MR. JUSTICE STEWART delivered the opinion of the Court.

Section 172 of the Internal Revenue Code of 1954, as amended, provides that a "net operating loss" experienced by a corporate taxpayer in one year may be carried as a deduction to the preceding three years and the succeeding five years to offset taxable income of those years. The entire loss must be carried to the earliest possible year; any of the loss that is not "absorbed" by that first year may then be carried in turn to succeeding years. The respondent, Foster Lumber Co., sustained a net operating loss of some $42,000 in 1968, which it carried back to 1966. In 1966 the respondent had had ordinary income of about $7,000 and a capital gain of about $167,000. The question presented is whether a loss carryover is "absorbed" by capital gain as well as ordinary income or is instead limited to offsetting only ordinary income. The taxpayer filed a refund suit in Federal District Court challenging the Commissioner's disallowance of its claim that the $35,000 of the 1968 loss not used to offset its 1966 ordinary income survived to reduce its 1967 tax liability. The trial court and the Court of Appeals for the Eighth Circuit agreed with the taxpayer. We granted certiorari to resolve a Circuit conflict on a recurring question of statutory interpretation.

I

The dispute in this case centers on the meaning of "taxable income" as used in § 172(b)(2) to govern the amount of carrybacks and carryovers that can be successively transferred from one taxable year to another. In relevant part, § 172(b)(2) requires the net operating loss to be carried in full to the earliest taxable year possible, and provides: "The portion of such loss which shall be carried to each of the other taxable years shall be the excess, if any, of the amount of such loss over the sum of the taxable income for each of the prior taxable years to which such loss may be carried." Thus when the loss has been carried back to the first year to which it is applicable, the loss "survives" for carryover to a succeeding taxable year only to the extent that it exceeds the taxable income of the earlier year. "Taxable income" is defined in § 63(a) of the Code to mean "gross income, minus the deductions allowed by this chapter." Gross income is in turn defined by § 61(a) of the Code as "all income from whatever source derived," and specifically includes "[g]ains derived from dealings in property." On its face the concept of "taxable income" thus includes capital gains as well as ordinary income. In the absence of a specific provision excluding capital gains, it thus appears that both capital gain and ordinary income must be included in the taxable income that § 172 directs must be offset by the loss deduction before any loss excess can be found to be available for transfer forward to the succeeding taxable year.

The respondent argues that the Code's prescribed method for calculating the taxes due on its taxable income conflicts with this natural reading of § 172. The Code provides two methods for computing taxes due on corporate income, and a corporation is under a statutory duty to employ the method that results in the lower tax. 26 U.S.C. § 1201 (a). Under § 11, the "regular method," ordinary income and capital gains income are added together to produce taxable income; during the period at issue a 22% tax rate was then imposed on the first $25,000 of taxable income and the remainder was taxed at a 48% rate. Section 1201(a) of the Code prescribes the "alternative tax," calculated in two steps and applied when resulting in a lower tax liability for the corporation. The first step computes a partial tax on the taxable income reduced by the net long-term capital gain at the regular corporate rates imposed by § 11. This step effectively subjects only ordinary income to the partial tax. The second step imposes a 25% tax on the net long-term capital gain. The alternative tax is the sum of the partial tax and the tax on capital gain. In practical terms, the alternative tax does not redefine taxable income, but it does result in a much lower effective tax rate for corporations whose income is in whole or substantial part composed of capital gain. It thus extends to corporations the longstanding statutory policy of taxing income from capital gain at a lower rate than that applicable to ordinary income.

The problem from the respondent's point of view is that the mechanics of the alternative tax work in such a way that the potential benefit of the loss deduction may not be fully reflected in reduced tax liability for the taxable year to which the loss is carried. The problem arises when, as in 1966 for the respondent, the "alternative method" governs the calculation of tax liability, and the ordinary income effectively subject to the partial tax under the first step is less than the loss deduction subtracted from it. The Code does not permit the excess loss to be subtracted from the capital gain income before the second step is carried out. Under the alternative method, therefore, the tax benefit of the loss deduction is effectively lost for the carryover year to the extent that it exceeds the ordinary income in that year. This can be seen simply by considering the taxpayer's circumstances in this case. Subtracting the loss deduction of $42,203.12 from the 1966 ordinary income of $7,236.05 under Step 1 resulted in a negative balance of $34,967.07; no partial tax was imposed and the 25% rate on the $166,634.81 of capital gains under Step 2 produced a tax of $41,658.70. If the loss deduction had been merely $7,236.05, and thus exactly offset the $7,236.05 of ordinary income, however, the tax due would still have been $41,658.70. The taxpayer therefore asserts that only $7,236.05 of the loss deduction was actually "used" in 1966 and that $34,967.07 remained to be carried forward to reduce its tax liability in 1967.

There can be no doubt that if the "regular method" had been applicable to the respondent's taxes in 1966, the loss deduction ($42,203.12) would have been fully "used" to offset capital gains ($166,634.81) as well as ordinary income ($7,236.05), leaving $131,667.74 to be taxed, and a tax bill of $58,200.52. It is clear that the alternative tax produced the lower tax liability despite the inability to fully "use" the loss deduction; the lower tax resulted directly from the favorable rate of taxation of capital gain income prescribed by the alternative method. The question is whether the two "tax benefit" provisions relied on by the respondent — low capital gain taxation under the alternative method and the loss carryback provision — must each be maximized independently of the other or whether Congress instead anticipated that the benefit provided by the loss deduction might on occasion be subsumed in the greater benefit provided by the alternative tax computation method.

Section 172 does not explicitly address the question of fit between these two tax benefits, providing simply that "[t]he portion of such loss which shall be carried to each of the other taxable years shall be the excess, if any, of the amount of such loss over the sum of the taxable income for each of the prior taxable years to which such loss may be carried." The respondent contends, and the Tax Court in Chartier Real Estate Co. v. Commissioner, 52 T.C. 346, aff'd per curiam, 428 F.2d 474 (CA1), held, that the phrase "to which such loss may be carried" modifies "taxable income" as well as "each of the prior taxable years." The Tax Court in the Chartier case further held that "`taxable income' in this context (as modified by the above phrase) means that taxable income to which the loss is actually applied in computing actual tax liability." 52 T. C., at 357-358. In other words, it was held, taxable income refers only to that ordinary income offset by a loss deduction that produces an additional reduction in tax liability under the alternative tax computation method.

It is, of course, not unusual in statutory construction to find that a defined term's meaning is substantially modified by an attached clause. But reading "taxable income to which . . . such loss may be carried" as equivalent to "taxable income to which such loss may be carried and deducted, resulting in a reduction of tax liability" gives these phrases a synergistic effect that goes well beyond their natural import. Such a construction subtly redefines "taxable income" in terms of the tax impact of a particular method of tax calculation. It thus implicitly departs from the "term of art" definition of taxable income given in § 63(a), while discovering a significance in the word "carry" that goes well beyond its usual connotation of a transfer of a loss from the year in which it occurred. Standing alone, this strained reading of the statute's terms falls considerably short of the explicit statutory support the Court has previously required of taxpayers seeking a tax benefit from losses suffered in other years. See, e. g., Woolford Realty Co. v. Rose, 286 U.S. 319, 326. If Congress had intended to allow a loss deduction to offset only ordinary income when the alternative tax calculation method is used, it could easily have said so.

II

The respondent further asserts that the legislative history and the broad policy behind the loss deduction section of the Code support its interpretation of "taxable income" under § 172(b). Although, for the reasons stated above, it can hardly be said that the benefit claimed by the respondent is fairly within the statutory language, it is not inappropriate to consider this contention — to consider, in short, whether "the construction sought is in harmony with the statute as an organic whole." See Lewyt Corp. v. Commissioner, 349 U.S. 237, 240.

The respondent relies on the Court's opinion in Libson Shops, Inc. v. Koehler, 353 U.S. 382, 386, for a description of the legislative purpose in allowing loss carryovers. In that case the Court said that the net operating loss carryover and carryback provisions "were enacted to ameliorate the unduly drastic consequences of taxing income strictly on an annual basis. They were designed to permit a taxpayer to set off its lean years against its lush years, and to strike something like an average taxable income computed over a period longer than one year."

There were, in fact, several policy considerations behind the decision to allow averaging of income over a number of years. Ameliorating the timing consequences of the annual accounting period makes it possible for shareholders in companies with fluctuating as opposed to stable incomes to receive more nearly equal tax treatment. Without loss offsets, a firm experiencing losses in some periods would not be able to deduct all the expenses of earning income. The consequence would be a tax on capital, borne by shareholders who would pay higher taxes on net income than owners of businesses with stable income. Congress also sought through allowance of loss carryovers to stimulate enterprise and investment, particularly in new businesses or risky ventures where early losses can be carried forward to future more prosperous years.

The respondent focuses on the equalizing purposes of § 172 to argue that the Commissioner's insistence on the absorption of the loss deduction by capital gain income is inconsistent with § 172's primary purpose of avoiding the subjection of similarly situated taxpayers to significantly different treatment solely on the basis of arbitrary timing. This argument is based on the observation that, unless it is accepted, the taxpayer's ability to fully benefit from the loss carryover deduction will turn on whether ordinary income in the first year to which the loss may be carried exceeds or is less than the loss deduction. If the ordinary income exceeds the loss, the taxpayer will get the full benefit of the deduction; if the ordinary income is less than the loss, the shortfall will be absorbed by capital gain income without providing an incremental tax reduction.

Congress may, of course, be lavish or miserly in remedying perceived inequities in the tax structure. While there is no doubt that Congress through the loss carryover provisions did intend to reduce the arbitrariness inherent in a taxing system based on annual accounting, the history of the loss offset provision does not support the respondent's vision of a Congress seeking perfection in the realization of its objective.

Over the years, Congress has shifted the definition of both the kinds of losses and the kinds of income that may be used in calculating the loss offset, indicating its ability in this area of the Internal Revenue Code as in others to make precise definitions and later to modify them in pursuing its broad policy goals. For example, Congress in 1924 specifically provided that a noncorporate taxpayer could use the excess of a loss deduction over ordinary income to reduce the amount of capital gain subject to tax, thus permitting full "use" of the loss deduction by the taxpayer. The inference can be drawn that Congress was aware of the potential "waste" of the deduction otherwise and acted to prevent it. That provision was in turn left out of the 1939 Code, leading to the contrary inference that Congress was aware of the "waste" of the deduction but decided not to remedy it.

The 1939 revision of the Code, in fact, tolerated even further "waste" of the loss deduction, providing not only that the loss must be offset against net income (ordinary income and capital gains), but that tax-exempt interest income must also be included in income that the loss was required to offset. This provision had the same arbitrary policy consequences that the respondent decries under the alternative tax computation method applicable here. It required the loss deduction to be "used up" in offsetting tax-exempt income, thus "wasting" a portion of the loss deduction's capacity to reduce overall tax liability. And it made the utility of the loss deduction turn on the accidents of timing. The loss deduction would be "wasted" in offsetting tax-exempt income realized in an early year, while if the tax-exempt income were not realized until a later year the full tax benefit of the loss deduction could have been garnered. Such results cut against any assertion that the loss-deduction provisions have consistently been used completely to minimize arbitrary timing consequences, and indicate that Congress has not hesitated in this area to limit taxpayers to the enjoyment of one tax benefit even though it could have made them eligible for two.

The 1954 Internal Revenue Code continued the 1939 Code's definition of ordinary and capital gain income as subject to setoff by the § 172 loss deduction. Although several substantive changes in the loss-deduction section were made and commented on in the legislative reports accompanying the 1954 Code, there was no indication that the addition to § 172(b) of the phrase "to which such loss may be carried" was meant to signal a willingness to condition the loss deduction's life on its ability to produce full tax benefits for the taxpayer. In view of the predecessor statutes' tolerance of a taxpayer's inability to maximize the tax benefit of a loss deduction, and the complete failure of the Committee Reports in any way to indicate the shift in policy the respondent claims to discern in the 1954 Code revision, the legislative history simply does not support the respondent's contention that the addition in 1954 of the phrase "to which such loss may be carried" was intended to eliminate the requirement that the loss deduction be used to offset capital gain under the alternative tax computation method.

We turn finally to an examination of § 172(b) in the context of the statute as it exists today. If the statute could be viewed as consistently minimizing the arbitrariness of timing consequences, a construction of § 172(b) inconsistent with that approach might be suspect. Section 172 as a whole has not, however, been drafted with the single-minded devotion to reducing arbitrary timing consequences that the respondent urges should control the decision in this case.

The most telling example of Congress' failure to remedy all timing accidents that "rob" a taxpayer of the full benefit of the loss deduction can be found in § 172(c). That provision defines a "net operating loss" as "the excess of the deductions allowed by this chapter over the gross income." A taxpayer does not have a loss for a particular year unless its deductions exceed its ordinary income and its capital gains. When an ordinary income loss is experienced in a year of negligible capital gains it gives rise to a net operating loss that can be carried over to other years. If that same ordinary income loss comes in a year when the net capital gains exceed that loss, there is no net operating loss under the statute to carry to another year. Because the statute also forbids setting off that ordinary loss against the capital gains before the capital gain tax is computed under the alternative method, the loss' potential tax benefit is arbitrarily "lost" to the taxpayer solely as a result of accidents of timing. Congress, of course, can and occasionally has in the past treated loss years differently from carryover years. But if Congress were intent on substantially eliminating accidents of timing from the calculation of income on an average basis, it would hardly have tolerated such a departure from that purpose at the very inception of the tax benefit provided by § 172.

The respondent's argument is further undercut by the holding in Chartier Real Estate Co., not challenged here, that the statute forbids using a loss deduction to offset capital gain income in a loss carryover year. If such an offset were permitted, the taxpayer would benefit by a further reduction in its capital gain tax liability already calculated at a preferential rate. The respondent in effect asks this Court to infer from that deliberate denial of the limited tax benefit that would accrue from using the loss to offset preferentially taxed capital gains, that Congress implicitly meant to confer the even greater tax benefit of using the loss to offset ordinary income taxed at the higher regular rates. In a statutory section that part by part manages explicitly to detail loss calculations on one hand and deductions on the other, such a leap in statutory construction must be much more firmly grounded in a consistently articulated and achieved congressional purpose than can be discerned here.

The respondent's broad argument, in short, boils down to a contention that "harmony with the statute as an organic whole" can be achieved in this area only by reading the Code provision so as to give the greatest possible benefits to all taxpayers. For the reasons we have discussed, that is a contention that cannot be accepted.

The judgment is

Reversed.