BEFORE: GREENBERG AND ROTH, Circuit Judges and POLLAK, District Judge*.

LOUIS H. POLLAK, Senior District Judge.

Before us is a petition from the United States for a writ of mandamus or prohibition directed to the Honorable John W. Bissell, United States District Judge for the District of New Jersey. The government's petition arises out of a pretrial order entered in a criminal tax fraud case against Victor Wexler which is to be tried before Judge Bissell. The order adopted a jury instruction on "genuine indebtedness" that, in the government's view, undermines a well-settled prohibition against deducting interest payments resulting from "sham transactions"--i.e. transactions entered into with no purpose other than to generate tax benefits. The government argues that the instruction adopted by the district court is clearly erroneous under settled law, and that the government will be unable to proceed with the present prosecution and will be severely prejudiced in other tax fraud prosecutions if the order remains in force. Wexler, responding to the petition1, contends that the proposed instruction is a proper statement of the law and that, in any event, the extraordinary appellate intrusion on trial court proceedings sought by the government is unwarranted. We conclude that the petition should be granted.

Between 1980 and 1985, the defendant in the underlying tax prosecution, Victor Wexler, served first as chief financial officer and subsequently as managing partner of McMahan, Brafman, Morgan & Co. ("MBM"), a limited partnership engaged in securities trading. Wexler was initially indicted on March 19, 1992. Subsequently a superseding indictment was filed. The superseding indictment consists of eight counts, and charges Wexler with, under count 1, conspiring (i) to defraud the United States by obstructing the lawful government functions of the I.R.S. in violation of 18 U.S.C. Sec. 371, and (ii) to aid and assist in the preparation of false tax returns in violation of 26 U.S.C. Sec. 7206(2); under count 2, aiding and assisting in the preparation of a U.S. Partnership Income Return, Form 1065, for MBM, for calendar year 1984, which falsely represented that MBM had incurred a loss of $75,491,898, in violation of 26 U.S.C. Sec. 7206(2); under count 8, making and subscribing a joint individual income tax return, Form 1040, falsely representing that Wexler was entitled to a deduction of $103,928 flowing from his MBM partnership interest, in violation of 26 U.S.C. Sec. 7206(1); under counts 3-7, aiding and assisting in the preparation by others of joint individual income tax returns, Form 1040, falsely representing that the taxpayers were entitled to deductions flowing from their MBM partnership interests, in violation of 26 U.S.C. Sec. 7206(2). Superseding Indictment, Appendix ("App.") at 5-21. The superseding indictment alleges that Wexler created over $160 million in fraudulent tax deductions for the MBM partnership from 1982 through 1986. According to the superseding indictment, the allegedly fraudulent deductions were the product of financial arrangements known as "repo to maturity" transactions.

"Repo" transactions: In order to be able to parse the charges against Wexler one needs to have a general understanding of what "repo" transactions are and how they work. In its brief in this court, as in its submissions to the district court, the government has described and provided examples of such transactions and their mechanics. Government Br. at 5-15. Since Wexler's brief does not quarrel with the government's exposition, we rely upon that exposition in this section of this opinion.

The word "repo" is an abbreviation for "repurchase agreement", the name given to a type of transaction commonly employed by firms dealing in government securities. The transaction--which may be consummated in a matter of days but may also span weeks or even a few months--is a sale of government securities, such as treasury notes, by one securities dealer to another, followed by their repurchase at a later date. But what is in form a sale and repurchase turns out in fact to constitute a loan for which the securities, during the interval between sale and repurchase, stand as collateral. An example may serve to illustrate how such a transaction works:

Firm A sells Treasury notes with a face value of $1,000,000 to Firm B; the price paid by B to A--the "repo principal"--is a negotiated figure presumably geared to the market value of the notes at the date of sale; A concurrently contracts with B to buy the notes back at the same price at an agreed future date--e.g. thirty days or sixty days hence--which is earlier than the maturity date of the notes; on that future date B returns the securities to A, A repays the repo principal, and A also pays "repo interest", a sum negotiated along with the repo principal at the outset of the transaction, and presumably geared to the short-term interest rate then governing loans for the particular time-period--thirty days, or sixty days, or whatever--covered by the transaction.

The extent to which the "repo" turns out to be financially advantageous to A depends on what happens, during the course of the transaction, to (a) the market value of the securities, and (b) short-term interest rates. A hopes that, at the transaction's closing date, the reacquired securities will be worth more and short-term interest rates will be lower than when the transaction began. Under that fortunate combination of circumstances A would have the capability of entering into a second repo on substantially more favorable terms than the first.2

The particular repo just described is one in which, as already noted, the transaction terminates on a date, agreed upon by the parties, which is earlier than the maturity date of the securities. That transaction is called an "open repo". If the repo's date of termination is the maturity date--the date on which the Treasury pays to the securities-holder the face value of the securities plus accrued coupon interest--the transaction is called a "repo-to-maturity". This litigation involves the tax consequences of repo-to-maturity transactions.

At its commencement a repo-to-maturity looks like an open repo, in the sense that A(a) transfers its Treasury securities to B in exchange for an agreed sum of repo principal and (b) promises that at the maturity date it will pay B an agreed amount of interest. But at that point the resemblance to an open repo ends. For, in a repo-to-maturity, A will not, at the maturity date, recover the securities it has transferred to B. The securities will have matured, and in lieu thereof B will pay A what the Treasury owes B as securities-holder--namely, the face value of the securities plus the accrued coupon interest.

In contrast with the open repo, in which A's profit or loss turns on what happens to the market value of the securities and to short-term interest rates while the transaction is pending, the repo-to-maturity dictates ex ante the payout at maturity. Indeed, from a profitability perspective A has no occasion to enter into a repo-to-maturity unless the coupon interest A will receive at maturity exceeds the interest it will, at the inception of a repo-to-maturity, have to obligate itself to pay to B at maturity. If coupon interest is equal to or less than the market interest rate prevailing at the time A is considering a repo-to-maturity, A would normally eschew the repo-to-maturity and either (a) hold the securities until they mature, or (b) sell them outright.

In this case, the government alleges that MBM, the firm of which defendant Wexler was, variously, chief financial officer and managing partner, entered into numerous repo-to-maturity transactions, designed by Wexler, in which MBM agreed to pay out, and in fact did pay out, more in market interest than it received in coupon interest. Moreover, according to the government's allegations, it was part of Wexler's design that his firm purchase the securities and "repo" them simultaneously. The entire purpose of the transaction, so the government contends, was to create interest expenses that could, for tax purposes, be treated as a cost of business to be offset against profits from other transactions.3

Key to the effectuation of the scheme as described by the government was that the repo-to-maturity period was constructed to span two tax years. In the hypothetical example proffered by the government to the district court, and renewed in its brief to this court, the three-month repo-to-maturity transaction would run from November 1 in the first tax year, the date on which the securities were purchased and immediately repoed, to February 1 in the second tax year, the date on which the securities were to mature. This calendar arrangement would make it possible for the interest owed in November and December to be treated as a deductible expense in the first tax year, while the November-December-January coupon interest--partially offset by the January interest payment--would not need to be reported as income until the second tax year. Although in the aggregate two months of interest deductions in the first tax year would appear to be balanced by two months of interest income in the second tax year, the apparent symmetry is illusory from a revenue standpoint: it would not be until the filing of the return for the second tax year that the government would recoup the taxes not paid in the first tax year, with the result that the government would for a year lose the use of the sum ultimately recouped.4

In sum, the government contends that the repo-to-maturity device illustrated above was the centerpiece of a conspiracy in which, according to p 9 of the superseding indictment, Wexler and others undertook, between 1982 and 1986, "to generate more than $740 million ($740,000,000) in fraudulent interest expenses chargeable to MBM", which "fraudulent expenses were in turn used by Wexler and his co-conspirators to create more than $160 million ($160,000,000) in bogus tax deductions for the partners of MBM."

The November 30, 1993 Order: In October, 1993, the government moved for a declaration of various principles of tax law in order to preclude certain defenses that prosecutors anticipated from Wexler. The government's motion contained the following proposed jury instruction:

I instruct you that, if [the government securities transactions in this case] were sham transactions, then interest and other expenses that MBM incurred as a result of the transactions were not deductible. Whether or not the transactions were shams is a question that is for you to decide. I instruct you, however, that you must find the transactions were sham transactions if you determine two things beyond a reasonable doubt:

First, that MBM had no business purpose for entering into the transactions other than to obtain tax benefits; and

Second, that there was no reasonable possibility that MBM could earn a profit on the transactions apart from tax benefits.

Defendant also proposed a jury charge to the court, which included the following instruction:

Assuming the Government has proven that the governmental security trades were sham transactions (i.e. no business purpose and devoid of economic substance), you cannot convict the defendant if you find that the interest expenses at issue were the product of a "genuine indebtedness". In other words, regardless of any conclusions you reach as to the motivation of the defendant or the lack of economic substance, the existence of a genuine debt between MBM and its trading partners, will permit the deduction of interest. Under the tax laws, even a sham transaction may have elements that have economic substance, and these elements of the transaction must be respected for tax purposes. Accordingly, if you find that the financing agreement to purchase the government securities is genuine and enforceable, the interest may be deducted.

In its November 30, 1993, Order the district court adopted an instruction on sham transactions to be delivered to the jury. The court's instruction adopted both parties' central suggestions, including the "genuine indebtedness" paragraph, above, proposed by defendant. The court's instruction, in relevant part, is as follows:

I instruct you that, if these transactions in government securities were sham transactions, then interest and other expenses that MBM incurred as a result of the transactions were not deductible, with certain exceptions I will explain in a moment. Whether or not the transactions were economic shams is a question that is for you to decide. I instruct you, however, that you must find the transactions were sham transactions if you determine two things beyond a reasonable doubt:

First, that MBM had no business purpose for entering into the transactions other than to obtain tax benefits; and

Second, that there was no economic substance to the transaction, that is there was no reasonable possibility that MBM could earn a profit on the transactions apart from tax benefits.

* * * * * *

Assuming the government has proven that the governmental security trades were sham transactions (i.e. no business purpose and devoid of economic substance), you may convict the defendant on the count you are considering only if you find that the interest expenses at issue were not the product of a "genuine indebtedness." In other words, regardless of any conclusions you reach as to the motivation of the defendant or the lack of economic substance in any transaction as a whole, the existence of a genuine debt between MBM and a trading partner, could permit the deduction of interest under the Internal Revenue Code. Under the tax laws, even an economic sham transaction may give rise to genuine indebtedness and could generate lawful interest deductions.

Genuine indebtedness exists if the parties intended to create and enforce a binding obligation on the part of MBM to pay interest on a loan. In determining whether genuine indebtedness exists in a particular transaction, you must consider the substance of the transaction, as intended by the parties, not only the form that the transaction took. You may consider all relevant evidence bearing on the parties' intent. Among your duties as jurors, you must carefully analyze the evidence applicable to each count to see if one or more obligations generating a genuine indebtedness with actual, deductible interest was or was not present. The government must prove the lack of genuine indebtedness beyond a reasonable doubt.

District Court Order of November 30, 1993.

The government objected to the district court's order, and moved for reargument. The district court denied the motion and similarly denied a stay of trial while the government petitioned this court for mandamus or prohibition. We subsequently granted a stay pending decision on the petition.

The government seeks the extraordinary relief of mandamus or prohibition (which we will, hereafter, treat simply as a petition for mandamus) because it contends that (1) the district court's announced jury instruction on genuine indebtedness is clear legal error; (2) the government will be unable to proceed to trial in good faith within the rubric of that instruction and will, therefore, have to move to dismiss the indictment; (3) the error is unappealable; and (4) the district court's ruling, if allowed to stand, will affect other cases by virtue of setting an erroneous rule that makes tax cases involving sham transactions far more difficult to prosecute.

Resolution of the petition before us hinges upon two questions: First, is the instruction on genuine indebtedness contained in the district court's November 30, 1993, Order clearly erroneous? Second, if the answer to the first question is "yes", is the error one whose probable consequences meet the strict standard for granting a writ of mandamus?

The government's principal argument is that the district court's November 30, 1993, Order adopts an instruction that is patently contrary to the law. Specifically, the government argues that economic sham transactions cannot generate lawful deductions of interest expense, even if they involve the payment of interest on "genuine indebtedness". Gov't Br. at 27.

1) General rule on sham transactions: The general rule on sham transactions in this circuit is well-established: "If a transaction is devoid of economic substance ... it simply is not recognized for federal taxation purposes, for better or for worse. This denial of recognition means that a sham transaction, devoid of economic substance, cannot be the basis for a deductible loss." Lerman v. Comm'r of Internal Revenue, 939 F.2d 44, 45 (3rd Cir.1991), cert. denied, --- U.S. ----, 112 S. Ct. 1940, 118 L. Ed. 2d 546 (1992). Lerman reflects the fundamental and long-standing rule that taxation depends on the substance, not the form, of transactions. See e.g. Gregory v. Helvering, 293 U.S. 465, 469-70, 55 S. Ct. 266, 267-68, 79 L. Ed. 596 (1935).

Where a transaction has no substance other than to create deductions, the transaction is disregarded for tax purposes. Knetsch v. United States, 364 U.S. 361, 366, 81 S. Ct. 132, 135, 5 L. Ed. 2d 128 (1960); DeMartino v. Commissioner, 862 F.2d 400, 401 (2nd Cir.1988). Deductions for expenses resulting from such transactions are not permitted. James v. Commissioner, 899 F.2d 905, 908 n. 4 (10th Cir.1990) ("transactions lacking an appreciable effect, other than tax reduction, on a taxpayer's beneficial interest will not be recognized for tax purposes"); Yosha v. Commissioner, 861 F.2d 494, 499 (7th Cir.1988) (where transactions "were devices whose only possible or contemplated effect was to avoid taxes, and a fortiori they were not engaged in for profit", resulting deductions violate the tax code).

Wexler does not challenge the sham-transaction doctrine in general. Rather, Wexler argues that the doctrine does not apply to deduction of interest payments pursuant to Sec. 1635 of the tax code if the taxpayer's obligation to pay the interest is binding and enforceable. Wexler grounds his arguments in both the statute and the case law. First, he argues that Sec. 163 differs from other sections of the code by not requiring that the underlying transaction be motivated by profit. Second, he argues that a modern trend in the case law favors allowing deductibility of genuine debt, even when related to a sham transaction. We address these arguments in order.

2) Deductibility under Sec. 163: Section 163(a) of the Internal Revenue Code, 26 U.S.C. Sec. 163(a), states that "[t]here shall be allowed as a deduction all interest paid or accrued within the taxable year on indebtedness."6 Wexler argues that "[i]n contrast to other sections of the Internal Revenue Code allowing for deductions, Section 163 does not require any profit motive or the conduct of a trade or business in order to claim the deduction for interest expense." Respondent's Br. at 20. The apparent thrust of the argument is that, unlike deductions taken pursuant to other sections of the Code, deductions under Sec. 163 are permitted for interest obligations resulting from sham transactions, so long as those obligations are "genuine"--i.e. binding and enforceable.

It is true, as Wexler argues, that other code sections expressly require that the deductions they provide for arise from transactions having a business purpose or profit motive, whereas Sec. 163 does not.7 Nonetheless, the case law construing Sec. 163 clearly establishes that the sham transaction doctrine also bars interest deductions under that section of the code. In Knetsch v. United States, 364 U.S. 361, 81 S. Ct. 132, 5 L. Ed. 2d 128 (1960), the Supreme Court held that interest deductions under the relevant language of Sec. 163(a) of the 1954 Code (and the identical language of the predecessor provision, Sec. 23(b) of the 1939 Code8) are not permissible when the underlying transaction is determined to have been a sham. Knetsch purchased $4,000,000 worth of 30-year, deferred annuity, bonds carrying a 2.5% interest rate. He borrowed the $4,000,000 purchase price at a rate of 3.5%, pledging the bonds as security for the loan. In each of the two years at issue, Knetsch paid the interest on the loan balance, and received a loan in the amount of the increased value of the annuity bonds. The Court found that he paid a total of $294,570 in interest, and received $203,000 in loans. Knetsch's net out-of-pocket costs were thus $91,570. Meanwhile, the equity value of the annuity bonds--the amount by which the value of the bonds exceeded Knetsch's debt--was only $1000. The Court found that, apart from tax benefits, the only benefit from Knetsch's $91,570 out-of-pocket expenditure was "the relative pittance" of $1000. Id. at 366, 81 S.Ct. at 135. The Court concluded that Knetsch's expenditures "did not appreciably affect his beneficial interest except to reduce his tax", and that there was "nothing of substance to be realized by Knetsch from this transaction beyond a tax deduction." Id. The Court therefore held that the loan transactions were shams and disallowed Knetsch's interest deductions.

Wexler does not undertake to address Knetsch in his brief. Yet Knetsch plainly contradicts respondent's claim that a sham transaction can give rise to deductions under Sec. 163. So do many other cases. The seminal sham-transaction case of Goldstein v. Commissioner, 364 F.2d 734 (2d Cir.1966), specifically addressed deduction of interest expenses pursuant to Sec. 163(a). Id. at 736. The underlying transaction involved a sweepstakes winner who borrowed money at 4.0% to purchase bonds yielding 1.5%, and then tried to deduct the interest payments on the loan from her taxable winnings. Notwithstanding the lack of an explicit requirement of a profit motive or business purpose in the text of Sec. 163, the Second Circuit ruled the deduction invalid because the interest had arisen from a transaction entered into "without any realistic expectation of profit and 'solely' in order to secure a large interest deduction". Id. at 740.

In Lifschultz v. Commissioner, 393 F.2d 232 (2nd Cir.1968), the court disallowed interest deductions under Sec. 163 for five transactions because the transactions were entered into without expectation of economic profit and had no purpose beyond creating tax deductions. The Second Circuit ruled that "the controverted payments did not constitute interest within the meaning of Section 163(a) and were not deductible." Id. at 234. Similarly, in Lukens v. Commissioner, 945 F.2d 92 (5th Cir.1991), the Fifth Circuit upheld denial of an interest deduction under Sec. 163 where the trial court had reasonably found that " 'petitioners did not have a profit objective, independent of tax savings' " for engaging in the underlying transaction. Id. at 100 (quoting trial court opinion); See also United States v. Manko, 979 F.2d 900, 910-11 (2nd Cir.1992) (approving, in dicta, jury instruction that would bar interest deduction if underlying transaction did not involve profit potential or market risk).

The Ninth Circuit has likewise held that a sham transaction cannot generate genuine indebtedness for Sec. 163 purposes. Shirar v. Commissioner, 916 F.2d 1414, 1417 (9th Cir.1990). The Shirar court stated that "ordinarily" Sec. 163 permits interest paid on indebtedness to be deducted, but only if the interest is paid on genuine indebtedness. Id. Citing Knetsch, the court held that no genuine indebtedness can result from a sham transaction, which it defined as a transaction from which "there is nothing of substance to be realized beyond a tax deduction." Id.

The Tax Court has also disallowed Sec. 163 deductions where the underlying transaction constituted an economic sham. In Sheldon v. Commissioner, 94 T.C. 738, 1990 WL 69233 (1990), the court addressed the question of whether interest obligations from repo transactions very similar to those Wexler allegedly engaged in could be deducted under Sec. 163. The court acknowledged that under Sec. 163 deductibility is not expressly contingent upon a profit objective in the underlying transaction. Id. at 760. The court held, however, that interest is nonetheless not deductible when it has resulted from a transaction that lacked economic substance--i.e. had no " 'purpose, substance or utility apart from [its] anticipated tax consequences.' " Id. at 761 (quoting Goldstein, 364 F.2d at 740).9

While there is no case in this court that specifically addresses the deductibility under Sec. 163 of interest payments on sham transactions, we have addressed whether economic shams could create deductible interest payments under Sec. 23(b) of the 1939 Internal Revenue Code, the predecessor of Sec. 163. Weller v. Commissioner, 270 F.2d 294 (3rd Cir.1959). The issue in Weller--a case decided a year before Knetsch--was the deductibility of prepaid interest on annuity contract loans. We found that "although in form the payments appear to constitute interest within the meaning of section 23(b) of the Internal Revenue Code of 1939, the entire transaction lacks substance." Id. at 296. We cited Gregory v. Helvering, 293 U.S. 465, 55 S. Ct. 266, 79 L. Ed. 596 (1935), for the general principle that no tax benefit could be created by a transaction entered into for no economic benefit other than tax avoidance. Weller, 270 F.2d at 296-97. Noting that Gregory involved what the Supreme Court termed "an elaborate and devious form of conveyance masquerading as a corporate reorganization," a camouflage the Court declined to recognize for tax purposes because to do so "would be to exalt artifice above reality." 293 U.S. at 470, 55 S.Ct. at 268. We said in Weller that "the principle laid down in the Gregory case is not limited to corporate reorganizations, but rather applies to the federal taxing statutes generally." Id. at 297. We rejected the petitioners' argument that the deduction should be allowed because the interest obligation was real: "That there may be an obligation which is binding under local law is not determinative of whether there is a true indebtedness within the meaning of Section 23(b)." Id. at 298.

In Weller, we held that interest payments are not deductible where the underlying transaction has no purpose other than tax avoidance, despite the fact that Sec. 23(b)--now Sec. 163(a)--had no express business-purpose requirement. We also held in Weller that the existence of a real obligation to pay interest--which is how Wexler defines "genuine indebtedness"--does not change the result. Weller remains good law. Our recent treatment of the sham transaction doctrine in Lerman states that transactions with no economic significance apart from tax benefits lack economic substance. 939 F.2d at 48 & n. 6. To be sure, Lerman did not involve Sec. 163; but we stated in Lerman that "economic substance is a prerequisite to any Code provisions allowing deductions." 939 F.2d at 52 (second emphasis added). We thus find no merit in Wexler's argument that economic substance in the underlying transaction is not required for deductibility under Sec. 163.

3) Alleged changes in the law: Wexler argues that recent decisions of the Fourth and Second Circuits and of the Tax Court reflect a modern trend towards a genuine indebtedness exception for Sec. 163 deductions--an exception that the district court was correct to follow. In making this argument, Wexler relies heavily on Rice's Toyota World v. Commissioner, 752 F.2d 89 (4th Cir.1985), a Fourth Circuit opinion also relied on by the district court in fashioning the jury instruction to which the government has taken such strong exception.

Rice's Toyota, an automobile dealership, purchased a used computer from a leasing company by issuing a recourse note and two non-recourse notes to the seller. Rice's Toyota leased the computer back to the leasing company, and then declared tax deductions for depreciation of the computer and for interest paid on the recourse and non-recourse notes. The Fourth Circuit affirmed the Tax Court's finding that Rice's transaction was a sham because "Rice subjectively lacked a business purpose and the transaction objectively lacked economic substance." Id. at 95. After finding that in substance Rice had not purchased a computer, but rather had paid a fee in exchange for tax benefits, the court went on to examine whether any part of the overall scheme was economically substantive. Id. The court agreed that the capital depreciation and the interest on the non-recourse notes should not be deductible. Id. However, the Fourth Circuit disagreed with the Tax Court's disallowance of the interest paid on the recourse note. The appellate court held that, despite the sham nature of the underlying transaction, Rice's Toyota purchased "something of economic value", and characterized "the 'cash payment' on the recourse note as a 'fee' for purchase of expected tax benefits." Id. at 96. The court held that the installment payments on the recourse notes were "genuine obligations" and that the "interest due upon their payment was equally an obligation of economic substance." Id. at 96. Wexler argues that Rice's Toyota therefore allows deduction of interest so long as it was paid on an enforceable obligation.

We reject Rice's Toyota as authority for the instruction adopted by the district court for three reasons.

First, the Fourth Circuit did not hold that interest was deductible in the absence of economic substance. Rather, the court expressly based deductibility of the recourse note interest on a finding that the recourse debt was "an obligation of economic substance," Id. at 96--a finding which, if warranted, would of course mean that the interest paid on the recourse note was properly deductible.

Second, and more important, we are unable to reconcile the Fourth Circuit's finding that the recourse note was "an obligation of economic substance" with its definition of "economic substance". According to the Fourth Circuit, "economic substance" involves "a reasonable possibility of profit from the transaction ... apart from tax benefits." Id. at 94. The court held, however, that the recourse note transaction had economic substance because Rice's payment on the note was a "purchase of something of economic value ... a 'fee' for purchase of expected tax benefits." Id. at 96. By the Fourth Circuit's own definition of economic substance, we do not see how the recourse note payment could possess economic substance given that the only purpose or gain was the tax benefit.

Third and finally, even if we were to accept respondent's interpretation of Rice's Toyota, Wexler's case is distinguishable from Rice's. The Fourth Circuit unbundled the sale-leaseback transaction and found that one small part of it, the recourse note, was economically substantive. Rice's transaction was unusual because the interest payments on the recourse note were separable from the interest payments and depreciation that would have created the principal tax benefits of the transaction. It appears that the Fourth Circuit did not view the interest expense on the recourse note as being, in itself, the purpose of Rice's transaction. The chief intended benefits of the sale-leaseback arrangement were the much larger deductions for depreciation and interest on the non-recourse notes. The court disallowed those larger deductions.

Wexler's case differs in a critical respect. There is no debt obligation that can be separated from the underlying repo scheme or that was undertaken for some reason other than the tax benefits of deducting interest on that obligation itself. The obligation that Wexler argues to be an economically substantive "genuine indebtedness", the loan secured by the government securities, is the very obligation that will generate the interest payments constituting the tax benefits of the entire transaction. Thus, if we were to follow Rice's Toyota, the transaction would be examined for a binding debt obligation even if the jury had found that Wexler's transaction was devoid of business purpose or profit potential other than tax avoidance--i.e. that, by the Fourth Circuit's own definition, the transaction was without economic substance. If such an obligation were found--i.e. that the MBM partnership was obligated to pay interest on the loan secured by the government securities--then the interest would be deductible because, under Rice's Toyota, it represents a fee paid for economically valuable tax benefits, thereby giving the transaction economic substance. We think such an outcome would undermine the sham transaction doctrine, for it would allow the MBM partnership to reap the entire tax benefit of its sham transaction simply because MBM was contractually obligated to make payments on the loan that was the centerpiece of the whole scheme. Such an outcome would be incompatible with the holdings of Weller, Knetsch, Gregory, Goldstein, and Lerman. This may explain why, five years after Rice's Toyota, the sixteen Tax Court judges deciding Sheldon v. Commissioner, 94 T.C. 738, 1990 WL 69233 (1990), all agreed that interest payments related to a sham transaction cannot be deducted under Sec. 163.10

Respondent argues that several recent decisions have followed Rice's Toyota on the genuine indebtedness issue, reflecting a current trend in the law. For example, Wexler cites Jacobson v. Commissioner, 915 F.2d 832 (2nd Cir.1990), as evidence that the Second Circuit has abandoned Goldstein and now allows deductions of interest payments notwithstanding an underlying sham transaction. But the Jacobson opinion does not support Wexler's argument.

In Jacobson, the court stated that "[e]ven if the motive for a transaction is to avoid taxes, interest incurred therein may still be deductible if it relates to economically substantive indebtedness." 915 F.2d at 840 (citing Rice's Toyota). The government does not quarrel with that formulation here: indeed, the two-part definition of a sham transaction proposed by the government and adopted by the district court would require the jury, in order to "find the transactions were sham transactions," to "determine two things beyond a reasonable doubt:"

First, that MBM had no business purpose for entering into the transactions other than to obtain tax benefits; and

Second, that there was no economic substance to the transaction, that is there was no reasonable possibility that MBM could earn a profit on the transactions apart from tax benefits.

Thus, what Jacobson holds is that interest will be deductible where the government proves lack of business purpose, but cannot prove lack of economic substance--i.e. if the underlying transaction is not a "sham" as defined by the district court's order in Wexler's case.11

Respondent also relies heavily on the Tax Court's recent decision in Lieber v. Commissioner, 66 T.C.M. (CCH) 722, 1993 WL 347353 (1993). In Lieber, the court addressed, inter alia, the deductibility of interest obligations incurred in connection with a computer purchase and lease transaction. The Tax Court first cited its own opinion in Rose v. Commissioner, 88 T.C. 386, 423, 1987 WL 49274 (1987), aff'd. 868 F.2d 851 (6th Cir.1989), in which it had stated that under Sec. 163 a taxpayer may deduct interest payments even absent a tax-independent motive for the underlying transaction. The Tax Court in Lieber went on, however, to acknowledge that its Rose dictum was in conflict with its more recent ruling in Sheldon that transactions "give rise to deductible interest only if there is some tax-independent purpose for the transactions." Sheldon, 94 T.C. 759.

Contrary to Wexler's argument, the Tax Court in Lieber did not reject Sheldon in favor of Rose, but expressly decided not to resolve the tension between Rose and Sheldon. Rather, it decided that, because an appeal in Lieber would be to the Second Circuit, which in the Tax Court's view had adopted the Rose dictum in Jacobson12, the Tax Court would follow Rose in Lieber. Thus, the Tax Court did not decide which position was correct, but rather decided the question with reference to what the Tax Court perceived the law to be in the relevant circuit court of appeals. Thus, the only relevance of Lieber to the instant case is to indicate that, were the instant case one that had arisen in the Tax Court, that court would have decided the issue at bar with reference to the prevailing law of the this circuit. And in this circuit, under Weller and Lerman, the interest arising from Wexler's transaction is not deductible if the underlying repo transaction is a sham.

We are convinced that there is no basis for finding that interest payments arising from the repo-to-maturity transactions allegedly arranged by Wexler should be deductible as "genuine indebtedness". Rice's Toyota, Jacobson and Lieber indicate that, in some circumstances, a sham transaction may have separable, economically substantive, elements that give rise to deductible interest obligations.13 Yet in each of those cases a key requirement is that the interest obligation be economically substantive, defined in every decision except Rose to mean that the transaction have a potential non-tax benefit. The jury instruction that the district court adopted in its November 30, 1993, Order goes even further: if, pursuant to the sham transaction instruction, the jury were to find beyond a reasonable doubt both "no business purpose ... other than to obtain tax benefits" and "no economic substance [because] there was no reasonable probability [of] a profit ... apart from tax benefits," deductibility would still follow unless the government proved beyond a reasonable doubt the "lack of genuine indebtedness."14 We find that this instruction would render the sham transaction doctrine inert, and that the instruction is at odds with the overwhelming weight of the relevant case law.

4) Clear error in the November 30, 1993, Order: On the basis of the foregoing discussion, we find that the "genuine indebtedness" instruction adopted by the district court constitutes clear error under established law. This circuit held in Weller that interest payments arising from a sham transaction are not deductible. 270 F.2d at 296-98. Our rule of disregarding sham transactions for federal taxation purposes continues in full force today, with no exception for Sec. 163 deductions. Lerman, 939 F.2d at 48 & n. 6, 52 ("economic substance is a prerequisite to any Code provisions allowing deductions") (second emphasis added); see also Peerless Industries v. United States, 1994 WL 13837 * 5, 1994 U.S.Dist. LEXIS 411 * 17, 94-1 U.S. Tax Cas. (CCH) p 50,043 (E.D.Pa.1994), (Sec. 163 "deduction is proper if there is some substance to the loan arrangement beyond ... the deduction."). The jury instruction adopted by the district court in its November 30, 1993, Order conflicts with the established law of this circuit, and with the dominant line of precedent following the Supreme Court's ruling in Knetsch and the Second Circuit's ruling in Goldstein.Whether a writ of mandamus should issue

The standard for issuing a writ of mandamus is stringent. As a threshold matter, the petitioner must prove that the district court committed a "clear abuse of discretion", Mallard v. U.S. District Court, 490 U.S. 296, 309, 109 S. Ct. 1814, 1822, 104 L. Ed. 2d 318 (1989), or a "clear error of law", In re Bankers Trust Co., 775 F.2d 545, 547 (3rd Cir.1985). We have found that the government has met that burden in this case.

In addition to demonstrating clear error, however, the petitioner must generally show that, other than mandamus, it has no means of adequate relief, Bankers Trust, 775 F.2d at 547, and that the error will cause irreparable injury, Cippolone v. Liggett Group, Inc., 785 F.2d 1108, 1118 (3rd Cir.1986).

We find in this case that the government has no alternative avenue of relief. The government sought rehearing on the intended jury instructions, but rehearing was denied. For double-jeopardy reasons, no appeal will be possible once trial begins. The government will not be able to interrupt the trial by filing an appeal or a renewed petition for mandamus when the district judge commences to give the erroneous instruction. And if--as the government anticipates, and Wexler does not contest--jury deliberations guided by the erroneous instruction end in an acquittal, the injury to the government will be irremediable.15

We are mindful of respondent's argument that mandamus is an extraordinary remedy, reserved for rare circumstances. We believe, however, that this case presents an appropriate occasion for exercise of the writ. Not only do we find clear error, likelihood of irreparable harm, and lack of alternatives, but we also note the absence of certain factors that would weigh against mandamus. A principal concern with mandamus is that it not be used as a substitute for appeal. In Re School Asbestos Litigation, 977 F.2d 764, 772 (3rd Cir.1992). Indeed, "[m]andamus is disfavored because its broad use would threaten the policy against piecemeal appeals." Id. In the case at bar, issuance of the writ will neither substitute for appeal nor bring this case piecemeal to this court, for the simple reason that appeal from the erroneous instruction is not an option for the government.16

Wexler would cabin our discretion to order mandamus to situations in which a district court either exceeded its lawful jurisdiction or declined to exercise a non-discretionary power. While those two situations constitute traditional reasons for the writ, they are not exclusive. Indeed, we have observed that "courts have not confined themselves to any narrow or technical definition of the term 'jurisdiction' " in using the writ to regulate proceedings in the district court. United States v. Santtini, 963 F.2d 585, 594 (3rd Cir.1992). Accordingly, mandamus may issue to correct clear abuses of discretion, to further "supervisory and instructional goals", and to resolve "unsettled and important" issues. In re School Asbestos Litigation, 977 F.2d at 773. While appellate courts must be parsimonious with the writ, it is also true that "[s]ome flexibility is required if the extraordinary writ is to remain available for extraordinary situations." Id.

We find that the adoption of a clearly erroneous jury instruction that entails a high probability of failure of a prosecution--a failure the government could not then seek to remedy by appeal or otherwise--constitutes the kind of extraordinary situation in which we are empowered to issue the writ of mandamus. Accordingly, the petition for a writ of mandamus is granted, the district court is directed to vacate its order of November 30, 1993, and the proceedings in the underlying case shall be consistent with this opinion.

I join in the opinion of the court because I believe that it correctly states the substantive tax law and because I further believe that it is appropriate for us to exercise our mandamus jurisdiction to direct the correction of the error in the proposed charge. Nevertheless I do so with some reluctance because I have difficulty reconciling Wexler's prosecution with the principles underlying Cheek v. United States, 498 U.S. 192, 111 S. Ct. 604, 112 L. Ed. 2d 617 (1991). In its petition for mandamus the government tells us that "the district court's order adopts an erroneous jury instruction predicated on a fundamentally incorrect reading of the tax law [which] if not corrected by this Court ... will force the government to dismiss the charges...." Petition at 1. Thus, the government explains, the "incorrect instruction would virtually preclude the government from obtaining a conviction [as] it would leave [it] with no means of proving a critical element of its case--that the interest deductions at issue were invalid." Petition at 2. The government adheres to this position in its reply brief, asserting that the district court's error "will force the government to dismiss the criminal charges." Reply brief at 4. In the circumstances, it is evident that the government's trial problem is that the district court intends to charge the jury that the deductions involved in this case were not unlawful, and the government quite understandably feels that in the face of that charge it cannot be successful at trial.

What bothers me in the government's position is that it is challenging a conclusion reached by the district court in a carefully crafted formal, though unpublished opinion, rendered after the court considered the appropriate authorities, including some of the very cases we cite, e.g., Lerman v. Comm'r, 939 F.2d 44 (3d Cir.1991), cert. denied, --- U.S. ----, 112 S. Ct. 1940, 118 L. Ed. 2d 546 (1992). Furthermore, the district court adhered to its conclusion on reconsideration. Therefore, Wexler is being prosecuted in a case in which the government intends to demonstrate that he acted in what the district court believed was a lawful manner. This circumstance almost has caused me to vote to deny the government's petition.

In Cheek the Supreme Court noted that "[t]he proliferation of statutes and regulations has sometimes made it difficult for the average citizen to know and comprehend the extent of the duties and obligations imposed by the tax laws." 498 U.S. at 199-200, 111 S.Ct. at 609. Thus, "[i]n the end [in a criminal case] the issue is whether, based on all the evidence, the Government has proved that the defendant was aware of the duty at issue, which cannot be true if the jury credits a good-faith misunderstanding and belief submission, whether or not the claimed belief or misunderstanding is objectively reasonable." Id. at 202, 111 S.Ct. at 611. The Supreme Court's formulation throws into question the fairness of Wexler's prosecution for even the district court did not believe that Wexler violated "the duty at issue." Thus, I have considered whether, notwithstanding the government's demonstration of a prima facie case for the issuance of a writ of mandamus, we should not exercise our discretion by denying its application. See In re School Asbestos Litig., 977 F.2d 764, 772 (3d Cir.1992).

However, I ultimately do join in the opinion for I do not question the government's good faith in instituting the criminal case as I have no doubt that it was surprised by the legal conclusions reached by the district court, and, on the limited record before us, I do not know all the circumstances which led the government to seek Wexler's indictment. Furthermore, I cannot possibly fault the government for seeking a writ of mandamus from this court as I accept its representation that, unless corrected, the district court's unreported opinion will be circulated and will stand as an impediment to the enforcement of the tax laws. Of course, now that the government has obtained its goal of overcoming the district court's opinion, the determination whether this criminal case should be continued must be made by the prosecuting authorities and not by this court.

I close, however, by pointing out that our opinion is narrow. We only give directions with respect to the charge to the jury. Accordingly, we do not address the possibility that Wexler may in some way be able to bring to the jury's attention the district court's opinion to support a contention that he acted in good faith with respect to the transactions and the deductions involved. That issue, however, is not before us and consequently I express no opinion with respect to it.

Before: SLOVITER, Chief Judge, and BECKER, STAPLETON, MANSMANN, GREENBERG, HUTCHINSON, SCIRICA, COWEN, NYGAARD, ALITO, ROTH, LEWIS, and McKEE, Circuit Judges, and POLLAK, Senior District Judge*.

The petition for rehearing filed by respondent Victor Wexler in the above captioned matter having been submitted to the judges who participated in the decision of this court and to all the other available circuit judges of the court in regular active service, and no judge who concurred in the decision having asked for rehearing, and a majority of the circuit judges of the circuit in regular active service not having voted for rehearing by the court in banc, the petition for rehearing is denied.