delivered the opinion of the Court.

During its tax year ending December 31,1958, respondent refunded $505,536.54 to two of its customers for overcharges during the six preceding years. Respondent, an Oklahoma producer of natural gas, had set its prices during the earlier years in accordance with a minimum price order of the Oklahoma Corporation Commission. After that order was vacated as a result of a decision of this Court, Michigan Wisconsin Pipe Line Co. v. Corporation Comm’n of Oklahoma, 355 U. S. 425 (1958), respondent found it necessary to settle a number of claims filed by its customers; the repayments in question represent settlements of two of those claims. Since respondent had claimed an unrestricted right to its sales receipts during the years 1952 through 1957, it had included the $505,536.54 in its gross income in those years. The amount was also included in respondent’s “gross income from the property” as defined in § 613 of the Internal Revenue Code of 1954, the section which allows taxpayers to deduct a fixed percentage of certain receipts to compensate for the depletion of natural resources from which they derive income. Allowable percentage depletion for receipts from oil and gas wells is fixed at 27%% of the “gross income from the property.” Since respond*680ent claimed and the Commissioner allowed percentage depletion deductions during these years, 27%% of the receipts in question was added to the depletion allowances to which respondent would otherwise have been entitled. Accordingly, the actual increase in respondent’s taxable income attributable to the receipts in question was not $505,536.64, but only $366,513.99. Yet, when respondent made its refunds in 1958, it attempted to deduct the full $505,536.54. The Commissioner objected and assessed a deficiency. Respondent paid and, after its claim for a refund had been disallowed, began the present suit. The Government won in the District Court, 255 F. Supp. 228 (D. C. N. D. Okla. 1966), but the Court of Appeals for the Tenth Circuit reversed, 392 F. 2d 128 (1968). Upon petition by the Government, we granted certiorari, 393 U. S. 820 (1968), to consider whether the Court of Appeals decision had allowed respondent “the practical equivalent of double deduction,” Charles Ilfeld Co. v. Hernandez, 292 U. S. 62, 68 (1934), in conflict with past decisions of this Court and sound principles of tax law. We reverse.

I.

The present problem is an outgrowth of the so-called “claim-of-right” doctrine. Mr. Justice Brandeis, speaking for a unanimous Court in North American Oil Consolidated v. Burnet, 286 U. S. 417, 424 (1932), gave that doctrine its classic formulation. “If a taxpayer receives earnings under a claim of right and without restriction as to its disposition, he has received income which he is required to return, even though it may still be claimed that he is not entitled to retain the money, and even though he may still be adjudged liable to restore its equivalent.” Should it later appear that the taxpayer was not entitled to keep the money, Mr. Justice Brandéis explained, he would be entitled to a deduction in the year of repayment; the taxes due for the year of receipt would *681not be affected. This approach was dictated by Congress’ adoption of an annual accounting system as an integral part of the tax code. See Burnet v. Sanford & Brooks Co., 282 U. S. 359, 365-366 (1931). Of course, the tax benefit from the deduction in the year of repayment might differ from the increase in taxes attributable to the receipt; for example, tax rates might have changed, or the taxpayer might be in a different tax “bracket.” See Healy v. Commissioner, 345 U. S. 278, 284-285 (1953). But as the doctrine was originally formulated, these discrepancies were accepted as an unavoidable consequence of the annual accounting system.

Section 1341 of the 1954 Code was enacted to alleviate some of the inequities which Congress felt existed in this area.1 See H. R. Rep. No. 1337, 83d Cong., 2d Sess., *68286-87 (1954); S. Rep. No. 1622, 83d Cong., 2d Sess., 118-119 (1954). As an alternative to the deduction in the year of repayment2 which prior law allowed, §1341 (a)(5) permits certain taxpayers to recompute their taxes for the year of receipt. Whenever § 1341 (a)(5) applies, taxes for the current year are to be reduced by the amount taxes were increased in the year or years of receipt because the disputed items were included in gross income. Nevertheless, it is clear that Congress did not intend to tamper with the underlying claim-of-right doctrine; it only provided an alternative for certain cases in which the new approach favored the taxpayer. When the new approach was not advantageous to the taxpayer, the old law was to apply under § 1341 (a)(4).

In this case, the parties have stipulated that § 1341 (a) (5) does not apply. Accordingly, as the courts below recognized, respondent’s taxes must be computed under § 1341 (a) (4) and thus, in effect, without regard to the special relief Congress provided through the enactment of § 1341. Nevertheless, respondent argues, and the Court of Appeals seems to have held, that the language used in § 1341 requires that respondent be allowed a deduction for the full amount it refunded to its customers. We think the section has no such significance.

*683In describing the situations in which the section applies, § 1341 (a)(2) talks of cases in which “a deduction is allowable for the taxable year because it was established after the close of [the year or years of receipt] that the taxpayer did not have an unrestricted right to such item . . . The “item” referred to is first mentioned in § 1341 (a)(1); it is the item included in gross income in the year of receipt. The section does not imply in any way that the “deduction” and the “item” must necessarily be equal in amount. In fact, the use of the words “a deduction” and the placement of § 1341 in subchapter Q — the subchapter dealing largely with side effects of the annual accounting system — make it clear that it is necessary to refer to other portions of the Code to discover how much of a deduction is allowable. The regulations promulgated under the section make the necessity for such a cross-reference clear. Treas. Reg. on Income Tax (1954 Code) § 1.1341-1 (26 CFR § 1.1341-1). Therefore, when § 1341 (a) (4) — the subsection applicable here — speaks of “the tax . . . computed with such deduction,” it is referring to the deduction mentioned in § 1341 (a) (2); and that deduction must be determined, not by any mechanical equation with the “item” originally included in gross income, but by reference to the applicable sections of the Code and the case law developed under those sections.

II.

There is some dispute between the parties about whether the refunds in question are deductible as losses under § 165 of the 1954 Code or as business expenses under § 162.3 Although in some situations the distinction may have relevance, cf. Equitable Life Ins. Co. of *684Iowa v. United States, 340 F. 2d 9 (C. A. 8th Cir. 1965), we do not think it makes any difference here. In either case, the Code should not be interpreted to allow respondent “the practical equivalent of double deduction,” Charles Ilfeld Co. v. Hernandez, 292 U. S. 62, 68 (1934), absent a clear declaration of intent by Congress. See United States v. Ludey, 274 U. S. 295 (1927). Accordingly, to avoid that result in this case, the deduction allowable in the year of repayment must be reduced by the percentage depletion allowance which respondent claimed and the Commissioner allowed in the years of receipt as a result of the inclusion of the later-refunded items in respondent’s “gross income from the property” in those years. Any other approach would allow respondent a total of $1.27% in deductions for every $1 refunded to its customers.

Under the annual accounting system dictated by the Code, each year’s tax must be definitively calculable at the end of the tax year. “It is the essence of any system of taxation that it should produce revenue ascertainable, and payable to the government, at regular intervals.” Burnet v. Sanford & Brooks Co., supra, at 365. In cases arising under the claim-of-right doctrine, this emphasis on the annual accounting period normally requires that the tax consequences of a receipt should not determine the size of the deduction allowable in the year of repayment. There is no requirement that the deduction save the taxpayer the exact amount of taxes he paid because of the inclusion of the item in income for a prior year. See Healy v. Commissioner, supra.

Nevertheless, the annual accounting concept does not require us to close our eyes to what happened in prior years. For instance, it is well settled that the prior year may be examined to determine whether the repayment gives rise to a regular loss or a capital loss. Arrow-*685smith v. Commissioner, 344 U. S. 6 (1952). The rationale for the Arrowsmith rule is easy to see; if money was taxed at a special lower rate when received, the taxpayer would be accorded an unfair tax windfall if repayments were generally deductible from receipts taxable at the higher rate applicable to ordinary income. The Court in Arrowsmith was unwilling to infer that Congress intended such a result.

This case is really no different.4 In essence, oil and gas producers are taxed on only 72%% of their “gross income from the property” whenever they claim percentage depletion. The remainder of their oil and gas receipts is in reality tax exempt. We cannot believe that Congress intended to give taxpayers a deduction for refunding money that was not taxed when received. Cf. O’Meara v. Commissioner, 8 T. C. 622, 634—635 (1947). Accordingly, Arrowsmith teaches that the full amount of the repayment cannot, in the circumstances of this case, be allowed as a deduction.

This result does no violence to the annual accounting system. Here, as in Arrowsmith, the earlier returns are not being reopened. And no attempt is being made to require the tax savings from the deduction to equal the *686tax consequences of the receipts in prior years.5 In addition, the approach here adopted will affect only a few cases. The percentage depletion allowance is quite unusual; unlike most other deductions provided by the Code, it allows a fixed portion of gross income to go untaxed. As a result, the depletion allowance increases in years when disputed amounts are received under claim of right; there is no corresponding decrease in the allowance because of later deductions for repayments.6 Therefore, if a deduction for 100% of the repayments were allowed, every time money is received and later repaid the taxpayer would make a profit equivalent to the taxes on 27%% of the amount refunded. In other situations when the taxes on a receipt do not equal the tax benefits of a repayment, either the taxpayer or the Government may, depending on circumstances, be the beneficiary. Here, the taxpayer always wins and the Government always loses. We cannot believe that Congress would have intended such an inequitable result.

The parties have stipulated that respondent is entitled to a judgment for $20,932.64 plus statutory interest for *687claims unrelated to the matter in controversy here; the District Court entered a judgment for that amount. Accordingly, the judgment of the Court of Appeals is reversed and the case is remanded to that court with instructions that it be returned to the District Court for re-entry of the original District Court judgment.

Reversed and remanded.

dissenting.

I share Mr. Justice Stewart’s views as to this case and add only a word.

If we sat in chancery reviewing tax cases, much of what the Court says would have appeal. But we do not sit to do equity in tax cases; that is one of Congress’ main concerns.

The search for equity in the tax laws is wondrous and elusive. As Edmond Cahn said: “[T]hose only are equal whom the law has elected to equalize.” E. Cahn, The Sense of Injustice 14 (1949).

Percentage depletion had its roots in granting a reward to men who go into undeveloped territory in search of oil and gas. But today it is granted anyone who has an interest in oil or gas; the beneficiary need not live the life of the oil wildcatter or bear his risks to obtain the benefits of percentage depletion.

When it comes to capital gains what “equities” are to be applied? Is it fair that earned income pay a heavier tax?

A son who spends $1,000 on his destitute father does not get the same tax benefit as he who pays a like sum to his alma mater. Louis Eisenstein pursues example after example of so-called inequities in tax laws in his book The Ideologies of Taxation (1961). For example, the profits on the sale of unbred pigs are taxable as ordinary income, while the profits on the sale of pigs once bred *688are taxable as capital gains. Id., 174. The same is true of turkeys but not of chickens, even though “a bred chicken and a bred turkey are similarly situated. Each has feathers and two legs." Ibid.

Treasury recently noted numerous basic inequities resulting in preferred tax treatment for some people’s dollars. Tax Reform Studies and Proposals, U. S. Treasury Dept-., Joint Publication of House Committee on Ways and Means and Senate Committee on Finance, 91st Cong., 1st Sess., pt. 1, pp. 13-17 (Comm. Print 1969).

Apart from certain aspects of percentage depletion were the reduced taxation on long-term capital gains and the exclusion of interest on state and local government bonds. The examples are legion. The Tax Reform study gives an unusual example:

“An individual had a total income of $1,284,718 of which $1,210,426 was in capital gains, the remaining $74,292 from wages, dividends, and interest. He excluded one-half of his capital gains, which he is allowed to do under present law, thereby reducing his present law (adjusted gross) income to $679,405 (after allowing for the $100 dividend exclusion). From this income he subtracted all his personal deductions, which amounted to $676,419 and which included $587,693 for interest on funds borrowed presumably for the purpose of purchasing the securities on which the capital gains were earned. As a result, after allowing $1,200 of personal exemptions his taxable income was reduced to $1,786 and he paid a tax of $274. His overall tax rate, therefore, was about two-hundredths of one percent.” Id., at 15.

This was made possible by using a taxpayer’s deductions only against that part of his income which is subject to the tax, ignoring the excluded part.

*689Tax laws are indeed arbitrary; the lines they draw are the products of pressures inside the Congress with compromises carrying the day.

The Court of Appeals held that the “item” here in question was properly included in “gross income” prior to 1958 and was an allowable “deduction” in 1958 because the taxpayer did not have “an unrestricted right” to a “portion of such item,” and that the amount of such deduction exceeds $3,000 — all as provided in § 1341.1 Skelly Oil Co. v. United States, 392 F. 2d 128, 131.

There is no irregularity on the face of the return. There is no conflict with any decision of any other Court of Appeals. We are asked, however, to put a gloss on the statute that Treasury desires. I would adhere to the construction given by the Court of Appeals leaving to Congress the correction of any inequities in the tax scheme.

*690The Congress many years ago created the Joint Committee on Internal Revenue Taxation, which is a standing committee. 26 U. S. C. §§ 8001-8005, 8021-8023. One of its statutory mandates is “[t]o investigate the operation and effects of the Federal system of internal revenue taxes.” Id,., § 8022.

In that connection a recent report states:

“[T]he Joint Committee staff has in recent years been used as a committee liaison with the Treasury Department in working on tax proposals for the committee. The staff aids the two tax committees in explaining provisions, in writing committee reports, and in aiding in the drafting of bills.”

The Joint Committee makes regular reports to Congress for revision of the tax laws. Inequities that arise as a result of interpretations that are given existing laws either at the administrative or judicial level can be quickly corrected by this agency of oversight.2

Treasury unhappily has developed the habit of jockeying in the courts, testing one theory against another. In California, it may take one position and in Massachusetts the opposite position, the issue in each being the same. The hope is that conflicts over litigious and important issues will develop and the case will be brought here.3

If we were trained in the art and science of taxation, we might serve a useful function. But taxation is a *691specialty in which we have only sporadic and no continuous experience. It has been said that one of our decisions is like a “lightning bolt” that “illuminates only a very small portion of the landscape,” leaving a darkness that later decisions do not remove. R. Paul, Studies in Federal Taxation 249-250 (3d series 1940). Our contributions, if such they can be called, are dubious indeed, for the Joint Committee can and does rewrite the Code frequently.

It is therefore the rare tax case4 we should consider, except the even rarer constitutional case. The present case has no constitutional overtones; the taxpayer followed the words of the tax law literally, using no new or strained construction of words to find a tax advantage; there is no conflict between this case and any other decision. The Solicitor General only claims that the result reached by the Court of Appeals does not fit the neat logic which he finds in a group of related tax cases.

An account of the cost, confusion, and inequity in tax administration that ensues while everyone waits for a conflict among the Circuits (which takes at least 10 years) is related in Griswold, The Need for a Court of Tax Appeals, 57 Harv. L. Rev. 1153 (1944). The role we presently play was stated as follows:

“Our present system of tax adjudication inevitably leaves nearly every question uncertain during the entire period while it must be dealt with, usually in thousands of instances, by the administrative officers. And yet that is just the period when there should be an authoritative rule if the system is to work smoothly, effectively, speedily, fairly, and *692without discrimination. Under our present system delay and discrimination are typical and inevitable.” Id., at 1161.

In absence of an unmistakably clear conflict among the Circuits, I would abide by the opinions of the Courts of Appeals in tax cases and leave to the Joint Committee whether the gloss which Treasury now tries to put on the statute is or is not desirable.

dissenting.

The Court today denies the respondent a tax benefit fairly provided by the Code for no other discernible reasons than that, under the statute as written, “the taxpayer always wins and the Government always loses,” 1 and that “the approach here adopted will affect only a few cases.” Ante, at 686. But we are not free, even in a few cases, to abandon settled principles of annual accounting and statutory construction merely to avoid what the Court thinks Congress might consider an “inequitable result.” 2

“[T]he rule that general equitable considerations do not control the measure of deductions or tax benefits cuts both ways. It is as applicable to the *693Government as to the taxpayer. Congress may be strict or lavish in its allowance of deductions or tax benefits. The formula it writes may be arbitrary and harsh in its applications. But where the benefit claimed by the taxpayer is fairly within the statutory language and the construction sought is in harmony with the statute as an organic whole, the benefits will not be withheld from the taxpayer though they represent an unexpected windfall.” Lewyt Corp. v. Commissioner, 349 U. S. 237, 240.

From any natural reading of § 1341, it is apparent that Congress believed the “deduction” in § 1341 (a) (2) would be in the amount of the “'item” described in § 1341 (a)(1). If that understanding is not manifest from the face of the statute and the legislative history,3 it is the unavoidable inference from a study of the pre-1954 law which *694the Court concedes § 1341 (a) (4) was intended to codify. In every case in this area previously decided by the Court the amount deductible in the year of repayment was considered to be exactly the same as the amount of the previously included item. In two of the cases most sharply in congressional focus in 1954, the Government had conceded without hesitation that the taxpayers were “entitled to a deduction for a loss in the year of repayment of the amount earlier included in income.” Healy v. Commissioner, 345 U. S. 278, 284. See also United States v. Lewis, 340 U. S. 590, 591. That has been the express position of the Treasury since at least 1936,4 and the Court today has not cited a single instance of deviation from that understanding.

The Court says that § 1341 is not alone controlling and that “it is necessary to refer to other portions of the Code to discover how much of a deduction is allow*695able.” Ante, at 683. I agree that § 1341 must be considered in the context of the Internal Revenue Code as an “organic whole.” But no other provisions of the Code in any manner bolster the Court’s argument. The Court assumes, quite correctly, that either § 162 or § 165 does permit a deduction for the refund. But it does not, and cannot, suggest that either of those sections— or any other statutory provision — limits the amount of the deduction for the undeniable loss of profits in the full amount of the repayment. Instead the Court assumes a broad equitable authority to weed out tax benefits which it calls “double deductions” — a characterization wholly inapposite to the facts of this case.

In prior decisions disallowing what truly were “double deductions,” the Court has relied on evident statutory indications, not just its own view of the equities, that Congress intended to preclude the second deduction. In those cases the taxpayers sought to benefit twice from the same statutory deduction.5 In this case, by con*696trast, the respondent has taken two different deductions accorded by Congress for distinct purposes. In the years 1952 through 1957 it deducted the proper amounts for depletion — a deduction which is allowed by Congress “on the theory that the extraction of minerals gradually exhausts the capital investment in the mineral deposit,” and which is “designed to permit a recoupment of the owner’s capital investment in the minerals so that when the minerals are exhausted, the owner’s capital is unimpaired.” Commissioner v. Southwest Exploration Co., 350 U. S. 308, 312. The respondent’s 1958 deduction was granted by Congress for the entirely different reason that the refund of previously reported income constituted a loss, or business expense. In purpose and effect the deductions are wholly unrelated, and each is sustainable on its own merits. Certainly it cannot be said either that the respondent did not in fact exhaust the capital assets for which the deductions were allowed in 1952 through 1957 or that it did not suffer a business loss by the 1958 repayment.

The sole nexus between these distinct transactions on which the Court constructs its “double deduction” theory is that the depletion deductions were computed as a percentage of gross income from the property. But this fact cannot distinguish percentage depletion from any other deduction. If the respondent had elected to take cost depletion in 1952 through 1957, for example, there would also have been a portion of the gross income in those years — perhaps less than 27%%, perhaps more— which was not included in taxable income. Whether a deduction is computed as a fixed percentage of income or *697in some other manner, it always reduces by some percentage the income which is ultimately taxed. There are other deductions, of course, whose amount is a function of a certain percentage of the taxpayer’s income. With respect to the individual taxpayer, the standard 10% deduction, § 141, and those for charitable contributions, § 170, and medical expenses, § 213, are doubtless the most frequent. Under the Court’s ruling today, any taxpayer who repays money included in gross income in a prior year in which he also took one of the above mentioned deductions will have to reduce his refund deduction by that portion of the previous year’s deduction attributable to the included income. Surely this result contravenes the purpose of the annual accounting concept to prevent recomputations of the prior year’s tax.

The Court says today that there can be no deduction “for refunding money that was not taxed when received.” Ante, at 685. This means nothing less than that, whenever a taxpayer seeks to deduct a refund of money received as income under a claim of right in a prior year, the deduction must be reduced by the percentage of gross income in that prior year which, for whatever reason, was not also taxable income. Otherwise there will be precisely the same kind of so-called “double deduction” as the Court finds in this case.

It is clear that the Court has wrought a major transformation of the deduction which has heretofore been allowed and which Congress recognized in § 1341 (a)(4). That deduction is permitted because, in the words of § 1341, the item “was included in gross income for a prior taxable year” (emphasis added), not because it was included in taxable income. It is no answer to say that the “annual accounting concept does not require us to close our eyes to what happened in prior years.” *698Ante, at 684. Of course we must look to the prior years to ascertain the amounts included in gross income and the nature of that income as it bears on the provision under which it is deductible in the year of repayment. Arrowsmith v. Commissioner, 344 U. S. 6.6 But the very purpose of the annual accounting concept is to preclude adjustments in the amount of the deduction to reflect the tax consequences of the item's inclusion in the prior year.

“Congress has enacted an annual accounting system under which income is counted up at the end of each year. It 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. This basic principle cannot be changed simply because it is of advantage to a taxpayer or to the Government in a particular case that a different rule be followed.” Healy v. Commissioner, 345 U. S. 278, 284-285.

One of the major factors, in addition to changes in tax rates and brackets, that determine who will benefit from adherence to the annual accounting principles embodied in § 1341 (a) (4) is the extent to which the taxpayer had deductions in the prior or subsequent taxable years to offset gross income. And it is no less incon*699sistent with annual accounting principles to pare down the allowable loss deduction in the year of repayment because of other deductions in the year of inclusion than because of a lower tax rate or bracket in that year.

Because I cannot agree that the Court’s equitable sensibilities empower it to depart from the sound principles of tax accounting specifically endorsed by Congress in § 1341, I respectfully dissent.