The United States and two agents of the Internal Revenue Service (IRS) petitioned the district court to enforce two summonses issued to the El Paso Company (El Paso) with regard to a tax audit. One summons sought El Paso’s “tax-pool analysis” — a summary of El Paso’s contingent liability for additional taxes should it ultimately be determined that El Paso owed more taxes than indicated on its return. After a hearing, the district court enforced the tax pool analysis summons and El Paso brought this appeal. We affirm.
*533I.
El Paso is the holding company for several large corporations, including 67 subsidiaries. The principal operations in the El Paso conglomerate are the El Paso Natural Gas Company, the El Paso Products Company, and the El Paso LNG Company. All are based in Texas. Because of the size and variety of El Paso’s business, the calculation of El Paso’s tax liability is an immense task. To prepare a return for a single year, El Paso’s in-house staff expends over 10,000 hours.
The IRS, as it does with most giant corporations, annually audits El Paso and has been doing so since the 1940s. To perform the audit, the IRS assembles a team of revenue agents. Even with a team specifically assigned to El Paso, however, the IRS has time to review and inspect only a small sample of the documents that underlie El Paso’s returns. The audit of El Paso that triggered the summons in this case was not conducted on suspicion of fraud; rather, it was a routine audit occasioned by the amount of taxes at stake.
Each audit of El Paso covers several years of tax returns. The relevant cycle in this case includes the years 1976, 1977, and 1978. Early in the audit of the 1976-1978 cycle, the IRS team coordinator delivered a document request to El Paso for “all analy-ses prepared by the El Paso Company regarding potential tax liabilities and tax problems” for the years covered by the cycle. Five days later, Jack McCarthy, the head of El Paso’s tax department, returned the request form marked “refused”.1
Shortly after receiving El Paso’s refusal to respond to the potential tax-liabilities document request, the IRS issued a summons to McCarthy covering “any document, memorandum, letter, or work papers which identify potential tax liabilities or tax problems for the period beginning January 1,1976 and ending December 31, 1978, inclusive.” McCarthy also declined to comply with the summons, stating his reasons in a letter written to the case manager of the El Paso audit.
On May 27,1981, the IRS filed its petition to enforce the summons under 26 U.S.C. § 7604.2 El Paso defended against enforcement on grounds of burdensomeness, relevance, attorney-client privilege, and work product doctrine. After a hearing at which both sides were permitted to call and cross-examine witnesses, the district court rejected El Paso’s defenses and concluded that El Paso must comply with the summons.
El Paso appealed and sought a stay of the district court’s judgment from this Court. Because of the importance and the novelty of the issues raised with respect to the IRS’s summons powers, we granted the stay pending our resolution of the case.
II.
This appeal is centrally concerned with documents known to the accounting profession under various names — the noncurrent tax account, the tax accrual work papers, and the tax pool analysis. Because the nomenclature is not standardized, the IRS chose to request El Paso’s version of these documents under a loose descriptive label— documents analyzing potential tax liabilities or tax problems. No matter what alias is used, however, the documents are of similar nature. It is useful to explain what these documents are before proceeding to determine whether the IRS may have access to them.
*534The income tax laws, as every citizen knows, are far from a model of clarity. Written to accommodate a multitude of competing policies and differing situations, the Internal Revenue Code is a sprawling-tapestry of almost infinite complexity. Its details and intricate provisions have fostered a wealth of interpretations. To thread one’s way through this maze, the business or wealthy taxpayer needs the mind of a Talmudist and the patience of Job.
Even endowed with these qualities, however, no taxpayer completes a return with the certainty that the IRS will agree with the bottom line, or the many steps taken to get there. There is no tax oracle one may consult to learn how a return will fare under the scrutiny of the revenue agents and the courts. The Code, after all, is a finite system of rules designed to apply flexibility to an infinite variety of situations. There are many “gray areas” in the tax world, twilight zones in which one may only dimly perceive how properly to treat a given accretion to wealth or given expenditure of funds.
When a large corporation like El Paso completes its return, the number of decisions in the gray areas is enormous. To characterize a sale as ordinary income or capital gain, to depreciate equipment over ten years or twenty, to attribute a transaction to this year or to the next: these decisions recur over and over in a course of preparing a return and guarantee that a large corporation has many opportunities to choose in good faith an interpretation of the tax code that leans toward lessening its taxes. The return is filed with the understanding, however, that the IRS may challenge some of these questionable positions and, through settlement or litigation, the corporation may end up owing more taxes than it initially acknowledged.
Business reality compels corporations to recognize on their financial sheets that the return as filed is not the last word in determining the taxes owed. Public companies subject to the securities laws must file financial statements with the Securities and Exchange Commission (SEC), 15 U.S.C. § 181. SEC regulations require that independent accountants verify these financial statements in accord with Generally Accepted Auditing Principles, 17 C.F.R. § 210.1-02(d). To demonstrate to the accountant that a balance sheet does not portray an overly-rosy view of a corporation’s financial health, the balance sheet must provide for contingent future tax liabilities. In short, the corporation must set aside an account to cover additional taxes that it may become liable to pay above and beyond the amount indicated on the initial return.
To comply with the securities laws, therefore, companies such as El Paso must prepare in-house or have prepared by outside auditors an analysis of their contingent tax liabilities.3 The analysis pinpoints the soft spots on the corporation’s tax returns and indicates those areas in which the taxpayer has taken a position that may, upon challenge, negotiation, or litigation, require the payment of more taxes. The analysis is known in the trade as the tax pool analysis, the noncurrent tax account, or tax accrual work papers.
Several points are worth noting about the tax pool analysis. First, it is not prepared to assist in filling out a tax return. The tax pool analysis is undertaken only after the return is filed. Although the same corporate employees or outside auditors may prepare both the return and the tax pool analysis, there is no necessary connection between the two jobs.
Second, the tax pool analysis may be performed either in house or by outside accountants. El Paso happens to prepare its own tax pool analysis, but many firms retain outside accountants to accomplish that task. See United States v. Arthur Young & Co., 677 F.2d 211 (2d Cir. 1982). It is not *535essential to consult an attorney to prepare the tax pool analysis.
Finally, the tax pool analysis is “[prepared for financial reporting purposes alone.... ” Caplin, IRS Toughens Its Stance on Summoning Accountant’s Tax Accrual Workpapers, J. Tax 130 (Sept. 1980). While the analysis must forecast the cumulative results of IRS audit, settlement, and litigation, the tax pool analysis itself is not prepared to respond to a specific charge by the IRS or to any pending or impending lawsuit. The tax pool analysis is undertaken solely to insure that the corporation sets aside on its balance sheet a sufficient amount to cover contingent tax liability.
III.
Before we turn to the objections raised by El Paso to the enforcement of the “potential tax liabilities” summons, we must determine the scope of the judgment below. El Paso contends that the judgment requires production of much more than the tax pool analysis and backup memoranda and files. In El Paso’s view, the district court’s judgment encompasses all El Paso documents that treat “tax problems.” This category, El Paso tells us, is so broad as to comprise all files in El Paso’s tax department.
El Paso also protests that it relied on IRS representations at trial that the government sought only the tax pool analysis and backup files, and defended the case on that basis. To expand the judgment beyond the tax pool analysis and backup files, El Paso concludes, is unjust.
Finally, El Paso contends that there are no “backup files” to the tax pool analysis. The tax pool analysis consists of four to five accounting pages each year listing items the tax treatment of which might be vulnerable to challenge. A dollar amount is indicated for each item. El Paso denies that specific backup files exist.
We first review the genesis of the judgment’s language. The subject matter of the original IRS summons was documents concerning “potential tax liabilities and tax problems.” According to the IRS, it chose this language because it did not know what name El Paso bestowed on its tax-accrual workpapers file. The district court issued an order to El Paso to show cause why the summons should not be enforced. In answer, El Paso revealed its terminology, at least in part. An affidavit filed by El Paso referred to the “vast majority” of the documents sought by the summons at issue as “documents containing or comprising the ‘tax pool analysis’ or discussing items reflected in the company’s financial statements as income or expense where the ultimate tax treatment was or is unclear.” While written in the disjunctive, this reference actually points to but a single set of documents. The more expansive language in the latter portion of the statement is drawn from the IRS’s own guidelines that define the term tax pool analysis.4
At the summons enforcement hearing, El Paso complained that it was unable to defend against enforcement because the summons lacked sufficient clarity to apprise it of the documents sought. The IRS responded that the El Paso affidavit, quoted above, adequately described what the government wanted. Since El Paso itself had used the term tax pool analysis, the IRS argued, El Paso could not deny that it knew what the IRS was after. The IRS conceded, however, that if the court found that the summons was too broad as written, the court could narrow it to the documents described in the El Paso affidavit.
The district court was understandably perplexed by the inability of the parties to agree on the subject matter of the summons. To give the IRS a chance to clarify what it sought, the court briefly recessed. In the recess, El Paso told the IRS that its *536reserve for contingent tax liability was called “the non-current tax account.” The size of the non-current tax account was determined through the tax pool analysis.
The IRS then reduced to writing a description of the documents it sought:
1. Non-current tax account plus backup.
2. Documents identifying tax problems or potential tax problems created in or referred to the Tax Department or its superiors.
The judgment enforcing the summons substantially tracks this language.
We must interpret the judgment in the context in which its specific language was chosen. The IRS and El Paso wrangled over the scope of the summons. To give shape to its request, the IRS sought to borrow language from an El Paso affidavit. When this language also produced confusion, the IRS turned to yet other clarifying terms. This linguistic pas de deux yielded the language that the judgment now embodies.
The first sentence of the clarification of the summons, we believe, is tolerably clear. The “noncurrent tax account” refers to the four to five page list of the items believed to be vulnerable to challenge. This list is also known as the tax pool analysis. “Backup” refers to the memoranda on which the assessment of vulnerability is based. Despite El Paso’s denial that back-up files exist, El Paso’s witnesses testified that, for many if not all items in the tax pool analysis, supporting memoranda can be found in various “subject” files scattered in the Tax Department. The summons seeks these memoranda in addition to the four to five page list.
The second sentence of the clarification of the summons is considerably less lucid. The language “tax problems or potential tax liabilities” had stimulated El Paso’s original concerns about the breadth of the summons. At trial, the IRS represented that this language could be faithfully recast in the words used in El Paso’s affidavit. The IRS is bound by this representation. The affidavit, as we have seen, only borrows the IRS’s own definition of tax pool analysis. This circle of definitions brings us back to square one. We are unable to give content to the second part of the IRS clarification in any fashion that avoids duplication of the first part.5
We conclude, therefore, that the judgment reaches only the tax pool analysis, as described herein, and the memoranda contained in various subject files that were used to prepare the tax pool analysis. On that understanding of the judgment, we proceed to the merits of this appeal.
IV.
The IRS exercised its summons power in this case under 26 U.S.C. § 7602. That section authorizes the IRS “[t]o examine any books, papers, records, or other data which may be relevant” to determining the correctness of the return. Under the Supreme Court’s seminal decision in United States v. Powell, 379 U.S. 48, 85 S.Ct. 248, 13 L.Ed.2d 112 (1964), the IRS need not show probable cause to obtain judicial enforcement of the summons. Rather, the government need only make a preliminary showing that
“The investigation will be conducted pursuant to a legitimate purpose, that the inquiry may be relevant to the purpose, that the information sought is not already within the Commissioner’s possession, and that the administrative steps required by the Code have been followed. ...”
379 U.S. at 57-58, 85 S.Ct. at 254-55; United States v. Grayson County State Bank, 656 F.2d 1070, 1073 (5th Cir. 1981), cert. denied, - U.S. -, 102 S.Ct. 1276, 71 L.Ed.2d 460 (1982).
As we observed in United States v. Davis, 636 F.2d 1028 (5th Cir. 1981),
Once the government has satisfied this minimal requirement, the burden shifts *537to the summonee either to disprove one of the four elements of the government’s prima facie showing or to demonstrate that judicial enforcement of the summons would otherwise constitute an abuse of the court’s process.
636 F.2d at 1034.
El Paso argues that the tax pool analysis summons is not relevant to any legitimate IRS inquiry. In El Paso’s view, the tax pool analysis represents “a compilation of opinion and speculation on the adequacy of the provision for taxes shown on El Paso’s financial statements.” Containing no source documents with the details of actual transactions, the tax pool analysis is only a forecast of contingent tax liabilities prepared for the company’s financial sheets. El Paso concedes that the IRS has a right to factual information upon which the return is based, but denies that the taxpayer’s predictions about the outcome of an audit have any bearing on the correctness of the return as filed. El Paso relies heavily on United States v. Coopers & Lybrand, 550 F.2d 615 (10th Cir. 1977), in which the court denied the relevance of a tax pool analysis to an IRS investigation.
We reject the view that the tax pool analysis fails the relevance test. An IRS summons is not judged by the relevance standards used in deciding whether to admit evidence in court. Rather, the well-established test is “whether the summons seeks information which ‘might throw light upon the correctness of the taxpayer’s return.’ ” United States v. Wyatt, 637 F.2d 293, 300 (5th Cir. 1981) (upholding the relevance of the so-called “eleven questions” designed to ferret out information on corporate payment of bribes). See also United States v. Noall, 587 F.2d 123, 125 (2d Cir. 1978) cert. denied, 441 U.S. 923, 99 S.Ct. 2031, 60 L.Ed.2d 396 (1979) (upholding the relevance of internal-audit work papers).
The IRS seeks here El Paso’s assessment of the questionable positions it took on its return, an assessment prepared for financial reporting purposes. That the tax pool analysis is neither used in preparing a return nor is a “source document” of an actual transaction does not bar a finding of relevance.6 A tax return is much more than a tabulation of transactions neatly fitted into an existing structure of tax concepts. The return largely represents an interpretation of transactions in the light most favorable to the taxpayer. The task of the Service in determining whether the return is correct is not only to verify the existence of the actual transactions cumulatively recorded in the return. The IRS must also determine whether the taxpayer has stretched a tax concept to cover a transaction to which the concept does not apply. Documents that focus and concentrate the Service’s energy on questionable positions in the return, therefore, are highly relevant. It is, in fact, hard to imagine a document more likely to shed light on the correctness of those aspects of a return that harbor doubts than the tax pool analysis.
Our holding accords with the similar conclusion on the relevance of tax-accrual work papers recently reached by the Second Circuit in United States v. Arthur Young & Co., 677 F.2d 211 (2d Cir. 1982). Facing an argument that tax-accrual work papers prepared by an independent auditor are not relevant, the court succinctly stated “[different tax positions lead to different amounts of liability. It is difficult to say that the assessment by the independent auditor of the correctness of positions taken by the taxpayer in his return would not throw ‘light upon’ the correctness of his return.” 677 F.2d at 219.7 We agree with *538this view and hold that the tax pool analysis is relevant to an IRS audit.
V.
We now turn to El Paso’s contention that the tax pool analysis is shielded by the attorney-client privilege. El Paso carries the burden of establishing this defense to the enforcement of the summons. United States v. Powell, 379 U.S. at 58, 85 S.Ct. at 255; United States v. Davis, 636 F.2d at 1034.
In Upjohn Co. v. United States, 449 U.S. 383, 101 S.Ct. 677, 66 L.Ed.2d 584 (1981), the Supreme Court delineated the purpose of the attorney-client privilege.8
The attorney-client privilege is the oldest of the privileges for confidential communications known to the common law. 8 Wigmore, Evidence § 2290 (McNaughton Rev. 1961). Its purpose is to encourage full and frank communication between attorneys and their clients and thereby promote broader public interests in the observance of law and administration of justice. The privilege recognizes that sound legal advice or advocacy serves public ends and that such advice or advocacy depends upon the lawyer being fully informed by the client. As we stated last term in Trammel v. United States, 445 U.S. 40, 51, 100 S.Ct. 906, 913, 63 L.Ed.2d 186 (1980), “the attorney-client privilege rests on the need for the advocating counselor to know all that relates to the client’s reasons for seeking representation if the professional mission is to be carried out.” And in Fisher v. United States, 425 U.S. 391, 403, 96 S.Ct. 1569, 1577, 48 L.Ed.2d 39 (1976), we recognized the purpose of the privilege to be “to encourage clients to make full disclosures to their attorneys.”
Id. 101 S.Ct. at 682.
The scope of the attorney-client privilege is shaped by its purposes. United States v. Pipkins, 528 F.2d 559, 563 (5th Cir. 1976), cert. denied, 426 U.S. 952, 96 S.Ct. 3177, 49 L.Ed.2d 1191 (1976). While the elements of the privilege have been comprehensively restated elsewhere,9 it is sufficient here to note that “[w]hat is vital to the privilege is that the communication be made in confidence for the purpose of obtaining legal advice from the lawyer." United States v. Kovel, 296 F.2d 918, 922 (2d Cir. 1961) (emphasis in original). See United States v. Miller, 660 F.2d 563 (5th Cir. 1981); Colton v. United States, 306 F.2d 633, 637 (2d Cir. 1962), cert. denied, 371 U.S. 951, 83 S.Ct. 505, 9 L.Ed.2d 499 (1963). Moreover, “disclosure of any significant portion of a confidential communication waives the privilege as to the whole.” United States v. Davis, 636 at 1043 n.18.10 *539Finally, we have made clear that the attorney-client privilege may not be tossed as a blanket over an undifferentiated group of documents. United States v. Davis, 636 F.2d at 1044 n. 20; United States v. Roundtree, 420 F.2d 845, 852 (5th Cir. 1969). The privilege must be specifically asserted with respect to particular documents.
The IRS argues that El Paso has failed to prove any of the elements of the attorney-client privilege. In its broadest theory, the IRS urges that preparing the tax pool analysis is not legal work and that El Paso’s attorneys in performing it are not giving legal advice. In the IRS’s view the tax pool analysis is a business document, drawn up solely to bring the company’s financial statements in line with generally accepted auditing principles. Because the tax pool analysis simply supports a record of the corporation’s finances, the IRS sees it as business work rather than legal work.
The line between accounting work and legal work in the giving of tax advice is extremely difficult to draw. See Kenderdine, The Internal Revenue Service Summons to Produce Documents: Powers, Procedures & Taxpayer Defenses, 64 Minn.L. Rev. 73, 100-01 (1979). We have held that the preparation of tax returns is generally not legal advice within the scope of the privilege. United States v. Davis, 636 F.2d at 1043. See Canaday v. United States, 354 F.2d 849 (8th Cir. 1966); Olender v. United States, 210 F.2d 795, 805-06 (9th Cir. 1954), cert. denied, 352 U.S. 982, 77 S.Ct. 382, 1 L.Ed.2d 365 (1957). Davis withheld the privilege from communications made to an attorney to prepare a tax return because such work is primarily an accounting service. The tax pool analysis may also be considered an accounting service since it is often performed by accountants, see Diss & Hanson, Tax Contingency Audit Work Papers: Current Developments, Observations, and Proposals, Tax Advisor 104 (Feb. 1981). Nevertheless, we would be reluctant to hold that a lawyer’s analysis of the soft spots in a tax return and his judgments on the outcome of litigation on it are not legal advice. We need not decide this issue, however, because we believe that El Paso’s attempt to claim the privilege fails on other grounds.
To retain the attorney-client privilege, the confidentiality surrounding the communications made in that relationship must be preserved. The purpose of the privilege is to foster full client disclosure to the lawyer; the privilege exists to assure the client that his private disclosures will not become common knowledge. The need to cloak these communications with secrecy, however, ends when the secrets pass through the client’s lips to others. Thus, a breach of confidentiality forfeits the client’s right to claim the privilege. Permian Corp. v. United States, 665 F.2d 1214, 1219 (D.C. Cir. 1981) (when confidential communications are voluntarily disclosed to one person, the privilege is waived as to all others); United States v. Miller, supra (previous delivery of accounting books to the IRS robbed them of the confidentiality needed to maintain the attorney-client privilege); McCormick on Evidence § 93 (2d ed. 1972). As Justice Holmes wrote when a member of the Massachusetts Supreme Court,
[T]he privacy for the sake of which the privilege was created was gone by the appellant’s own consent, and the privilege does not remain in such circumstances for the mere sake of giving the client an additional weapon to use or not at his choice.
Green v. Crapo, 181 Mass. 55, 62 N.E. 956, 959 (1902).
The district court found in this case that El Paso discussed “some of the information and many of the potential tax liability issues” in its tax pool analysis with *540the independent auditors who certify the corporation’s books. This finding is not clearly erroneous.11 Pullman-Standard v. Swint, - U.S. -, 102 S.Ct. 1781, 72 L.Ed.2d 66 (1982). The evidence showed that in the course of the audit, the accountants determine whether the company has set aside an adequate reserve for contingent taxes. This task carries the auditors into the tax pool analysis and into at least some of the supporting memoranda. Confidentiality as to these documents is neither expected nor preserved, for they are created with the knowledge that independent accountants may need access to them to complete the audit. See United States v. Pipkins, 528 F.2d at 563.
El Paso generates the tax pool analysis to portray its financial condition accurately. As the securities laws require, independent accountants then verify the financial statements by probing the corporation’s reasons for allocating a given amount to its noncurrent tax account.12 In El Paso’s case, the tax pool analysis is revealed to the independent accountants as part of their audit. Our Circuit does not recognize an accountant-client communications privilege, United States v. Davis, 636 F.2d at 1043; Falsone v. United States, 205 F.2d 734 (5th Cir. 1953), cert. denied, 346 U.S. 864, 74 S.Ct. 103, 98 L.Ed. 375 (1953), and, as the Supreme Court has acknowledged, neither does any other federal court. Couch v. United States, 409 U.S. 322, 335, 93 S.Ct. 611, 619, 34 L.Ed.2d 548 (1973) (“.. . we note that no confidential accountant client privilege exists under federal law, and no state created privilege has been recognized in federal cases....”); 2 J. Weinstein & M. Burger, Weinstein’s Evidence ¶ 503(a)(3)[01] (1981). Under these circumstances, El Paso’s disclosure of the tax pool analysis to the auditors destroys confidentiality with respect to it. With the destruction of confidentiality goes as well the right to claim the attorney-client privilege.
We recognize that the Second Circuit has denied enforcement of IRS summons to the taxpayer’s accountant seeking tax accrual workpapers that the accountant has prepared. United States v. Arthur Young & Co., supra. The logic of Arthur Young implies that the taxpayer’s revelation of tax accrual workpapers to an accountant should be *541considered a communication in confidence.13 Arthur Young does not persuade us, however, that El Paso’s disclosure of its tax pool analysis to its independent accountants is consistent with the confidentiality required to assert the attorney-client privilege. To extend Arthur Young to this case would, in effect, create an accountant-client communications privilege. No such privilege exists under federal law, however, as we noted above. In the absence of a contrary rule established by law, we cannot view El Paso’s discussions with its auditors as confidential. The attorney-client privilege is, therefore, waived.
We believe that El Paso may not withdraw behind the shield of the attorney-client privilege for an additional reason. El Paso failed to particularize its assertion of the privilege and prove its case with respect to any specific document. El Paso made only a blanket assertion of the privilege as to all documents in, or backing up, its tax pool analysis. Such a showing is simply inadequate.
In United States v. Davis, we reiterated the unacceptability of blanket assertions of the attorney-client privilege. 636 F.2d at 1044 n.20. Such assertions disable the court and the adversary party from testing the merits of the claim of privilege. In Davis we relaxed the rule because of confusion about when the attorney-client privilege should be raised in a tax summons enforcement proceeding. We cautioned, however, that “future litigants who make only blanket assertions of privilege at enforcement proceedings should not expect such grace. ’ Id. El Paso had fair notice of the obligation to make its privilege claim precise.14 We cannot excuse its failure to do so.
The reasons for refusing to tolerate blanket assertions of the privilege apply fully to this case. El Paso’s tax department employs eighty accountants and ten attorneys. The tax department as a whole has responsibility for preparing the various memoranda that underpin the tax pool analysis. Both the head of the tax department and the general counsel testified at the enforcement hearing, but neither witness had reviewed the backup memoranda prior to testifying. As a result, neither was able to state how many or which memoranda were prepared by attorneys rather than by accountants.
El Paso’s failure to prove which documents were prepared by attorneys considerably undermines its claim to the attorney-client privilege. Because the privilege does not attach to tax work prepared by accountants unless the accountant is translating complex tax terms into a form intelligible to a lawyer at the lawyer’s behest, United States v. Kovel, 296 F.2d at 922, the memoranda prepared by accountants do not qualify for the privilege. We cannot say that any single memorandum in the El Paso’s tax department is stamped with an attorney’s seal. A general claim that the tax department funnels tax work through its attorneys will not do, and El Paso’s proof amounted to little more.
*542We hold, therefore, that El Paso has breached the confidentiality needed to shield its attorney-client communications, and in any event, has failed to meet its burden to assert the privilege specifically. The attorney-client privilege does not protect El Paso against disclosing the documents that the IRS seeks.
VI.
El. Paso next contends that the work product doctrine immunizes the tax pool analysis from production.15 It is settled, of course, that a work product defense may be asserted against enforcement of an IRS summons. Upjohn Co. v. United States, 101 S.Ct. at 686.
The work product doctrine traces its lineage to the Supreme Court’s decision in Hickman v. Taylor, 329 U.S. 495, 67 S.Ct. 385, 91 L.Ed. 451 (1947). In Hickman, the plaintiff’s attorney sought to discover notes that the defendant’s attorney had taken in interviews of prospective witnesses in preparation for trial. The Court rejected this discovery attempt, which was made without any special showing of need. The Court held that the work of preparing for trial demands insulation from opposing counsel’s inquiries on a lawyer’s research, analysis, legal theories, and mental impressions. To adopt a more lenient approach to discovery of an attorney’s work product would deter lawyers from committing to paper their work in preparation for trial and would thus diminish the quality of representation. Moreover, the invasion of a lawyer’s private thoughts would demoralize the profession. 329 U.S. at 510-512, 67 S.Ct. at 393-94. “An attorney’s thoughts, hithertofore inviolate, would not be his own.” Id. at 511, 67 S.Ct. at 393-94.
The accent in Hickman was on a lawyer’s need for a sphere of privacy in preparing a lawsuit. “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 ease.” United States v. Nobles, 422 U.S. 225, 238, 95 S.Ct. 2160, 2170, 45 L.Ed.2d 141 (1975) (applying the work product doctrine to criminal trials).
The work product doctrine is not an umbrella that shades all materials prepared by a lawyer, however. The work product doctrine focuses only on materials assembled and brought into being in anticipation of litigation. Excluded from work product materials, as the advisory committee notes to Rule 26(b)(3) make clear, are “[m]aterials assembled in the ordinary course of business, or pursuant to public requirements unrelated to litigation. .. . ” 48 F.R.D. at 501.
Even assuming that El Paso’s tax pool analysis otherwise qualifies for work product protection, we hold the doctrine unavailable here because the tax pool analysis is not prepared “in anticipation of litigation.” But in doing so we concede that determining whether a document is prepared in anticipation of litigation is a slippery task. See In Re Grand Jury Investigation, 599 F.2d 1224, 1229 (3d Cir. 1979); 8 C. Wright & A. Miller, Federal Prac. & Pro. § 2024 at 198 (1970). In United States v. Davis we phrased the test in the following terms: “[L]itigation need not be imminent ... as long as the primary motivating purpose behind the creation of the document was to aid in possible future litigation.” *543636 F.2d at 1040. Davis held that an attorney’s work in preparing a tax return did not fall under the work product doctrine because the work was not primarily motivated to assist in future litigation over the return. We reach a similar conclusion with respect to El Paso’s tax pool analysis.
El Paso establishes its non-current tax account to bring its financial books into conformity with generally accepted auditing principles. The desire to please the accountants, in turn, is compelled by the securities laws. The primary motivating force behind the tax pool analysis, therefore, is not to ready El Paso for litigation over its tax returns. Rather, the primary motivation is to anticipate, for financial reporting purposes, what the impact of litigation might be on the company’s tax liability. El Paso thus creates the tax pool analysis with an eye on its business needs, not on its legal ones.
We recognize that the tax pool analysis involves weighing legal arguments, predicting the stance of the IRS, and forecasting the ultimate likelihood of sustaining El Paso’s position in court. Nevertheless, this analysis (assuming it is legal rather than accounting work) is only a means to a business end. The legal analysis is not an end in itself: the memoranda underlying the tax pool analysis do not map out El Paso’s actual litigating strategy, should litigation occur. In fact, no single item in the tax pool analysis is specifically under scrutiny by the IRS when the memoranda are drafted. Cf. Kent Corp. v. NLRB, 530 F.2d 612, 123 (5th Cir. 1976), cert. denied, 429 U.S. 920, 97 S.Ct. 316, 50 L.Ed.2d 287 (1976) (investigation reports of the NLRB prepared after a charge has been filed are the NLRB attorney’s work product because the charge contemplates litigation). Business imperatives, not the press of litigation, call these documents into being.
Moreover, El Paso’s tax litigation is handled by outside counsel. Cf. United States v. Gates, 35 F.R.D. 524 (D.Colo.1964) (IRS documentary files on the taxpayer were not work product when referred to Justice Department attorneys who were prosecuting a tax suit against the taxpayer). Although an attorney from El Paso’s tax department serves as co-counsel, outside counsel takes the lead in directing the conduct of El Paso’s tax suits. There is no evidence in the record that the tax pool analysis or underlying memoranda are referred to outside counsel or used by El Paso’s attorneys to prepare for trial or negotiations.
In tax refund suits against the United States, documents generated by the IRS in connection with its defense have been denied work product protection even though much more closely tied to litigation then El Paso’s tax pool analysis. Abel Investment Co. v. United States, 53 F.R.D. 485 (D.Neb.1971); Peterson v. United States, 52 F.R.D. 317 (S.D.Ill.1971); United States v. San Antonio Portland Cement Co., 33 F.R.D. 513 (W.D.Tex.1963). In Abel Investment, the court denied the work product privilege to documents prepared by the IRS in response to a tax refund claim made by the taxpayer. The documents impartially evaluated the strengths and weaknesses of the IRS’s and the taxpayer’s positions. The court held that the documents were not prepared in anticipation of litigation.
In support of its conclusion, the Abel Investment court observed that the reports were prepared routinely before any lawsuit began and were not developed by the attorney who would try the case, should litigation ensure. Moreover, “[t]he documents in all probability do not fix the government’s theory of the case to be used at trial, because trial counsel should and undoubtedly would set the defense from all available facts and theories whether or not conceived or expressed by personnel at the various stages of the settlement process.... ” 53 F.R.D. at 49. El Paso’s tax pool analysis is even more distant from actual litigation than the IRS documents in Abel Investment. The factors mentioned in Abel Investment also point to a denial of work product protection here.
In sum, we believe that the tax pool analysis does not contemplate litigation in the sense required to bring it within the work product doctrine. The tax pool analy*544sis concocts theories about the results of possible litigation; such analyses are not designed to prepare a specific case for trial or negotiation. Their sole function, from all that appears in the record, is to back up a figure on a financial balance sheet. Written ultimately to comply with SEC regulations, the tax pool analysis carries much more the aura of daily business than it does of courtroom combat. We hold, therefore, that the tax pool analysis and backup mem-oranda are not protected work product materials.
VII.
El Paso contends that we should refuse to enforce the tax pool analysis summons, even though it seeks otherwise relevant and nonprivileged documents, on grounds of public policy. El Paso urges us to constrict the sweep of the IRS summons power because of the policy of the securities laws. To permit routine summoning of tax pool analyses from companies will have a chilling effect on the companies’ willingness to prepare such analyses searchingly and critically. Moreover, companies will conceal tax pool analyses from their auditors and thereby thwart the accountants’ attempts to measure the adequacy of the contingent tax accounts. These consequences obstruct the full and frank disclosure of financial information that the securities laws envision. The public policy of protecting investors, therefore, demands that we deny the IRS access to a company’s tax pool analysis.
We are not swayed by the argument that the public policy of the securities laws implicitly overrides the clear grant of summons power to the IRS. While the IRS does not enjoy untrammeled authority to direct the production of documents, see Upjohn Co., supra, (attorney-client privilege and work product doctrine are defenses to the enforcement of the summons in proper cases); United States v. LaSalle National Bank, 437 U.S. 298, 98 S.Ct. 2357, 57 L.Ed.2d 221 (1978) (good faith requirements); United States v. Powell, supra, Congress has endowed the IRS with broad authority to conduct tax investigations. The Supreme Court has “consistently construed congressional intent to require that if the summons authority claimed is necessary for the effective performance of con-gressionally imposed responsibilities to enforce the tax code, that authority should be upheld absent express statutory prohibition or substantial countervailing policies.” United States v. Euge, 100 S.Ct. at 878; United States v. Bisceglia, 420 U.S. 141, 95 S.Ct. 915, 43 L.Ed.2d 88 (1975). We are unwilling to make inroads in the plainly-announced congressional policy to allow the IRS broad access to relevant, nonprivileged documents on the basis of El Paso’s claim of a conflict with the policies underlying the securities laws.
The main premise of El Paso’s argument is that the accuracy of financial reports will suffer if companies must divulge their tax pool analyses to the IRS. We find this premise wholly speculative. El Paso paints a picture of corporations evading their responsibilities under the securities laws to prepare their financial books properly and to lay open their books and records to independent auditors. We do not join El Paso in assuming that corporations will dishonor their legal obligations by discontinuing the preparation of tax pool analyses. El Paso fails to recognize that “[i]f the corporation blocks the efforts of the accountants to ascertain the true state of the company’s contingent tax liabilities, the accountant would be obligated to decline to certify the financial statements.” United States v. Arthur Young, 677 F.2d at 224 (dissenting opinion). The powers of the SEC suffice to ensure that companies will comply with law.16
*545We note as well that Congress granted the IRS “broad latitude to adopt enforcement techniques helpful in the performance of ... tax collection and assessment....” United States v. Euge, 100 S.Ct. at 880 n.9. This latitude enables the IRS, without judicial intervention, to temper the perceived ill-effects of summoning tax pool analyses. The IRS has recognized resistance to the production of tax-pool analyses and responded by tightening its internal requirements for the issuance of such summonses. See Internal Revenue Manual § 4024.4 (Guidelines for Requesting Audit or Tax Accrual Workpapers) (May 14, 1981). These guidelines have no retroactive effect on the IRS’ power to summon the tax pool analysis in this case. They do, however, evidence administrative sensitivity to the concerns evoked by routinely summoning a company’s tax pool analysis. This demonstration of administrative flexibility reenforces our decision not to trim the IRS’ authority irrevocably.
By upholding the IRS’ authority to reach the tax pool analysis, we do not unleash an inquisition. The tax pool analysis may be useful to the IRS as a “roadmap” through a company’s tax return; it is not, however, an admission of guilt. The actual tax liability of El Paso is hammered out in negotiations and ultimately is subject to resolution in the courts. We therefore decline to hamper the efficiency that access to the tax pool analysis can bring to an IRS audit.
In the absence of a more profound clash between congressional policies, we cannot cut back on the summons power that Congress has given to the Service. We do not feel free to reweave the fabric of national legislation in accord with our notions of how various statutory schemes mesh. Such policy choices belong to the Congress. Accordingly, we find no public policy barrier to the enforcement of this summons.
CONCLUSION
Our tax collection system depends primarily on the voluntary compliance of taxpayers acting in good faith. Congress has chosen to enforce this system of self-assessment through an agency vested with broad investigatory powers. The enforcement of the revenue laws contemplates that the IRS may verify the basis of a return by a thorough audit of a taxpayer’s records. Certainly, IRS access to a document, prepared for business reasons, that guides and channels the Service’s energy and limited resources to those areas of the return deserving of special scrutiny fits within the congressional scheme of tax enforcement. Because we find the summons here to be authorized by Congress and not otherwise prohibited, we enforce the tax pool analysis summons.
We restate our holdings. First, we interpret the judgment to extend only to El Paso’s tax pool analysis and supporting memoranda in various subject files. Second, we do not find El Paso’s tax pool analysis shielded by the attorney-client privilege. Third, the documents in question are not protected work product materials. Fourth, we find a lack of public policy in the securities laws demanding that we deny enforcement. The stay of the district court’s judgment is lifted, and the judgment is
AFFIRMED.