The much-anticipated decision of the First Circuit Court of Appeals in the Textron case, involving the application of the work-product privilege to tax accrual workpapers, finally came down last week. See United States v. Textron, Inc., 2009 U.S. App. LEXIS 18103 (1^st Cir. 2009). The bad news is that the court held that the privilege did not apply. The good news is that the majority opinion is so poorly reasoned—and the dissent is so persuasive—that the case will probably carry little weight outside the First Circuit.

Background

The underlying facts are as follows. In 2001, Textron participated in nine sale-in, lease-out (SILO) transactions, which the IRS considers to be abusive "listed transactions." Textron's in-house tax attorneys analyzed the company's tax positions and, together with its in-house CPAs, used the analyses to prepare tax accrual workpapers. The workpapers included a spreadsheet listing the issues, percentage estimates of the hazards of litigation, and a tax reserve amount. Textron disclosed the tax accrual workpapers to its auditors (Ernst & Young) with the understanding that they were confidential. It did not allow E&Y to retain copies.

The IRS commenced an audit of Textron and became aware of the SILO transactions. IRS policy is to: (1) request tax accrual workpapers specifically related to a taxpayer's participation in a listed transaction if the taxpayer engaged in a single listed transaction and properly disclosed its participation; and (2) request all tax accrual workpapers if the taxpayer has engaged in multiple listed transactions or in a single listed transaction that was not properly disclosed. Because Textron engaged in nine SILOs, the IRS issued summonses for all Textron's tax accrual workpapers for 2001. Textron declined to produce them, and the IRS then moved to enforce the summons.

Textron Asserts Privilege

In its defense, Textron asserted the following: (1) the attorney-client privilege; (2) the Section 7525 privilege (which in limited circumstances applies to accountant communications); and (3) the work-product privilege. The district court easily disposed of the first two claims because Textron's disclosure to E&Y was a waiver. However, the district court correctly found that that disclosure was not a waiver of the work-product privilege. Only disclosures that are inconsistent with the adversary system are deemed waivers of the work-product privilege. Typically, this means disclosure to an adversary. E&Y was not an adversary.

The work-product privilege is defined by Rule 26(b)(3) of the Federal Rules of Civil Procedure and applies to documents "prepared in anticipation of litigation or for trial." The privilege is a "qualified" one, which means that an opponent can have the privilege set aside by a showing of "substantial need" and an inability "without undue hardship to obtain the substantial equivalent" by other means. However, "mental conclusions, opinions, or legal theories of an attorney or other representative of a party" are generally protected as core work product.

Prepared in Anticipation of Litigation

What does it mean for a document to be "prepared in anticipation of litigation or for trial" if, at the time of preparation, there is no ongoing litigation? Most courts hold that the privilege applies if a document was prepared because of the prospect of litigation, even if there were other purposes. However, the Fifth Circuit (in a 1982 case involving tax accrual workpapers) held that the privilege applies only if the documents are "primarily motivated to assist in future litigation." United States v. El Paso Co., 682 F.2d 530, 542 (5th Cir. 1982), cert. denied, 466 U.S. 944 (1984). No other circuit follows the Fifth Circuit rule.

In Textron, the lower court applied the "because of" test and concluded that the tax accrual workpapers were privileged:

On appeal, the IRS argued that the mere presence of a business or regulatory purpose defeats work-product protection—in other words, it argued that the tax accrual workpapers were not prepared in anticipation of litigation because they were prepared to evaluate uncertain tax positions for financial accounting reasons. Initially, the First Circuit rejected this IRS argument and affirmed the district court's holding that documents could have a dual purpose. After all, as the court noted, they would not have been prepared for financial accounting purposes if there had been no prospect of litigation. However, that decision was vacated and the case was heard again by all five First Circuit judges.

The First Circuit Adopts a "For Use" Test

Following the rehearing, the majority issued an opinion that purported to apply the "because of" test. However, as the dissent observed, the court actually adopted a new test that is even more restrictive than the Fifth Circuit's "primary purpose" test. Under the majority holding, a in order to qualify as work product a document must be prepared for use in possible litigation. The majority then observed:

The dissent, quite accurately, compared this "touch and feel" approach to Justice Potter Stewart's "famously unhelpful" test for identifying obscenity: "I know it when I see it."

One problem with the majority's analysis is that it unduly limits "litigation" to court trials. Other courts interpret the term more broadly to include adversarial proceedings before an administrative agency, as did the dissent in Textron. The dissent correctly noted that IRS audits can be adversarial. Indeed, that was especially true in Textron which involved listed transactions. To paraphrase the majority, anyone with experience of an IRS examination of a listed transaction would understand that the examination is simply the initial IRS discovery phase in a lengthy litigation process. After all, unless the taxpayer agrees to the often non-negotiable terms of an IRS global settlement offer (if there is one), the taxpayer will have to litigate.

Another problem with the majority's analysis is that the text of the rule explicitly covers more than just materials prepared "for trial;" it also covers material "prepared in anticipation of litigation." The First Circuit's "for use" test does not comport with the language of the rule.

Favorable Law in Other Circuits

The majority's "for use" test is flatly contrary to cases in other circuits, in particular the Second Circuit's decision in Adlman, where the court said:

United States v. Adlman, 134 F.3d 1194, 1198-99 (2d Cir. 1998). Adlman held that a memorandum discussing various legal issues relating to taxation and predicting the outcome of possible litigation that was prepared in connection with the sale of a subsidiary could qualify as work product.

Textron is also contrary to the Roxworthy case. U.S. v. Roxworthy, 457 F.3d 590 (6th Cir. 2006). The court there held that two "more likely than not" memoranda prepared by KPMG regarding the creation of a captive insurance company and related stock transfers were work product. The "anticipation of litigation" requirement was met because the taxpayer (Yum Brands) had an objectively reasonable belief there could be litigation, based on the fact that it was audited annually, the dollars at issue were significant, and there was a large discrepancy between a tax loss and a book loss. See also Regions Financial Corp. v. United States, 2008 WL 2139008 (N.D. Ala. 2008) (tax accrual workpapers held to be protected by the work-product privilege).

Textron also is at odds with several non-tax cases that have involved loss reserves. For example, in an Eighth Circuit case, a corporation's in-house lawyers established individual loss reserves (including settlement value) after notice of a claim or suit. Its risk management department then aggregated those reserves for a variety of business purposes. The court held that the aggregate numbers were not protected, but that the individual case reserves "reveal the mental impressions, thoughts, and conclusions of an attorney in evaluating a legal claim" and thus were core work product. See Simon v. G.D. Searle & Co., 816 F.2d 397 (8^th Cir. 1987).

Are Tax Cases Different?

Textron, of course, was a tax case, and this made all the difference to the majority which, as the dissent noted, engaged in "outcome determinative reasoning." The majority made no attempt to hide its partiality. It noted that the IRS has not automatically requested tax accrual workpapers, but has been driven to it "in the wake of Enron." It rejected Textron's argument that providing an adversary with evaluations of its legal positions is unfair with the bromide that "tax collection is not a game." It observed that under-reporting of corporate taxes "is likely endemic." It noted that the IRS asked for the workpapers "only after finding a specific type of transaction that had been shown to be abused by taxpayers," notwithstanding that only one SILO case (decided after the summons in Textron was issue) has been decided to date. The dissent chided the majority for its favoritism, stating: "The scope of the work-product doctrine should not depend on what party is asserting it."

What Did the IRS Really Want?

The dissent also noted that "under the majority's rule one party in a litigation will be able to discover an opposing party's analysis of the business risks of the instant litigation, including the amount of money set aside in a litigation reserve fund, created in accordance with similar requirements as Textron's tax reserve fund." Courts in non-tax cases have likewise noted "the need for protection against forced revelation of a party's evaluation of his case," and that "it would be an intolerable intrusion on the bargaining process to allow one party to take advantage of the other's assessment of his prospects." Rhone-Poulenc Rorer Inc. v. The Home Indemnity Co., 139 F.R.D. 609 (E.D. Pa. 1991).

However, in Textron, given the hard-line settlement posture the IRS has taken in SILO cases, it seems unlikely that the IRS was primarily motivated to understand the taxpayer's settlement strategy. Why, then, was the IRS so interested in obtaining reserve calculations that will not help it to understand the transactions and will not provide a roadmap to Textron's legal theories? The most likely answer is that the IRS was hoping to uncover estimates of loss at a high enough level that they can be used to prejudice a judge or jury in a future trial. Such tactics can be effective. Indeed, the majority in Textron noted that some of the "spreadsheet entries estimated the probability of IRS success at 100 percent," which it evidently took as an indication that Textron was flouting the tax law. The majority ignored Textron's briefs, which explained that the 100 percent estimates were made after its returns were filed to indicate that a position should be conceded as a result of post-return adverse authority (or concessions in a prior audit).

Textron Will Not Be the Last Word

Many observers expected Textron to become a landmark decision clarifying the law of work-product privilege for tax accrual workpapers. However, it is anything but that. The dissent got it exactly right when it said:

Textron is a case that deserves Supreme Court review; however, the Court rarely agrees to decide tax cases. Nonetheless, even if Textron stands as the last word in the First Circuit, taxpayers outside of that circuit should not view it as in any way definitive. The First Circuit's result-oriented decision is out of step with the law and will likely be considered an anomaly by other courts.