This is an action brought by the United States to enforce an Internal Revenue Service summons issued to Monroe Adlman, as Officer and Representative of Sequa Corporation ("Sequa"). See 26 U.S.C. Sec. 7604 (conferring power upon the district courts to enforce an IRS summons). The summons sought production of a preliminary and final draft of a memorandum prepared by Sequa's auditors, Arthur Andersen & Co. ("AA"), at Adlman's request. Adlman, who is an attorney and Sequa's Vice President for Taxes, refused to produce the memoranda, invoking Sequa's attorney-client privilege and the work product doctrine. Sequa contended, among other things, that the memoranda were prepared in anticipation of litigation with the IRS over a $290 million loss claimed by Sequa in its 1989 tax return, resulting from a corporate reorganization.

By order entered May 16, 1994, the United States District Court for the Southern District of New York, Whitman Knapp, Judge, enforced the summons, ruling that the memoranda were protected by neither attorney-client privilege, nor the work product rule. We affirm the district judge's finding that attorney-client privilege does not apply to the memoranda at issue. As to the applicability of the work product doctrine, we vacate the order and remand for further findings.

Sequa, a large corporation generating annual revenues of nearly $2 billion, is primarily engaged in the business of manufacturing products for the aerospace industry. In 1987, Sequa purchased Atlantic Research Corporation ("ARC"), which develops lightweight composite materials for aircraft. Shortly thereafter, Sequa's management began considering a plan to combine ARC with another of its subsidiary corporations, Chromalloy Gas Turbine Corporation ("Chromalloy"). Chromalloy manufactures and repairs component parts for jet aircraft engines.

In the spring of 1989, Sequa's management assessed the likely tax consequences of the proposed plan to combine ARC and Chromalloy. Adlman consulted with Paul Sheahen, an accountant and partner at AA. AA serves as Sequa's accountant and auditor. Sheahen specializes in evaluating the tax implications of corporate reorganizations. After numerous discussions, Adlman directed Sheahen to prepare a draft memorandum discussing the probable tax consequences of the proposed restructuring. Sheahen presented the draft memo to Adlman in August of 1989. After making several changes to address questions raised by Adlman, Sheahen sent Adlman the memorandum in final form on September 5, 1989. The final draft is 58 pages long and contains a detailed technical analysis of the tax implications of the proposed reorganization. The preliminary and final versions of this memorandum are the subject of this appeal.

On September 7, 1989, two days after submitting the final memo to Adlman, Sheahen wrote to Gerald Gutterman, Sequa's Executive Vice-President in charge of Finance and Administration, summarizing "our recommendations and conclusions with respect to the proposed plan of restructuring." The letter stated that "[t]he relevant tax implications ... have been reviewed with and approved by Monroe Adlman." The letter also contained AA's recommendation as to the form the transaction should take--urging Sequa to sell a substantial majority interest in ARC to Chromalloy, recapitalize the remaining shares as preferred stock, and then sell the newly preferred stock to a third party. AA advised Sequa that it should retain an investment banker to assist in valuing ARC and recapitalizing and selling the retained shares. AA "strongly urged" Sequa's management to document all relevant business considerations motivating the reorganization. Finally, the letter offered "to discuss all related matters with a view toward ... prompt and successful implementation [of the restructuring plan]."

The transaction occurred in essentially the manner recommended by AA and was consummated by the end of 1989. On December 12, 1989, Sequa sold 93% of its holding in ARC to Chromalloy for $167.4 million. Six days later, ARC amended its certificate of incorporation to create two classes of stock. One class, which consisted entirely of the 7% holding retained by Sequa, was given substantially increased rights. The next day, on December 19, 1989, Sequa sold this remaining 7% stake to Bankers Trust New York Corporation for $12.6 million.

AA formalized its analysis of the transaction in two opinion letters dated April 20, 1990, and addressed to Adlman. In these letters, AA evaluated the reorganization in light of the federal tax code and concluded that the reorganization should result in a substantial capital loss for Sequa. On its 1989 tax return, Sequa claimed a tax loss of approximately $290 million from the transaction. Part of that loss was carried back to offset millions of dollars of Sequa's capital gains from 1986, thereby generating a large tax refund.

The IRS undertook an audit of Sequa's tax returns from 1986 to 1989. The IRS sent numerous informal information document requests for materials related to the restructuring. Sequa acknowledged the existence of the draft and final versions of the AA memo, but claimed that they were protected from disclosure by the attorney-client privilege and the work product doctrine. On September 23, 1993, the IRS served a summons on Adlman to compel disclosure of the AA memos. Adlman again refused to provide them on grounds of privilege and work product.

On February 3, 1994, the government filed a petition in the Southern District of New York to enforce the summons. See 26 U.S.C. Secs. 7402(b), 7604(a) (providing for district court jurisdiction to enforce an IRS summons). Both parties submitted affidavits and exhibits. The district court conducted a hearing at which it heard argument and accepted the AA memoranda for in camera review.

Sequa's proof before the district court consisted mainly of affidavits of Adlman, Sheahen, Gutterman, and Norman Alexander, Sequa's Chief Executive Officer. Adlman stated that he is an attorney and "in charge of advising management of the consequences under the tax laws of a particular transaction"; that Sequa's management had asked him for advice on the tax consequences of combining ARC and Chromalloy; and that, "[b]ecause I was not expert in the highly complex corporate reorganization provisions [of the tax code], I needed Mr. Sheahen to interpret them for me." According to Adlman, "Mr. Sheahen and I both knew that the AA Memo was prepared to assist me in rendering my advice to Sequa's management and that, because I was acting as a lawyer, the AA Memo was private and confidential." Adlman claims that, after reviewing the final AA memo, he was able to come to his own conclusions regarding the tax consequences of the reorganization, and that he advised Sequa's management about the transaction in several meetings.

Adlman also asserted that he had "no doubt that Sequa would end up in litigation with the IRS" over the reorganization. Adlman claimed three reasons for this belief. First, he believed that the transaction would not escape the attention of the IRS because the IRS audited Sequa every year. Second, the large size of the claimed tax loss would require review by the Congressional Joint Committee on Taxation. Finally, Adlman anticipated litigation because "there was no case or ruling directly on point."

The other Sequa affiants corroborate Adlman's version of the events, each attesting to Sequa's reliance on Adlman for legal advice and to the common understanding that the AA memo was to be treated as confidential. Sheahen's affidavit similarly asserted that Adlman "hired Arthur Andersen, and me in particular, to interpret and help him understand the complex tax provisions affecting corporate reorganizations." Sheahen "knew that Mr. Adlman was using the AA Memo to help him evaluate the Transaction and to assist him in giving legal advice to Sequa's management." Sheahen therefore "assumed" that the AA memo was private and confidential.

Gutterman stated that he relied on Adlman to analyze the tax consequences of the reorganization and advise Sequa's management. He claimed never to have read the AA memorandum, instead relying on Adlman to evaluate AA's analysis and to determine whether the transaction would have favorable tax consequences. Gutterman further stated that Sequa would not have gone forward with the deal if Adlman had not advised management that the transaction was appropriate under the tax laws. Similarly, Alexander stated in his affidavit that he understood that Adlman was acting as Sequa's lawyer and that he considered all discussions regarding the transaction to be confidential. He too stated that he would have not approved the reorganization without Adlman's advice.

Thus, Sequa contended that AA's advice came within the attorney-client privilege under United States v. Kovel, 296 F.2d 918 (2d Cir.1961), because it was rendered to Adlman to assist him in giving legal advice to his client Sequa.

The government characterized the relationship between Adlman, Sheahen, and Sequa quite differently. It contended that the AA memos were tax advice given to Sequa as part of AA's larger role as a consultant to Sequa's management. The government argued that AA acted as a tax and business advisor to Sequa's management, and not merely as an accountant assisting Adlman in rendering legal advice. Gov't Mem. in Support of Petition at 16. The government presented evidence that AA rendered various and extensive services to Sequa in connection with the reorganization, and that these were billed to Sequa without differentiation.

The district court held that the AA memos were not covered by the attorney-client privilege. The court found that "the objective evidence contradicts [Sequa's] assertion of privilege." Noting that Sequa had not produced any contemporaneous evidence, such as a separate retainer agreement, distinguishing Sheahen's work on the memos from other work, the court concluded that Sequa had failed to meet its burden of establishing the attorney-client privilege.

The district court also rejected Sequa's claim of work product protection. It stated that "[t]he possibility of future litigation based on a transaction which has not yet occurred ... fails to establish that the work-product doctrine applies to a communication." The court accordingly ordered production of the AA memoranda in response to the IRS summons.

We review district court rulings on claims of attorney-client privilege and work product for abuse of discretion. See Cruden v. Bank of New York, 957 F.2d 961, 972 (2nd Cir.1992) (attorney-client privilege); Horn & Hardart Co. v. Pillsbury Co., 888 F.2d 8, 12 (2nd Cir.1989) (work product). The district court's factual findings are reviewed for clear error. See Petereit v. S.B. Thomas, Inc., 63 F.3d 1169, 1176 (2d Cir.1995).

Sequa contends that the district court erred in finding it had not established a privilege under Kovel. The attorney-client privilege forbids an attorney from disclosing confidential communications obtained from the client during the course of professional consultations. United States v. Schwimmer, 892 F.2d 237, 243 (2d Cir.1989). The privilege is intended to encourage clients to be forthcoming and candid with their attorneys so that the attorney is sufficiently well-informed to provide sound legal advice. See Upjohn Co. v. United States, 449 U.S. 383, 389, 101 S. Ct. 677, 682, 66 L. Ed. 2d 584 (1981); United States v. Bilzerian, 926 F.2d 1285, 1292 (2d Cir.), cert. denied, 502 U.S. 813, 112 S. Ct. 63, 116 L. Ed. 2d 39 (1991). The privilege applies only to communications between lawyer and client; in general, communications between accountants and their clients enjoy no privilege. See United States v. Arthur Young & Co., 465 U.S. 805, 817, 104 S. Ct. 1495, 1502-03, 79 L. Ed. 2d 826 (1984); United States v. Bein, 728 F.2d 107, 112 (2d Cir.), cert. denied, 469 U.S. 837, 105 S. Ct. 135, 83 L. Ed. 2d 75 (1984).

Under certain circumstances, however, the privilege for communication with attorneys can extend to shield communications to others when the purpose of the communication is to assist the attorney in rendering advice to the client. Thus, in Kovel, 296 F.2d at 918, we recognized that the privilege would extend to communications by an attorney's client to an accountant hired by the attorney to assist the attorney in understanding the client's financial information. Recognizing that the privilege would surely apply where a client who spoke only a foreign language furnished his confidential information to an interpreter employed by the attorney to translate for the attorney's benefit, Judge Friendly observed that accounting concepts can be as incomprehensible as a foreign language to attorneys. "Hence the presence of an accountant ... while the client is relating a complicated tax story to the lawyer, ought not destroy the privilege...." Id. at 922. "What is vital to the privilege is that the communication be made in confidence for the purpose of obtaining legal advice from the lawyer. If what is sought is not legal advice but only accounting service ... or if the advice sought is the accountant's rather than the lawyer's, no privilege exists." Id. (emphasis in original).

Adlman contends that his communications with AA fall within the principle of Kovel. He asserts that Sequa imparted its information to him in confidence for the purpose of receiving his legal advice on the tax implications of its contemplated transactions, and that he, in turn, hired Sheahen to help him understand the problem and shared the client's information with the accountant for that limited purpose.

The party claiming the benefit of the attorney-client privilege has the burden of establishing all the essential elements. von Bulow by Auersperg v. von Bulow, 811 F.2d 136, 144 (2d Cir.), cert. denied, 481 U.S. 1015, 107 S. Ct. 1891, 95 L. Ed. 2d 498 (1987). We cannot say that the district court abused its discretion or made clearly erroneous findings when it concluded that Sequa had failed to demonstrate that it came within the principle of Kovel. The problem with Adlman's argument is that the facts are subject to competing interpretations. In many respects, the evidence supports the conclusion that Sequa consulted an accounting firm for tax advice, rather than that Adlman, as Sequa's counsel, consulted AA to help him reach the understanding he needed to furnish legal advice. Adlman himself serves not only as counsel to Sequa but also as one of its officers. AA, furthermore, is regularly employed by Sequa to furnish auditing, accounting, and advisory services. AA furnished extensive advisory services to Sequa in connection with the merger, including tax advice. If the facts were that Sequa furnished information to AA to seek AA's expert advice on the tax implications of the proposed transaction, no privilege would apply.

There is virtually no contemporaneous documentation supporting the view that AA, in this task alone, was working under a different arrangement from that which governed the rest of its work for Sequa. As noted, AA's billing statements lump the work done in this consultation together with its other accounting and advisory services to Sequa. By his own admission, Adlman lacked the expertise necessary to assess the tax implications of corporate reorganizations. It was AA, not Adlman, that prepared all the written analyses of the tax consequences of the transaction. These included not only the memos in question, but also two formal opinion letters dated April 20, 1990. Furthermore, by its letter dated September 7, 1989, AA sent a summary of its recommendations and conclusions directly to Sequa's management, offering "to discuss all related matters with a view toward ... prompt successful implementation [of the restructuring plan]." All of these factors favored the IRS contentions as to the proper understanding of the transaction. Sequa's interpretation was suggested primarily by litigation affidavits prepared by interested persons four years after the fact and lacking any support in contemporaneous documentation. The district court did not abuse its discretion or make clearly erroneous findings when it concluded that Sequa had failed to sustain its burden of showing that the facts come within the principle of Kovel.1

Sequa contends the AA memoranda are shielded from discovery by the "work product" rule originally laid down in Hickman v. Taylor, 329 U.S. 495, 514, 67 S. Ct. 385, 395, 91 L. Ed. 451 (1947), and subsequently codified in Fed.R.Civ.P. 26(b)(3). The work-product rule shields from disclosure materials prepared "in anticipation of litigation" by a party, or the party's representative, absent a showing of substantial need. Fed.R.Civ.P. 26(b)(3). The purpose of the doctrine is to establish a zone of privacy for strategic litigation planning and to prevent one party from piggybacking on the adversary's preparation. See United States v. Nobles, 422 U.S. 225, 238, 95 S. Ct. 2160, 2170, 45 L. Ed. 2d 141 (1975) ("At its core, the work-product doctrine shelters the mental processes of the attorney, providing a privileged area within which he can analyze and prepare his client's case."); Hickman v. Taylor, 329 U.S. 495, 516, 67 S. Ct. 385, 396, 91 L. Ed. 451 (1947) (Jackson, J., concurring) (work product rule intended to insure that one side does not "perform its functions ... on wits borrowed from the adversary"); Restatement of the Law Governing Lawyers, Chap. 5 at 1 (Tent.Draft No. 6 1993) ("the doctrine seeks to preserve a zone of privacy in which a lawyer can work free from intrusion by opposing counsel").

Sequa contends it caused the memoranda to be written "in anticipation of litigation" with the IRS over Sequa's claimed tax loss. The district judge rejected Sequa's claim of entitlement to work-product protection for the sole reason that at the time of AA's creation of the memoranda, neither the litigation nor the events giving rise to it (the proposed merger) had yet occurred.

Although the non-occurrence of the events giving rise to the anticipated litigation is a factor that can argue against application of the work product doctrine, especially when the expected litigation is merely a vague abstract possibility without precise form, see Gould Inc. v. Mitsui Mining & Smelting Co., Ltd., 825 F.2d 676, 680 (2d Cir.1987) (work product rule requires the "existence of a real, rather than speculative, concern"), there is no rule that bars application of work product protection to documents created prior to the event giving rise to litigation. Nor do we see any reason for such a limitation. In many instances, the expected litigation is quite concrete, notwithstanding that the events giving rise to it have not yet occurred. If, for example, a party expects to be sued by a particular adverse claimant in the event it undertakes to trade under a disputed mark or publishes a book the copyright of which is contested, we see no reason why work-product protection should not apply to preparatory litigation studies undertaken by that party (or its representatives) before it begins to trade under the contested mark or publishes the book. Nor should the answer be different for a party that prepares to initiate the lawsuit in the expectation that another is about to begin trade under the party's mark, or about to publish a work to which the party claims the copyright.

The district judge relied on the authority of two cases which, in our view, do not support his view of the law. Neither case denied work product protection because the actionable events had yet to occur. In National Union Fire Ins. Co. v. Murray Sheet Metal Co., Inc., 967 F.2d 980 (4th Cir.1992), it is true that the Fourth Circuit wrote a passage apparently conveying that message: "[in order for work product protection to apply], [t]he document must be prepared because of the prospect of litigation when the preparer faces an actual claim or a potential claim following an actual event or series of events that reasonably could result in litigation." Id. at 984. Given the context, however, it appears that the court's real message was to stress the necessary causal relationship between the anticipated litigation and the creation of the document, rather than a requirement that the litigation-causing events have already occurred. The dispute involved no issue of distinguishing between documents prepared before and after the litigation causing events. The event in National Union was a fire; the anticipated litigation involved fire insurance claims; the fire had occurred prior to the preparation of the documents sought to be protected. The court's statement that the document must have been prepared "following an actual event" was pure dictum.

The other case cited by the district court, SCM Corp. v. Xerox Corp., 70 F.R.D. 508 (D.Conn.), appeal dismissed, 534 F.2d 1031 (2d Cir.1976), also provides little support for the district court's position. Plaintiff claimed the defendant had violated the antitrust laws through, among other things, the acquisition of patents, and sought discovery of memoranda prepared by attorneys in defendant's patent department. The district court denied defendant's claim of work-product protection finding no showing that the documents were prepared in anticipation of litigation. Then-District Judge Newman wrote, "A specific claim must have arisen to make the prospect of litigation identifiable in order for the work product rule to apply." Id. at 515. We read this language as emphasizing the need for focus on a specific claim, rather than on the abstract possibility that the patent may someday become the focus of some as-yet unidentified claim. We do not construe the court's statement that "a specific claim must have arisen" to require that the actionable facts have occurred, as opposed to being immediately contemplated. Cf. Delaney, Migdail & Young v. I.R.S., 826 F.2d 124, 126-27 (D.C.Cir.1987).

We conclude that the district court barred work-product protection on the basis of an incorrect standard. We must therefore remand for a determination whether the protection of Rule 26(b)(3) should apply. Our vacating the district court's ruling should not be understood to imply that work-product protection is applicable--only that the particular reason for denying it was invalid. The district court will need to consider whether, within the meaning of the rule, the AA memoranda were "prepared in anticipation of litigation," and if so, whether the government has shown "substantial need of the materials." Fed.R.Civ.P. 26(b)(3). If discovery is warranted on the basis of these showings, the district court "shall protect against the disclosure of the mental impressions, conclusions, opinions, or legal theories of an attorney or other representative of a party concerning the litigation." Id.

We affirm the district court's order insofar as it denied Sequa's claim of attorney-client privilege. We vacate the order enforcing the summons and remand for further consideration of Sequa's claim of work product protection.