with whom Justice Marshall joins, concurring in the judgment in No. 81-485 and dissenting in No. 81-930.
These two cases should be decided in the same way. The taxpayer in each case is a corporation. In 1972 each taxpayer made a deductible expenditure, and in 1973 its shareholders received an economic benefit. Neither corporate taxpayer ever recovered any part of its 1972 expenditure. In my opinion, the benefits received by the shareholders in 1973 are matters that should affect their returns; those benefits should not give rise to income on the 1973 return of the taxpayer in either case.
Both cases require us to apply the tax benefit rule. This rule has always had a limited, but important office: it deter*404mines whether certain events that enrich the taxpayer— recoveries of past expenditures — should be characterized as income.1 It does not create income out of events that do not enhance the taxpayer’s wealth.
Today the Court declares that the purpose of the tax benefit rule is “to approximate the results produced by a tax system based on transactional rather than annual accounting.” Ante, at 381. Whereas the rule has previously been used to determine the character of a current wealth-enhancing event, when viewed in the light of past deductions, the Court now suggests that the rule requires a study of the propriety of earlier deductions, when viewed in the light of later events. The Court states that the rule operates to “cancel out” an earlier deduction if the premise on which it is based is “fundamentally inconsistent” with an event in a later year. Ante, at 383.2
The Court’s reformulation of the tax benefit rule constitutes an extremely significant enlargement of the tax collector’s powers. In order to identify the groundbreaking character of the decision, I shall review the history of the tax benefit rule. I shall then discuss the Bliss Dairy case in some detail, to demonstrate that it fits comfortably within *405the class of cases to which the tax benefit rule has not been applied in the past. Finally, I shall explain why the Court’s adventure in lawmaking is not only misguided but does not even explain its inconsistent disposition of these two similar cases.
I
What is today called the “tax benefit rule” evolved in two stages, reflecting the rule’s two components. The “inclu-sionary” component requires that the recovery within a taxable year of an item previously deducted be included in gross income. The “exclusionary component,” which gives the rule its name, allows the inclusionary component to operate only to the extent that the prior deduction benefited the taxpayer.
The inclusionary component of the rule originated in the Bureau of Internal Revenue in the context of recoveries of debts that had previously been deducted as uncollectible. The Bureau sensed that it was inequitable to permit a taxpayer to characterize the recovery of such a debt as “return of capital” when in a prior year he had been allowed to reduce his taxable income to compensate for the loss of that capital. As one commentator described it, “the allowance of a deduction results in a portion of gross income not being taxed; when the deducted item is recouped, the recovery stands in the place of the gross income which had not been taxed before and is therefore taxable.”3 This principle was quickly endorsed by the Board of Tax Appeals and the courts. See Excelsior Printing Co. v. Commissioner, 16 B. T. A. 886 (1929); Putnam National Bank v. Commissioner, 50 F. 2d 158 (CA5 1931).
*406The exclusionary component was not so readily accepted. The Bureau first incorporated it during the Great Depression as the natural equitable counterweight to the inclusionary component. G. C. M. 18525, 1937-1 Cum. Bull. 80. It soon retreated, however, insisting that a recovery could be treated as income even if the prior deduction had not benefited the taxpayer. G. C. M. 22163, 1940-2 Cum. Bull. 76. The Board of Tax Appeals protested, e. g., Com Exchange National Bank & Trust Co. v. Commissioner, 46 B. T. A. 1107 (1942), but the Circuit Courts of Appeals sided with the Bureau. Helvering v. State-Planters Bank & Trust Co., 130 F. 2d 44 (CA4 1942); Commissioner v. United States & International Securities Corp., 130 F. 2d 894 (CA3 1942). At that point, Congress intervened for the first and only time. It enacted the forerunner of § 111 of the present Code, ch. 619, Title I, § 116(a), Act of Oct. 21, 1942, 56 Stat. 812, using language that by implication acknowledges the propriety of the inclusionary component by explicitly mandating the exclusionary component.4
The most striking feature of the rule’s history is that from its early formative years, through codification, until the 1960’s, Congress,5 the Internal Revenue Service,6 courts,7 *407and commentators,8 understood it in essentially the same way. They all saw it as a theory that appropriately characterized certain recoveries of capital as income. Although the rule undeniably helped to accommodate the annual accounting system to multiyear transactions, I have found no sugges*408tion that it was regarded as a generalized method of approximating a transactional accounting system through the fabrication of income at the drop of a fundamentally inconsistent event.9 An inconsistent event was always a necessary condition, but with the possible exception of the discussion of the Board of Tax Appeals in Barnett v. Commissioner, 39 B. T. A. 864, 867 (1939), inconsistency was never by itself a sufficient reason for applying the rule.10 Significantly, the first case from this Court dealing with the tax benefit rule emphasized the role of a recovery.11 And when litigants in *409this Court suggested that a transactional accounting system would be more equitable, we expressly declined to impose one, stressing the importance of finality and practicability in a tax system.12
*410In the 1960’s, the Commissioner, with the support of some commentators and the Tax Court, began to urge that the tax benefit rule be given a more ambitious office.13 In Nash v. United States, 398 U. S. 1 (1970), the Commissioner argued that the rule should not be limited to cases in which the taxpayer had made an economic recovery, but rather should operate to cancel out an earlier deduction whenever later events demonstrate that the taxpayer is no longer entitled to it. The arguments advanced, and rejected, in that case were remarkably similar to those found in the Court’s opinion today.14
*411The Nash case arose out of the sale of a partnership business to a corporation. The partnership had taken deductions for ledger entries in a “bad debt reserve” — an account that reflected the firm’s estimate of its future losses from accounts receivable that would eventually become uncollectible. When the partnership business was sold to a corporation, the Commissioner sought to apply the tax benefit rule, arguing that even though the partnership had made no recovery of the amount in the bad debt reserve, the deductibility of the presale additions to the taxpayer’s reserve had been justified on the basis of an assumption that was no longer valid after the business was sold.15
This Court flatly rejected the Commissioner’s position. Rather than scrutinizing the premises of the prior deduction in the light of subsequent events, the Court used the subsequent events themselves as its starting point. Since the transfer of the bad debt reserve did not enrich the taxpayer, there was no current realization event justifying the application of the tax benefit rule. “[Although the ‘need’ for the reserve ended with the transfer, the end of that need did *412not mark a ‘recovery’ within the meaning of the tax benefit cases.” Id., at 5.16
Today, the Court again has before it a case in which the Commissioner, with the endorsement of some commentators and a closely divided Tax Court, is pushing for a more ambitious tax benefit rule.17 This time, the Court accepts the invitation. Since there has been no legislation since Nash suggesting that our approach over the past half-century18 has been wrong-headed, cf. n. 32, infra, the new doctrine that emerges from today’s decision is of the Court’s own making.
In the Bliss Dairy case, the Court today reaches a result contrary to that dictated by a recovery theory. One would not expect such a break with the past unless it were apparent that prior law would produce a palpable inequity — a clear windfall for the taxpayer. Yet that is not the case in Bliss Dairy. Indeed, the tax economics of the case are indistinguishable from those of the Mash case.
*413Three statutory provisions, as interpreted by the Commissioner, interact in Bliss Dairy. First, pursuant to § 162(a),19 the Commissioner allowed the corporation to deduct the entire cost of all grain purchased in 1972. That deduction left it with a basis of zero in that grain. Second, under the terms of §336,20 the corporation was not required to recognize any gain or loss when it went through a § 333 liquidation in 1973. And third, pursuant to the regulations implementing § 334,21 the shareholders were allowed to assign'some portion of their basis in the corporation’s stock to the grain they received in the liquidation. Admittedly, this combination of provisions could in some cases cause a “step-up” in the grain’s basis that is not reflected in the income of either the corporation or the shareholders. That possibility figured strongly in the decision of the Court of Appeals for the Sixth Circuit to endorse an inconsistent-event theory in a precursor of this case. See Tennessee-Carolina Transportation, Inc. v. Commissioner, 582 F. 2d 378, 382, and n. 14 (1978). And it is stressed by the Solicitor General in his argument in this case. Brief for United States in No. 81-930 and for Respondent in No. 81-485, pp. 39-42. Yet close analysis reveals that the potential untaxed step-up is not the sort of extraordinary and inequitable windfall that calls for extraordinary measures in this case.
*414As a factual matter, the record does not include the tax returns of Bliss Dairy’s shareholders. We have no indication of how much, if any, step-up in basis actually occurred. And as a legal matter, a § 333 liquidation expressly contemplates steps-up in basis that are not reflected in income. Thus, even if the corporation had behaved as the Court believes it should have and had fed all the grain to the cows before liquidating, whatever shareholder stock basis was assigned to the grain in this case would have been used to step up the basis of some other asset that passed to the shareholders in the liquidation.
I suppose it might be argued that this sort of untaxed step-up is acceptable if it happens accidentally, but not if a taxpayer manipulates business transactions solely to take advantage of it. Yet here again we have too little information to conclude that there has been any such manipulation in the case of Bliss Dairy. To begin with, the Government has never questioned the propriety of the 1972 deduction, viewed in the light of 1972 events.22 Moreover, the record before us on appeal does not tell us how much feed the Dairy’s cattle consumed in 1972, whether 1972 consumption exceeded 1972 purchases, or how the volume purchased in 1972 compared with purchases in prior years. Indeed, it is quite possible that in 1971 the Dairy had made abnormally large purchases as a hedge against a possible rise in the market price, and that its 1972 consumption of grain actually exceeded its $150,000 in purchases during that year.
It is no doubt for these reasons that the Court never relies on the untaxed step-up argument in its opinion today.23 Un*415fortunately, the only argument the Court offers in its place is an ipse dixit: it seems wrong for a taxpayer not to realize income if it fails to use up an asset, when it was allowed to deduct the value of that asset in a prior year. We rejected that precise proposition in Nash. In both Nash and Bliss Dairy, the transfer of the business in a subsequent year revealed that a business asset matching a prior deduction (i. e., grain matching the expense deduction, or the account receivable matching the bad debt deduction) would not be used up (i. e., consumed or become uncollectible) until it had passed to a different taxpayer.24 The only explanation for today’s decision to detach the tax benefit rule from the recovery mooring appears to be the challenge to be found in an open sea of troublesome and inconclusive hypothetical cases.25
*416I — I I — I I — I
Because tax considerations play such an important role in decisions relating to the investment of capital, the transfer of operating businesses, and the management of going concerns, there is a special interest in the orderly, certain, and consistent interpretation of the Internal Revenue Code. Today’s decision seriously compromises that interest. It will engender uncertainty, it will enlarge the tax gatherer’s discretionary power to reexamine past transactions, and it will produce controversy and litigation.
Any inconsistent-event theory of the tax benefit rule would make the tax system more complicated than it has been under the recovery theory.26 Inconsistent-event analysis *417forces a deviation from the traditional pattern of calculating income during a given year: identify the transactions in which the taxpayer was made wealthier, determine from the history of those transactions which apparent sources of enrichment should be characterized as income, and then determine how much of that income must be recognized. Of course, in several specific contexts, Congress has already mandated deviations from that traditional pattern,27 and the additional complications are often deemed an appropriate price for enhanced tax equity. But to my knowledge Congress has never even considered so sweeping a deviation as a general inconsistent-event theory.
Nonetheless, a general inconsistent-event theory would surely give more guidance than the vague hybrid established by the Court today. The dimensions of the Court’s newly fashioned “fundamentally inconsistent event” version of the *418tax benefit rule are by no means clear. It obviously differs from both the Government’s “inconsistent event” theory and the familiar “recovery” theory, either of which would require these two cases to be decided in the same way. I do not understand, however, precisely why the Court’s theory distinguishes between these cases, or how it is to be applied in computing the 1973 taxes of Bliss Dairy, Inc.
The Government describes its test as whether “subsequent events eliminate the factual premise on which the deduction was originally claimed.” Brief for United States in No. 81-930 and for Respondent in No. 81-485, p. 18. The Court describes its test as whether “the later event is indeed fundamentally inconsistent with the premise on which the deduction was initially based.” Ante, at 383. One might infer that the difference between these tests is a difference between “inconsistent events” and “fundamentally inconsistent events.” The Court attempts to place the line more precisely “between merely unexpected events and inconsistent events.” Ante, at 383, n. 15. I am afraid the attempt fails because, however it is described, the line does not cleanly and predictably separate the Court’s position from the Government’s.
The Court presents its test as whether “the occurrence of the [later year’s] event in the earlier year would have resulted in the disallowance of the deduction.” Ante, at 389. But in Hillsboro, the Court rejects the Government’s claim. The Court holds that if this Court had decided Lehnhausen v. Lakeshore Auto Parts Co., 410 U. S. 356 (1973), in 1972, the Bank would still have been entitled to deduct a dividend that was not used for the payment of taxes. It attempts to read the legislative history of § 164(e) as establishing that Congress did not care about the use to which the dividend payment was put, but only about the Bank’s reason for the dividend. I would simply note that I find Justice Blackmun’s interpretation of § 164(e) far more plausible.
The Court’s analysis of Bliss Dairy is equally unsatisfying. Without any mention of a same-year principle, the Court re*419solves the case in a single sentence: a §336 liquidation is assimilated to a dividend distribution, which is deemed “the analog of personal consumption,” and therefore “it would seem that it should take into income the amount of the earlier deduction.” Ante, at 396, and n. 31.28 It is not obvious to me why the change in ownership of a going concern is more “the analog of personal consumption” than the gift of an asset.
The new rule will create even more confusion than that which will accompany efforts to reconcile the Court’s disposition of these two cases. Given that Nash is still considered good law by the Court, it is not clear which prior expenses of Bliss Dairy, Inc., will give rise to income in 1973. Presumably, all expenses for the purchase of tangible supplies will be treated like the cattle feed. Thus, all corporate paper towels, paper clips, and pencils that remain on hand will become income as a result of the liquidation. It is not clear, however, how the Court would react to other expenses that provide an enduring benefit. I find no limiting principle in the Court’s opinion that distinguishes cattle feed and pencils from prepaid rent, prepaid insurance, accruals of employee vacation time, advertising, management training, or any other expense that will have made the going concern more valuable when it is owned directly by its shareholders.
The Court’s opinion also leaves unclear the amount of income that is realized in the year in which the fundamentally inconsistent event occurs. In most of its opinion, the Court indicates that the taxpayer is deemed to realize “the amount of his earlier deduction,” ante, at 383, but from, time to time the Court equivocates,29 and at least once suggests that when *420an expensed asset is sold, only the “amount of the proceeds on sale,” ante, at 395, is income.30 Even in Bliss Dairy, which involves a revolving inventory of a fungible commodity, I am not sure how the Court requires the “cost” of the grain, ante, at 403, to be computed. If the corporation’s 1972 consumption matched its 1972 purchases, one might think that the relevant cost was that in the prior years when the surplus was built up. I cannot tell whether or why the fundamentally inconsistent-event theory prefers LIFO accounting over FIFO.
IV
Neither history nor sound tax policy supports the Court’s abandonment of its interpretation of the tax benefit rule as a tool for characterizing certain recoveries as income. If Congress were dissatisfied with the tax treatment that I believe Bliss Dairy should be accorded under current law, it could respond by changing any of the three provisions that bear on this case. See supra, at 413. It could modify the manner in which deductions are authorized under § 162.31 It could
*421legislate another statutory exception to the annual accounting system, much as it did when it made the depreciation recapture provisions, §§ 1245, 1250, apply to § 336 liquidations.32 Or it could modify the manner in which basis is allocated under § 334.33 But in the absence of legislative action, I cannot join in the Court’s attempt to achieve similar results by distorting the tax benefit rule.34
*422Accordingly, I concur in the Court’s judgment in No. 81-485 and respectfully dissent in No. 81-930.
Cf. H. Simons, Personal Income Taxation 50 (1938) (defining income as net accumulation of property rights over the course of a year, plus consumption during that year).
Notwithstanding this focus on the legitimacy of the 1972 deductions, scrutinized with the aid of hindsight after the completion of a multiyear transaction, the Court is careful not to require that “transactional inequities” be dealt with in the most “precise way” imaginable. Ante, at 378, n. 10. It rejects Justice Blackmun’s suggestion that the 1972 tax returns be reopened, quite properly noting our past observations that “ ‘[i]t would be disruptive of an orderly collection of the revenue to rule that the accounting must be done over again to reflect events occurring after the year for which the accounting is made, and would violate the spirit of the annual accounting system.’” Ante, at 380, n. 10, quoting Healy v. Commissioner, 345 U. S. 278, 285 (1953). See also ante, at 380-381, n. 12.
Plumb, The Tax Benefit Rule Today, 57 Harv. L. Rev. 129, 131, n. 10 (1943). Accord, Estate of Collins v. Commissioner, 46 B. T. A. 765, 769 (1942), rev’d sub nom. Harwich v. Commissioner, 133 F. 2d 732 (CA8), rev’d sub nom. Dobson v. Commissioner, 320 U. S. 489 (1943).
Section 111(a) provides:
“Gross income does not include income attributable to the recovery during the taxable year of a bad debt, prior tax, or delinquency amount, to the extent of the amount of the recovery exclusion with respect to such debt, tax, or amount” (emphasis added).
Ibid. See also the detailed provisions of ch. 619, Title I, § 156(a) of the Act of Oct. 21, 1942, 56 Stat. 852, amended, Act of Aug. 16, 1954, ch. 736, 68A Stat. 343, repealed, Pub. L. 94-455, Title XIX, § 1901(a)(145)(A), Act of Oct. 4, 1976, 90 Stat. 1788, establishing a mechanism for including the recoveries of previously claimed war losses in current income to the extent of tax benefit.
Consider the following passages from the regulations under § 111.
“General. Section 111 provides that income attributable to the recovery during any taxable year of bad debts, prior taxes, and delinquency amounts shall be excluded from gross income to the extent of the ‘recov*407ery exclusion’ with respect to such items. The rule of exclusion so prescribed by statute applies equally with respect to all other losses, expenditures and accruals made the basis of deductions from gross income for prior taxable years, including war losses . . . , but not including deductions with respect to depreciation, depletion, amortization, or amortizable bond premiums. . . .
“Definition of ‘recovery’. Recoveries result from the receipt of amounts in respect of the previously deducted or credited section 111 items, such as from the collection or sale of a bad debt, refund or credit of taxes paid, or cancellation of taxes accrued.” Treas. Reg. § 1.111 — 1(a), 26 CFR § 1.111 — 1(a) (1982) (emphasis added).
Consider also:
“If the taxpayer deducted a loss in accordance with the provisions of this paragraph and in a subsequent taxable year receives reimbursement for such loss, he does not recompute the tax for the taxable year in which the deduction was taken but includes the amount of such reimbursement in his gross income for the taxable year in which received, subject to the provisions of section 111, relating to recovery of amounts previously deducted.” Treas. Reg. § 1.165-l(d)(2)(iii), 26 CFR § 1.165-l(d)(2)(iii) (1982).
E. g., National Bank of Commerce v. Commissioner, 115 F. 2d 875 (CA9 1940); South Dakota Concrete Products Co. v. Commissioner, 26 B. T. A. 1429 (1932).
See Costigan, Income Taxes on Recoveries from Civil Litigation, Proceedings of the U. S. C. Tax Inst. 559, 567-570 (1954); Atlas, Tax Free Recoveries: The Tax Benefit Rule, N. Y. U. 9th Inst, on Fed. Tax. 847 (1951); Tye, The Tax Benefit Doctrine Reexamined, 3 Tax L. Rev. 329 (1948); Plumb, The Tax Benefit Rule Today, 57 Harv. L. Rev. 129, 131, n. 10, 176 (1943); Lassen, The Tax Benefit Rule and Related Problems, 20 Taxes 473, 475 (1942) (statute of limitations not a problem because “all these cases have a new element, namely, a recovery or increment in value or decrease of liability in the year in which income was determined by the Commissioner to have been received); Note, 56 Harv. L. Rev. 434, 436 (1942); Zysman, Income Derived From the Recovery of Deductions, 19 Taxes 29 (1941); Ayers, Bad Debts — Deductions and Recoveries, 18 Taxes 549 (1940).
Except for Barnett v. Commissioner, 39 B. T. A. 864 (1939), all of the early cases cited by the Court, ante, at 386-387, involved a recovery. In Estate of Block v. Commissioner, 39 B. T. A. 338, 341 (1939), the Board of Tax Appeals made the following statement:
“When recovery or some other event which is inconsistent with what has been done in the past occurs, adjustment must be made in reporting income for the year in which the change occurs. No other system would be practical in view of the statute of limitations, the obvious administrative difficulties involved, and the lack of finality in income tax liability, which would result. The foregoing principles, which have been established by the following cases, require that the refund here be included in the income of this estate for the year of recovery.”
Notwithstanding the general reference to an inconsistent event in the first quoted sentence, it is obvious from the two succeeding sentences that the Board was not intending to lay the groundwork for a new theory of the tax benefit rule. Rather, it was attempting to respond to the suggestion that the adjustment be made in the year of deduction rather than the year of recovery. This conclusion is confirmed by the fact that the third quoted sentence speaks of “the year of recovery,” not “the year of inconsistent event,” and by the fact that each of the 14 eases cited by the Board following the conclusion of the quoted passage, like Estate of Block itself and like South Dakota Concrete Products Co. v. Commissioner, supra, involved recoveries in the traditional sense.
It should be noted that even in Barnett, the Board of Tax Appeals included its discussion of inconsistent events only after emphasizing that the inclusion in income of a prior oil depletion deduction was required by Treasury Regulations that had been ratified by Congress. 39 B. T. A., at 867.
In Dobson v. Commissioner, 320 U. S. 489 (1943), the taxpayer had bought stock, sold it at a loss, and then claimed a deductible loss on his tax return. Eight years later, the taxpayer had sued for rescission of the stock purchase, claiming fraud; he settled the suit and received approxi*409mately $30,000 for the stock on which he had sustained the loss. We upheld the Tax Court’s determination that the $30,000 recovery did not need to be reported as income, since the earlier deductible losses had not reduced the taxpayer’s taxes in the year he had claimed them. For present purposes, the holding in Dobson was less significant than the way it endorsed the Tax Court’s analysis:
“The Tax Court has not attempted to revise liability for earlier years closed by the statute of limitation, nor used any expense, liability, or deficit of a prior year to reduce the income of a subsequent year. It went to prior years only to determine the nature of the recovery, whether return of capital or income.” Id., at 493 (emphasis added).
The tax benefit question was not one of inconsistent events, but whether a recovery should be characterized as return of capital or as income.
Burnet v. Sanford & Brooks Co., 282 U. S. 359 (1931), was a mirror image of this case. The taxpayer argued that a recovery of previously deducted funds should not be income because, seen from a transactional view, no net profits had been realized. The Court framed the issue as whether net profits are to be determined “on the basis of fixed accounting periods, or... on the basis of particular transactions of the taxpayer when they are brought to a conclusion.” Id., at 363. The answer was unanimous and unflinching:
“A taxpayer may be in receipt of net income in one year and not in another. The net result of the two years, if combined in a single taxable period, might still be a loss; but it has never been supposed that that fact would relieve him from a tax on the first, or that it affords any reason for postponing the assessment of the tax until the end of a lifetime, or for some other indefinite period, to ascertain more precisely whether the final outcome of the period, or of a given transaction, will be a gain or a loss.
“The Sixteenth Amendment was adopted to enable the government to raise revenue by taxation. It is the essence of any system of taxation that it should produce revenue ascertainable, and payable to the government, at regular intervals. Only by such a system is it practicable to produce a regular flow of income and apply methods of accounting, assessment, and collection capable of practical operation. It is not suggested that there has ever been any general scheme for taxing income on any other basis. . . . While, conceivably, a different system might be devised by which the tax could be assessed, wholly or in part, on the basis of the finally ascertained results of particular transactions, Congress is not required by the amend*410ment to adopt such a system in preference to the more familiar method, even if it were practicable.” Id., at 364-365.
Healy v. Commissioner, 345 U. S. 278 (1953), dealt with a more extreme effort to approximate a transactional accounting system — more closely analogous to Justice Blackmun’s approach than to the Court’s today. In that case, a taxpayer received money under a claim of right in an early year and was forced to disgorge it in a later year. The Government was perfectly willing to allow the taxpayer to take a deduction in the year of disgorgement (presumably under a “tax detriment” theory, since repayments of loans are usually not deductible), but the taxpayer wanted to be able to go back and reopen the prior year, in which he had included the receipt in income. The Court refused to allow the reopening, extolling the virtues of an annual accounting system. Id., at 284-285. See n. 2, supra.
See Rev. Rui. 62-128, 1962-2 Cum. Bull. 139; Note, 21 Vand. L. Rev. 995 (1968); Estate of Schmidt v. Commissioner, 42 T. C. 1130 (1964), rev’d, 355 F. 2d 111 (CA9 1966).
For example:
“The [tax benefit] rule rests on the notion that a taxpayer should not be permitted to retain the tax benefit of a deduction when later events demonstrate that he no longer is entitled to it. . . . To limit application of the rule to cases in which there has been an economic recovery would frustrate its purpose, which is to insure that a taxpayer not retain the benefit of a deduction to which he is no longer entitled. Fulfillment of that purpose requires application of the rule, whether the lack of need for a bad debt reserve arises from a sale or collection of accounts receivable, or merely by reason of the termination of the existence of the owner of the receivables. . . . The proper analysis of the transaction where accounts receivable are sold for their net value by a reserve method taxpayer is to restore *411the reserve to his income and accord him a loss on the sale of property. That this loss (face value less amount realized) equals the amount of the restoration to income does not militate against the basic principle that the reserve must be restored to income when it is no longer needed, irrespective of whether there has been an economic recovery.” Brief for United States in Nash v. United States, O. T. 1969, No. 678, pp. 8-14.
The Commissioner argued that (a) an accrual basis taxpayer is normally only allowed to deduct accounts receivable that are uncollectible, (b) the taxpayer had been allowed to take a deduction because the taxpayer was being allowed to represent that it would eventually find some presently collectible accounts to be uncollectible, and (c) those accounts had not in fact become uncollectible at the time they left the taxpayer’s hands. The Commissioner asserted that the earlier deduction had been allowed on the assumption that it fairly represented the taxpayer’s continuing “need” for bad debt deductions across taxable years. He argued that once the accounts receivable left the taxpayer’s hands, it had no further “need” for the bad debt reserve and that the premise for the prior deduction had become invalid. See ibid.
The taxpayer’s receipts from the sale of the business was not a recovery because, to the extent the accounts receivable were offset by the bad debt reserve, the corporation had not paid the taxpayer a penny for them. It was this fact that distinguished Nash and Schmidt from earlier cases applying the tax benefit rule to bad debt reserves such as Arcadia Savings & Loan Assn. v. Commissioner, 300 F. 2d 247 (CA9 1962), Citizens Federal Savings & Loan Assn. v. United States, 154 Ct. Cl. 305, 290 F. 2d 932 (1961), West Seattle National Bank v. Commissioner, 288 F. 2d 47 (CA9 1961), and S. Rossin & Sons, Inc. v. Commissioner, 113 F. 2d 652 (CA2 1940).
See Rev. Rul. 74-396, 1974-2 Cum. Bull. 106; Feld, The Tax Benefit of Bliss, 62 B. U. L. Rev. 443 (1982); Tennessee-Carolina Transportation, Inc. v. Commissioner, 65 T. C. 440 (1975), aff’d, 582 F. 2d 378 (CA6 1978), cert. denied, 440 U. S. 909 (1979). Contra, O’Hare, Application of Tax Benefit Rule in New Case Threatens Certain Liquidations, 44 J. Taxation 200 (1976); Broenen, The Tax Benefit Rule and Sections 332, 334(b)(2) and 336, 53 Taxes 231 (1975).
Nash and Dobson are the only tax benefit rule cases ever decided in this Court. On one other occasion, the Court invoked the tax benefit rule by analogy. United States v. Shelly Oil Co., 394 U. S. 678, 688, and n. 5 (1969).
In relevant part, 26 U. S. C. § 162(a) provides that “[tjhere shall be allowed as a deduction all the ordinary and necessary expenses paid or incurred during the taxable year in carrying on any trade or business . . . .”
In relevant part, 26 U. S. C. § 336 provides that “no gain or loss shall be recognized to a corporation on the distribution of property in partial or complete liquidation.”
In relevant part, 26 U. S. C. § 334 provides that “[i]f. . . property was acquired by a shareholder in the liquidation of a corporation in cancellation or redemption of stock, and . . . the extent to which gain was recognized was determined under § 333, then the basis shall be the same as the basis of such stock cancelled or redeemed in the liquidation, decreased in the amount of any money received by the shareholder, and increased in the amount of gain recognized to him.” The relevant implementing regulations are found at Treas. Reg. § 1.334-2, 26 CFR § 1.334-2 (1982).
The Court’s partial quotation of Treas. Reg. § 1.162-3, 26 CFR 1.162-3 (1982), ante, at 395, suggests that it may regard the deduction for unconsumed feed as improper because the feed was not “actually consumed and used in operation in the taxable year,” ibid. Of course, if the Court really believed that such a deduction should not be allowed, the proper course of action would be to modify the rules authorizing the deduction, see n. 31, infra, rather than to modify the tax benefit rule.
It should also be noted that the potential for an untaxed step-up, which may give rise to a second deduction in cases such as this, is analytically the *415same as the potential for a double deduction that was present in Nash. See Brief for United States in Nash v. United States, O. T. 1969, No. 678, pp. 20-31.
The Solicitor General made this timing point explicit in Nash. “Since a reserve for bad debts represents losses that are estimated will be sustained in subsequent taxable years, . . . any unabsorbed amounts in such a reserve must be restored to income when ... it becomes clear that the taxpayer will not suffer some or all of the estimated losses as a result of the uncollectibility of accounts receivable.” Id., at 8-9.
The flaws in the Court’s approach are exemplified by its discussion, ante, at 384-385, of a hypothetical situation involving a tenant who has paid the entire cost of a 30-day lease that straddles two taxable years on December 15 of the first year.
The Court first invites consideration of what tax consequences would result if the premises burn down in January of the second year. I would think it obvious that a taxpayer does not realize income under such circumstances, and the Court manages to accommodate this result to its theory. Even though the original explanation for the deduction (the business would make use of the premises) is no longer valid, the Court finds no fundamental inconsistency because “the loss is attributable to the business.”
The Court goes through this exercise in order to reach the next hypothetical, wherein the taxpayer voluntarily stops using the leasehold for a business purpose during the second year. Having assumed that the entire cost of the lease was deductible during the first year, the Court now de-*416dares that the tax benefit rule must be invoked to prevent a tax inequity. The Court’s methodology in this regard is quite revealing. It has presumed the validity of the deduction in the first year, citing Zaninovich v. Commissioner, 616 F. 2d 429 (CA9 1980). Yet Zaninovich is still being debated in the lower courts, in part because of hypothetical cases such as this one, and has not been endorsed by either the Commissioner or the Tax Court. See Keller v. Commissioner, 79 T. C. 7, 40, n. 24 (1982); Dunn v. United States, 468 F. Supp. 991, 994, and n. 17 (SDNY 1979) (Weinfeld, J.). See also Van Raden v. Commissioner, 71 T. C. 1083, 1107 (1979) (suggesting a distinction between “period” costs and “product” costs), aff’d, 650 F. 2d 1046 (CA9 1981). Thus, the Court creates its own problem by blithely allowing a deduction in the initial period, with the intention of continuously second-guessing that decision in subsequent years. I do not see the advantage of this approach over Justice Blackmun’s suggestion, which is criticized by the Court ante, at 378-380, n. 10. Instead, I would prefer to think carefully about whether or not the deduction should be allowed in the first place (taking into account such factors as the ease with which a lease can be prorated, the likelihood of nonbusiness uses, and the bright-line recovery rule) and, if that results in a decision to grant the deduction, to abide by whatever consequences follow from the application of traditionally accepted tax principles.
The Court suggests, ante, at 381-383, that a recovery requirement is both too narrow and too broad to give certain guidance. The Court suggests it is too narrow because the Court believes that the cancellations of indebtedness found in Mayfair Minerals, Inc. v. Commissioner, 456 F. 2d 622 (CA5 1972), Bear Manufacturing Co. v. United States, 430 F. 2d *417152 (CA7 1970), Haynsworth v. Commissioner, 68 T. C. 703 (1977), and G. M. Standifer Construction Corp. v. Commissioner, 30 B. T. A. 184 (1934), “[do] not fit within any ordinary definition of ‘recovery.’ ” Ante, at 382. I disagree. As the Court concedes, ibid., cancellation of a legally enforceable liability quite obviously increases the taxpayer’s net worth. The Code therefore explicitly requires a discharge of indebtedness to be included in income. § 61(a)(12). Cf. § 108. It does no damage to the English language to say that a taxpayer who has previously incurred an expense by assuming a liability recovers that expense when the liability is canceled.
The Court suggests that the term “recovery” is too broad because two courts have claimed to find a recovery in situations the Court finds surprising. Tennessee-Carolina Transportation, Inc. v. Commissioner, 582 F. 2d 378 (CA6 1978) (alternative holding); First Trust and Savings Bank of Taylorville v. United States, 614 F. 2d 1142 (CA7 1980). Since I believe both cases were wrongly decided (Tennessee-Carolina applied the tax benefit rule to a case closely analogous to Bliss Dairy, and First Trust applied the rule to a ease closely analogous to Hillsboro Bank), I do not find the Court’s criticism any more persuasive than I would find a suggestion that someone might incorrectly think there was a “recovery” in the cases before us today.
E. g., §§ 1245, 1250 (mere dispositions of certain depreciable property and certain depreciable realty may give rise to income).
It is noteworthy that the Court cites no authority for the assumption— critical under its interpretation of the tax benefit rule — that if the liquidation of Bliss Dairy, Inc., had occurred in 1972, the deduction for purchased but unconsumed cattle feed would have been disallowed.
In a footnote, ante, at 402, n. 37, the Court suggests that it is not addressing the issue of whether some figure less than the amount previously deducted might be appropriate. Two possibilities have been sug*420gested: lesser-of-prior-deduction-and-current-fair-market-value, see Tennessee Carolina Transportation, Inc. v. Commissioner, 65 T. C. 440, 448 (1975), and lesser-of-prior-deduction-and-shareholder’s-basis, see Feld, The Tax Benefit of Bliss, 62 B. U. L. Rev. 443 (1982). Given the uniform tenor of the Court’s theory that the tax benefit rule serves to “ ‘cancel out’ an earlier deduction,” ante, at 388,1 think it unlikely that it would endorse either possibility.
This view, of course, conforms to what a recovery theory would dictate. See Rosen v. Commissioner, 611 F. 2d 942 (CA1 1980) (taxpayer realizes the value of recovered asset, determined at time of recovery).
If it were so inclined, it could modify § 162(a) to provide that no deduction would be allowed for the purchase of materials and supplies; instead, a deduction would be allowed only at the time of consumption. Such a sentiment clearly underlies the Court’s statement, ante, at 395, that “[t]he deduction is predicated on the consumption of the asset in the trade or business.”
Alternatively, it could provide that a purchase of materials and supplies not be considered an “ordinary and necessary expense” to the extent it includes items that will probably not be consumed during the taxable year. As the Solicitor General notes in his brief before this Court, “income might *421be more accurately reflected through the use of inventory accounting for such supplies,” Brief for United States in No. 81-930 and for Respondent in No. 81-485, pp. 37-38. It bears mention that the Commissioner presently takes the position that an expenditure for feed that will be consumed in a subsequent year will not be allowed unless three tests are satisfied: it must be a payment (not a refundable deposit), it must be made for a business purpose and not merely for tax avoidance, and the deduction must not result in a material distortion of income. Rev. Rul. 79-229, 1979-2 Cum. Bull. 210. Cf. Treas. Reg. 1.162-12, 26 CFR §1.162-12 (1982). The courts have divided over whether the Commissioner’s position is consistent with the present § 162(a). Compare Clement v. United States, 217 Ct. Cl. 495,580 F. 2d 422 (1978), cert. denied, 440 U. S. 907 (1979); Dunn v. United States, 468 F. Supp. 991 (SDNY 1979) (Weinfeld, J.) (supporting the Commissioner), with Frysinger v. Commissioner, 645 F. 2d 523 (CA5 1981); Commissioner v. Van Raden, 650 F. 2d 1046 (CA9 1981) (supporting the taxpayer).
In 1975, Congress had before it, but was unable to pass, such legislation. H. R. 10936, 94th Cong., 1st Sess. (1975).
In particular, it could prohibit the allocation of any shareholder basis to an expensed asset, thereby completely eliminating the possibility of the step-up discussed supra, at 413. Indeed, the Internal Revenue Service could do so consistently with the present text of § 334 by modifying Treas. Reg. § 1.334-2, 26 CFR § 1.334-2 (1982).
Because I disagree with the Court’s conclusion that Bliss Dairy, Inc., realized income upon liquidation, I obviously do not reach the issue of how §336, the nonrecognition provision, should be construed. I would observe, however, that in order to justify its conclusion, the Court is forced to override the plain language of § 336. See ante, at 397-402; n. 20, supra. The Court justifies this course of action by invoking the lower court decisions that have held the nonrecognition language of § 337 superseded by the tax benefit rule in the context of recoveries. E. g., Commissioner v. Anders, 414 F. 2d 1283 (CA10), cert. denied, 396 U. S. 958 (1969); Anders v. United States, 199 Ct. Cl. 1, 462 F. 2d 1147, cert. denied, 409 U. S. 1064 (1972). Those cases are relevant because when Congress enacted § 337, it hoped to allow taxpayers to enjoy some of the tax benefits of § 336 even when they do not precisely satisfy the formal prerequisites for applying *422that section. But assuming that the Anders construction of §337 is correct (a point this Court has never decided), I would think the tail wags the dog if one construes §336 in light of §337, rather than vice versa. Cf. Tennessee-Carolina Transportation, Inc. v. Commissioner, supra, at 453 (Tannenwald, J., dissenting) (“Section 337 was designed to be a shield for taxpayers and not a sword to be used against them in applying other sections of the Code”).