Under Rules of Evidence 802 and 804(b)(1), deposition testimony is admissible into evidence only if adverse parties had an opportunity to cross-examine the deponent. This opportunity must, of course, be roughly comparable to that which would have been available at trial. See Mancusi v. Stubbs, 408 U.S. 204, 213-16 , 92 S.Ct. 2308, 2313-14 , 33 L.Ed.2d 293 (1972); United States v. Franklin, 235 F.Supp. 338, 341 (D.D.C.1964). Accordingly, the majority is implicitly holding that the district court would commit a mandamus-worthy abuse of discretion if it failed to quash a subpoena duces tecum aimed at the “work product” for use in impeaching an expert testifying at trial, and is also suggesting that the unavailability of such documents during discovery does not so deprive the defendants in this case of the right to cross-examine that the deposition testimony would be inadmissible under Rules 802 and *598 804(b)(1). 6 These are different — and much more difficult — questions than those the majority actually addresses. Whether the district court under such circumstances would commit an abuse of discretion either by failing to quash the subpoena duces tecum or by excluding the deposition transcript depends on the circumstances, I believe. This very fact suggests that it is primarily a matter for the exercise of district court discretion in the crucible of trial.

Were we in an ordinary discovery situation, I might go as far as the majority does in protecting the work product. Since the primary value of the communications between expert and attorney does not come from its capacity to lead to admissible evidence, see Fed.R.Civ.P. 26(b)(1), but rather in its value as impeachment, there is no great need early in the litigation for these documents. Thus, in the ordinary situation where the deponent will testify at trial, if at all, concerns over work product doctrine may best be resolved at an in limine hearing just prior to or during the trial. By that time, the problem may have disappeared or be more carefully focused than it is at discovery, when the issues in the case and the probative value of testimony is often uncertain. Where as here, however, the witness will apparently be unavailable at trial, the district court may not have this luxury.

Turning to the exercise of discretion, one circumstance that must guide the district court in determining whether to quash the subpoena or, equivalently in this case, to deny discovery, is the adverse parties’ need for the work product. I disagree with the majority’s pronouncement — which it apparently believes to be one of law — that evidence demonstrating that an economist’s theory did not originate or evolve as a result of his own research, but rather as a result of the hiring lawyer’s suggestion, is of only “marginal value.” Rather, such a revelation could, in some cases, critically alter the finder of fact’s assessment of the expert’s testimony. While the majority would (I think) concede that the expert could be cross-examined concerning the attorney’s role in the development of this theory, 7 see Fed.R.Evid. 705, the issue here is whether (and when) extrinsic evidence can be used to impeach an economist who denies or minimizes the lawyer’s role in “shaping” his “expert” findings.

The majority would apparently exclude the extrinsic evidence under almost all circumstances. As I read his opinion, Judge Van Artsdalen apparently would have admitted it under almost all circumstances. I would take a more flexible and more particularized approach. Specifically, I would suggest that the judge inspect the documents in camera and decide whether their impeachment value in the particular case would significantly outweigh the chill on development of legitimate attorney work product that would admittedly accompany disclosure. As I have acknowledged, this latter consideration is entitled to great weight. Ordinarily this inspection would take place at an in limine hearing before trial; 8 under the unusual circumstances of this case, it could take place either before *599 or at the economist’s deposition. 9 In this case, the district court undertook no such inspection and made no such balancing. Instead, it simply declared there to be a need for the material to facilitate cross-examination and thereupon ordered production. This order amounted to clear legal error and requires the writ of mandamus to issue.

While the test I have proposed admittedly suffers the defects of balancing incommensurables, it is at least sensitive' to the interests at stake. The majority’s holding, however, is in significant and unacknowledged tension with contrary authority from impressive sources discussing the relationship between work product doctrine and Federal Rule of Evidence 612, which allows production of material used to refresh a witness’ recollection. In Berkey Photo, Inc. v. Eastman Kodak Co., 74 F.R.D. 613 (S.D.N.Y.1977), Judge Frankel held that whenever core work product is shown by counsel to a witness in preparation for trial, the protection of the work product doctrine 10 is waived. 11 Judge Schwartz held similarly in his excellent opinion in James Julian, Inc. v. Raytheon Co., 93 F.R.D. 138 (D.Del.1982). See also Wheeling-Pittsburgh, Inc. v. Underwriters Laboratories, Inc., 81 F.R.D. 8 (N.D.Ill.1978). Academic commentators have also recognized that work product may be significantly less protected when its content becomes a subject of testimony. See Note, Interactions Between Memory Refreshment and Work Product Protections Under the Federal Rules, 88 Yale L.J. 390 , 404-06 (1978). While these cases do not deal with the precise issue that is faced in this case — the collision of Rule 705 requiring disclosure of the basis for an expert’s opinion and the work product doctrine — they seem similar enough to render the majority’s confident tone somewhat suspect. I would not fully apply the holdings of these cases to the situation at bar — even assuming Berkey and James Julian are right, there is a difference between use of work product to refresh memory and to create beliefs in the first plaee. I believe, however, that the interests those courts focused upon must guide the district courts here. 12

In summary, I would grant the writ of mandamus in this case not because communications between the lawyer and his expert are absolutely (or almost absolutely) immune from discovery and privileged, but because the district court did not conduct an in camera inspection and make the factual findings needed to determine whether the need for the documents in order to impeach in this case outweighs the general and weighty presumption against disclosure of work product. The majority’s failure to issue the writ may well leave the district court confused as to its charter on remand. I would issue mandamus and advise the district court to conduct an in camera inspection and make findings on these matters. Therefore, I respectfully dissent.

. In particular, I agree that the district court erred in ordering production of the “work product" without apparent consideration of factors set forth in Upjohn Co. v. United States, 449 U.S. 383 , 101 S.Ct. 677 , 66 L.Ed.2d 584 (1981). I also agree with the majority that Rule 26(b)(4) cannot be read as wholly overriding Rule 26(b)(3), and that the conclusion that it does is the linchpin of the district court’s decision.

. Though it might conceivably be so read, I do not understand the majority opinion to be creating an exception to the “clear legal error” rule of our mandamus jurisprudence for cases in which there is an order for production over a litigant’s claim of privilege or immunity.

. Were the majority to issue the writ of mandamus (as its opinion suggests), my opinion would be a concurrence and not a dissent.

. At the very least, the majority has created a problem for the Supreme Court. If the defendants’ petition for certiorari on grounds that as a practical matter they have lost this dispute, the Supreme Court will have to decide whether its certiorari statute, 28 U.S.C. § 1257 (1982), permits it to review the record upon petition from a party that in the most technical sense has won. While I would anticipate that the Supreme Court would respond that its certiorari jurisdiction did not extend so far — it review judgments, not opinions — or would deny certiorari on discretionary grounds, that body has far better things to worry about. Supreme Court review of the issue in this case might also be obtained were the plaintiffs (who technically have lost) to petition for certiorari. I regard this possibility, however, as remote. Indeed, the scenario can be extended. In the event that the district court were to read the entire majority opinion as dicta and, in light of the failure of the court of appeals to issue a writ of mandamus, continued to order discovery of the “core work product,” our denial of mandamus might, by law of the case and our Internal Operating Procedures Chapter VIIIC prevent any subsequent panels of this circuit from granting the new mandamus petition from the plaintiffs that would surely follow. Such a result would be as unfair to the plaintiffs in this litigation as is the possible foreclosure of certiorari to the defendants. While the example is fanciful here, it points up the problem with the majority’s approach.

. What happens, for example, if the defendants revive their attempt to subpoena these documents under Fed.R.Civ.P. 45(d)?

. Though I doubt it, it is, of course, possible that the majority is not ruling at all on either the admissibility of the deposition transcript at trial or on the propriety of a subpoena duces tecum aimed at obtaining these materials for trial. It may simply be ruling that no discovery of the work product may be had now. I suspect, however, the plaintiffs would be astonished with the Pyrrhic victory they would have obtained today were they to find that, because of the inadequate opportunities for cross-examination, the deposition of their expert could not be introduced at trial or to discover that a subpoena duces tecum aimed at their work product would not be quashed.

. I would so hold, as did Judge Van Artsdalen. On the other hand, under the majority’s approach without benefit of any examination of the documents, the adverse parties’ attorney might not have the requisite good-faith basis for cross-examination of the expert on this point.

. The use of pre-trial in limine hearings to resolve difficult evidentiary problems in complex cases was approved in In re Japanese Electronics Products, 723 F.2d 238, 260 (3d Cir.1983).

. The judge may, of course, preside at trial depositions.

. The Supreme Court has been assiduous in terming the protection offered work product as a "doctrine" and as an “immunity," not as a "privilege.” See United States v. Arthur Young & Co., — U.S. -, 104 S.Ct. 1495 , 79 L.Ed.2d 826 (1984); Upjohn Co. v. United States, 449 U.S. 383, 397-402 , 101 S.Ct. 677, 686-89 , 66 L.Ed.2d 584 (1981).

. It is not clear whether these cases proceeded on the theory that the work product doctrine’s protections were waived when the attorney refreshed his witness’ memory, or whether the limited protections of the doctrine were overcome by the need of adverse parties to cross-examine the witness effectively concerning his testimonial capacities. I do not believe this issue matters much, however.

. The majority’s opinion is also in unacknowledged opposition to the decision in Boring v. Keller, 97 F.R.D. 404, 406-08 (D.Colo.1983). Although the Boring court relied in part on two circuit court opinions in reaching its conclusion, United States v. Meyer, 398 F.2d 66 (9th Cir.1968), and United States v. McKay, 372 F.2d 174 (5th Cir.1967) (per Maris, J., sitting by designation), my review of those cases suggests that the issues there were somewhat different. Thus, I do not think our opinion today creates a circuit conflict.