UNITED STATES OF AMERICA, PETITIONER V. THE GOODYEAR TIRE AND RUBBER COMPANY AND AFFILIATES No. 88-1474 In the Supreme Court of the United States October Term, 1989 On Writ of Certiorari to the United States Court of Appeals for the Federal Circuit Reply Brief for the United States 1. Respondent mistakenly asserts (Br. 5-7) that the government in its opening brief has taken "new position(s)" for the first time in this Court. There is manifestly no substance to respondent's suggestion (Br. 6-7, 16, 38) that the government has argued that "accumulated profits" in this case should be computed on an aggregate basis, as under the 1986 legislation, rather than on an annual basis. In our opening brief (at 4 n.4, 17 n.6, 25 n.11), we repeatedly note that the Section 902 credit for the years at issue is computed on an annualized basis and that the dividends must be "sourced" to the profits of the years out of which they are paid. We also explain (Br. 7 n.5) that the difference between respondent's and the government's calculation of tax liability is caused in large part by a dispute over whether the dividend in question can be "sourced" to respondent's relatively high foreign tax years of 1968 and 1969. We do, however, question respondent's statement (Br. 16) that the concept of "sourcing" is "critical" to the resolution of this case; it is actually quite secondary to the question presented. The critical question here is whether the term "accumulated profits" in Section 902 of the Internal Revenue Code (26 U.S.C.) -- i.e., the profits to which the dividends are "sourced" -- is determined in accordance with principles of U.S. law or of foreign law. Once that determination is made, the "sourcing" calculation and the application of the Section 902 apportionment formula follow as a matter of course. There is likewise no merit to respondent's assertion (Br. 5) that, in stating that "accumulated profits" should be computed according to the principles governing the determination of "earnings and profits" under U.S. law, the government is arguing for a different definition from that reflected in the stipulation. As respondent states (Br. 5), the language used in the stipulation was "computed under United States income tax concepts," and this phrase clearly encompasses the concept of "earnings and profits." Nothing in the stipulation suggests an intent not to invoke that concept, which is the one explicitly identified in the relevant regulation (Treas. Reg. Section 1.902-3(c) (1971 ed.)) and understood by the commentators (see, e.g., E. Owens, The Foreign Tax Credit 121-122 (1961)), by the Tax Court (H.H. Robertson Co. v. Commissioner, 59 T.C. 53, 78-79 (1972), aff'd, 500 F.2d 1399 (3d Cir. 1974)), and by Congress (see U.S. Br. 39 n.17) to be the benchmark for defining "accumulated profits." /1/ 2. Conspicuously absent from respondent's submission is any analysis of the congressional purpose in establishing the allocation formula of Section 902. Respondent merely states that the purpose of the foreign tax credit in general is to prevent double taxation, and respondent also notes that the indirect credit is intended to equate the treatment of subsidiary operations with that of branch operations -- in the limited sense that the domestic owner of a branch always receives full credit for the foreign tax in the year paid. See Resp. Br. 7, 11-13, 20. This incomplete view of Section 902 conveniently leads to the conclusion that the allocation formula is to be interpreted so as to maximize the allowable credit; in respondent's words, "the computation that produces the greater credit must necessarily approach most closely the result applicable to a branch operation" (Resp. Br. 47). Obviously, respondent's limited view of the purpose of Section 902 is mistaken. If Congress wanted to advance the goal of maximizing the credit (because branch operations get the full credit in the year the tax is paid), it could simply have allowed the parent of a foreign subsidiary to take the full credit. There would be no need for the apportionment formula of Section 902 at all. But there is a need for apportionment, and respondent completely ignores its basic function -- namely, to prorate the foreign tax deemed paid when the subsidiary distributes less than all of its earnings to the U.S. parent. Congress realized that when a subsidiary does not distribute all of its earnings to its parent, the parent is not currently taxed on those earnings and should not be entitled to a credit for all of the foreign tax paid. In that situation, in sharp contrast to the branch operation, the parent has not received all of the earnings subjected to the foreign tax; some of those earnings have been retained by the foreign subsidiary and thereby insulated from U.S. tax. Thus, as we explained in detail in our opening brief (at 26-28 & n.12), the purpose of Section 902 is to guard not only against double taxation, but also against undertaxation. In cases of partial distributions, Congress determined that equal treatment of branch and subsidiary operations would be achieved if the parent could take a partial credit; if no credit were allowed, the parent would be subjected to double taxation, but if the full credit were allowed, there would be undertaxation and the subsidiary operation would receive preferential treatment as compared to the branch operation. The allocation formula of Section 902 is designed to allow a prorated credit that bears the same proportional relationship to the total foreign tax paid as the dividend distribution bears to the total earnings. Respondent seeks to have subsidiary operations treated the same as branch operations with respect to the amount of the credit received, without focusing on whether the two operations are reporting an equivalent amount of income. But it is critical to the goal of nondiscrimination to consider the income that is distributed to the U.S. parent. In the case of a branch operation, all of the earnings are immediately received by the parent and subjected to U.S. taxation. In the case of a subsidiary operation, like respondent's, the profits earned by the foreign subsidiary are generally not included in the income of the U.S. taxpayer until they are distributed as dividends. As this Court has noted, "(s)uch dividends, not its subsidiary's profits, constitute it income to be returned for taxation." American Chicle Co. v. United States, 316 U.S. 450, 452 (1942). Here, during the years in question, respondent's subsidiary did not distribute all of the earnings that could have been distributed as dividends; rather, it retained a substantial portion of those earnings as surplus. Respondent's approach to the apportionment formula essentially ignores these retained earnings (to the extent they exceed the income recognized under foreign law) or pretends that they have been distributed (see U.S. Br. 45-46), thus allowing respondent both to receive credit for all of the foreign tax paid by the subsidiary and also to have its subsidiary retain earnings immune from U.S. taxation. That approach would give the subsidiary operation a decided advantage over a branch operation in derogation of Congress's clear purpose in enacting Section 902. See also point 7, infra. 3. Respondent's mistake in focusing solely on the credit received in a branch operation, but not on the income subjected to U.S. tax, leads it to a fallacious premise that seriously undermines its argument. Respondent states (Br. 12) that a 40% British tax rate is the "touchstone" for determining whether its tax credit is "consistent" with that of a branch operation. Respondent then proceeds to argue (id. at 25-27) that the government's approach to Section 902 must be incorrect because it yields an effective British tax rate significantly lower than 40%. Respondent confuses the "actual" foreign tax rate with the "effective" foreign tax rate applied to U.S. income. In fact, when income computed under British law differs from that computed under U.S. law, the effective British tax rate on U.S. income will never equal 40%, and the divergence from the actual rate identified by respondent is precisely what should be expected if the Section 902 credit is properly serving its function of promoting equal treatment of branch and subsidiary operations. The British tax is laid at 40% on income as computed by Britain; it necessarily follows that, in cases where the U.S. definition of income differs from that of Britain, the British tax will not be 40% of the U.S. income. Clearly, it is the effective rate imposed on U.S. income that is significant to the taxation of a branch operation. Foreign concepts of taxable income are irrelevant in determining how much of the earnings of a foreign branch constitute income to its domestic owner; those earnings flow directly to the owner and are included in its income under U.S. tax principles. Thus, if the income of the branch is higher under United States law than under British law, the foreign tax imposed at 40% on the foreign tax base necessarily will be less than 40% of the larger sum that the owner must report as income on its United States return. And, consequently, the direct tax credit available to the owner of the branch will also be less than 40% of the income. /2/ It is this "effective" rate of foreign tax, i.e., the foreign tax paid expressed as a percentage of the branch's income computed under United States law, that is relevant in measuring whether the Section 902 tax credit for a domestic company with a British subsidiary is essentially equivalent to that available to a domestic company with a branch. /3/ Contrary to respondent's contention, examining the effective rate of tax on the facts of this case demonstrates the correctness of the government's position. The Section 902 credit that the Commissioner allowed to respondent is entirely consistent with the credit that would have been granted had it been conducting the same business as a branch. If respondent had operated its business through a branch, it would have received a credit for all of the foreign tax paid by the branch, but it also would have been required to include on its U.S. income tax return all of the branch's profits as determined under United States law -- the loss and carryback allowed only by Britain could not have been used to offset those profits. The result in 1970 and 1971 would have been a relatively small British tax (and therefore tax credit) because of the reduced British tax base, yet a relatively large amount of U.S. income; consequently, the effective British tax rate on the total income of a branch operation would have been quite small. Because respondent operated its business through a subsidiary, its U.S. income for these years includes only the dividends distributed. When the dividends are "sourced to these years, however, comparability with a branch requires the parent to recognize income according to U.S. concepts, which will yield the same small effective British tax rate that would have occurred in the case of a branch operation. On these facts, computation of "accumulated profits" under U.S. concepts yields an effective tax rate of only 19.9% for 1970 and 1.6% for 1971, which is precisely what respondent would have paid in 1970 and 1971 had it operated as a branch. See Resp. Br. 26; J.A. 31. /4/ The reason that the rates for those years were so low is that the generous British loss and carryback allowances virtually eliminated the British tax base for 1971 and substantially reduced it for 1970. 4. As we explain in detail in our opening brief (at 15-25), the apportionment formula of Section 902 is incongruous unless accumulated profits are defined in terms of U.S. principles. When a foreign subsidiary makes a distribution to its domestic parent, the concept of "profits" is relevant at three stages of the Section 902 inquiry. First, the U.S. parent must determine how much of the distribution is a dividend reportable as income on its U.S. return, which requires determining the extent to which it was paid out of "earnings and profits." Second, the amount of "dividends" to be placed in the numerator of the Section 902 fraction must be determined, i.e., the dividends paid out of "accumulated profits." Third, the amount of "accumulated profits" to be placed in the denominator of the fraction must be determined. It is our submission that all of these quantities must be determined in a consistent manner, which means according to U.S. tax principles. The "earnings and profits" relevant to the first determination clearly are determined according to U.S. law pursuant to Sections 312 and 316 of the Internal Revenue Code (26 U.S.C.). If the apportionment formula is to fulfill its purpose, the "dividends" in the numerator of the fraction must correlate with the "dividens" reported on the U.S. tax return, i.e., be based on profits computed under U.S. principles. And, as the Claims Court noted (Pet. App. 21a), "(a)lgebraic consistency" then demands that the profits in the denominator of the fraction also be based on U.S. tax principles. Respondent agrees with much of this analysis. Respondent agrees that the determination of the amount of dividends reported on the parent's U.S. return must be based upon profits computed under U.S. tax principles (see Resp. Br. 37, 44; see also National Foreign Trade Council Amicus Br. 29). Respondent also appears to agree that "algebraic consistency" demands that accumulated profits in the numerator be determined according to the same principles as in the denominator (see Br. 46). Respondent diverges from the above analysis, however, in insisting (Br. 44) that U.S. law does not govern the definition of the numerator. Although respondent does not explicitly state what does govern the definition of "dividends" in the numerator, presumably its position is that they are distributions made out of profits as determined by foreign law. This position is untenable. The ultimate goal of Section 902 is to determine the appropriate tax credit that should be allowed on the parent's U.S. return to reflect the foreign tax that is "deemed paid" on the dividend income reported on that return. In the words of the statute, the foreign tax shall be allowed as a credit in "the same proportion" as the "amount of such dividends * * * bears to the amount of such accumulated profits" (Section 902(a)(1)). Clearly, the dividends in the fraction must correlate with the dividends reported on the U.S. return in order to obtain the correct proportion for allowing a partial credit against the U.S. income tax on those dividends. As this Court explained in American Chicle Co. v. United States, 316 U.S. 450 (1942), the foreign tax is to be "proportioned to the dividends made available to the parent in this country" (id. at 454), which means that "the numerator is the dividends received by the parent" (id. at 452). If the dividends in the ratio do not correlate with the dividends being taxed on the parent's return, then the apportionment exercise becomes meaningless. /5/ See generally E. Owens, The Foreign Tax Credit 121-122 (1961). Thus, as the Tax Court has held on several occasions (see U.S. Br. 19-20), the "dividends" in the numerator of the Section 902 fraction must be determined on the basis of profits computed under U.S. concepts. /6/ If so, the principle of "algebraic consistency" recognized by respondent demands that the term "accumulated profits" in the denominator must also be defined under U.S. concepts, and therefore the decision below cannot stand. 5. Respondent relies heavily (Br. 20-24) on two examples (numbered 2 and 3), which it claims illustrate the distortions caused by the Commissioner's interpretation of Section 902. Amicus curiae National Foreign Trade Council presents its own examples (Br. 15-17, 21-25), also claiming that they demonstrate the error of the Commissioner's approach. These examples will not bear the weight ascribed to them, however, for they do not illustrate anything about the general operation of the Section 902 credit. The crucial feature common to all of these examples is hypothesizing a year in which the subsidiary pays foreign tax, but has zero profits under U.S. principles. Because a dividend cannot be "sourced" to a year in which there are no accumulated profits, the foreign tax paid in that year can never be matched to a dividend and therefore it can never be credited as "deemed paid." Thus, the narrow point illustrated by the various examples is that defining accumulated profits in terms of U.S. concepts yields an anomalous result in one specific situation: where foreign tax is paid in a year in which the subsidiary's accumulated profits under U.S. concepts are zero, the credit for the foreign tax paid in that year will be "blocked," resulting in some degree of double taxation after a full dividend distribution. The examples, however, illustrate nothing about a case in which there is not a zero year; so long as there exists even one dollar in after-tax profits under U.S. concepts, the anomalous result highlighted by respondent will disappear because the Section 902 formula will credit the foreign tax paid in that year in full when a dividend distribution is "sourced" to the U.S. profits for that year. These examples provide no basis for affirmance of the judgment below. First, respondent's assertion (Br. 22) that one of the examples "presents the very essence of Goodyear's case" is insupportable. The theoretical possibility of an unrecoverable payment of foreign tax is not present in most cases, and it certainly is not present here. The accumulated profits of respondent's subsidiary for 1970 and 1971, as computed by the Commissioner under U.S. law, are far in excess of zero. The Commissioner's computation already credited respondent with a portion of the foreign taxes paid for those years (based on the ratio of the dividends paid to the accumulated profits), and the rest of those taxes, as well as the foreign taxes paid in 1968 and 19969 whose credit respondent seeks to accelerate, remain fully accessible to be claimed as credits whenever the balance of the retained profits may be distributed. See J.A. 31. /7/ Thus, these examples bear no relation to the facts of this case. In any event, the "zero income" anomaly illustrated by respondent's examples provides no support should for the conclusion that accumulated profits should be determined under foreign law principles. At most, the examples show that Section 902 does not work perfectly in a carefully structured set of circumstances that is not likely to occur frequently. If that anomaly is unacceptable, the appropriate way to address it is through a narrowly tailored adjustment, not to misinterpret and eviscerate the statute by defining accumulated profits in terms of foreign law principles. As we have explained, respondent's interpretation would incorrectly apportion the credit, not just in unusual "zero income" cases, but in every case in which profits are different under U.S. and foreign law concepts. Indeed, even with respect to the specific problem of a "blocked" or unrecoverable credit, respondent's interpretation is inferior. If accumulated profits are determined according to foreign law, in every case in which the foreign definition of income yields a larger amount than the U.S. definition, the taxpayer will be restricted from recovering a portion of the foreign tax paid. This is because a distribution in the full amount of the U.S. income will yield only a partial credit, and any additional distribution will not be a dividend (because it is not out of "earnings and profits") and therefore will not give rise to any Section 902 credit. See U.S. Br. 22; point 7, infra. Thus, although the "zero income" anomaly has long been recognized, it has never been suggested that it should be "cured" by defining accumulated profits in terms of foreign law. See, e.g., B. Bittker & J. Eustice, Federal Income Taxation of Corporations and Shareholders Paragraph 17.11, at 17-44, 17-47 (5th ed. 1987); 1 E. Owens & G. Ball, The Indirect Credit 164-165, 174 (1975). /8/ Clearly, this discrete hypothetical situation, which is not presented here, cannot form the basis for resolving one of the fundamental questions relating to the general applicability of Section 902. 6. Respondent argues that the term "total taxable income" used in the original 1918 version of Section 902 in lieu of "accumulated profits" clearly refers to a foreign definition of income, and therefore the statute at issue here should be construed the same way. See Resp. Br. 30; National Foreign Trade Council Amicus Br. 13-14; Vulcan Materials Co. Amicus Br. 5-11. The premise of this argument is highly questionable. /9/ As we have explained, it would have been illogical for Congress to have based the statute's apportionment formula upon a foreign definition of income because that approach would not correctly relate the apportioned credit to the dividend reported on the U.S. return, which comes out of U.S. profits. Moreover, since the statute was designed to provide a credit equivalent to branch operations, which report income determined under U.S. concepts, it would have upset the balance for Congress to have contemplated the use of foreign profits in allocating the credit to subsidiary operations. Because the concept that Congress had to incorporate in the statute was a hybrid -- a measurement of the profits of a foreign corporation intended to be used to determine the appropriate tax credit allowed to a U.S. corporation -- no commonly used term in the Code was appropriate, and it was reasonable for Congress to invent a new one. /10/ The phrase that it chose, "total taxable income," does not by its terms clearly refer either to foreign or U.S. concepts of income, and there is no persuasive evidence that Congress, simply by using this term in the 1918 legislation, deliberately chose the illogical and counterproductive course of using a foreign definition of income. /11/ Similarly, Congress's failure in 1962 to amend Section 902 to conform with language being used in the new Subpart F gives rise to no inference that Section 902 refers to profits computed under foreign law principles. Cf. National Foreign Trade Council Amicus Br. 10-12. The meaning of "accumulated profits" was by that time recognized as referring to U.S. law, and there was no reason for Congress to tamper with the existing language. See, e.g., I.T. 2676, XII-1 C.B. 48, 50 (1933); Steel Improvement & Forge Co. v. Commissioner, 36 T.C. 265, 276-282 (1961), rev'd on other grounds, 314 F.2d 96 (6th Cir. 1963). Indeed, the 1962 legislative history shows Congress's understanding that Section 960 of the Code would operate under "rules * * * consistent with Section 902," which means that Congress must have understood that "accumulated profits" are computed under U.S. tax concepts. H.R. Rep. No. 1447, 87th Cong., 2d Sess. A103-A104 (1962). And that same understanding is reflected in the history of subsequent enactments in the foreign tax credit area. /12/ 7. In our opening brief (at 21-22), we provide some simple examples to illustrate how respondent's interpretation of Section 902 would disrupt the proper operation of the credit by failing to correlate the percentage of the credit allowed with the percentage of earnings distributed by the foreign subsidiary. Respondent contends (Br. 39-40; see also National Foreign Trade Council Amicus Br. 18-20) that these examples are misleading, arguing that the incorrect credits yielded by the use of a foreign definition of "accumulated profits" in the government's examples are a consequence not of respondent's misinterpretation of Section 902, but rather of the government's failure to use the same foreign tax rate in all of the examples. This objection is completely misconceived. The problem illustrated by the examples is that, under respondent's interpretation, disparities between the U.S. and foreign tax base would yield either an overstatement or understatement of the credit that should be apportioned to a particular dividend. Specifically, when the U.S. tax base is lower than the foreign tax base, the credit would be too low, resulting in double taxation; the taxpayer would be denied full credit for the foreign taxes paid even when it distributes all of its earnings as a dividend. Conversely, when the U.S. tax base is higher than the foreign tax base, the credit would be too high; the taxpayer would be able to take the full credit when it does not distribute all of its earnings as a dividend. These faulty results are illustrated by the individual examples standing alone, not by a comparison of the various examples to a "base case," and therefore the results clearly are not attributable to a failure to use the same foreign tax rate in each example. Thus, the same point can be illustrated by changing the examples (U.S. Br. 21-22) to use a constant 20% foreign tax rate. In the second example, where the foreign definition of income yields $180 and the U.S. definition yields $100, the foreign tax paid at a 20% rate will be $36, which leaves after-tax U.S. profits of $64. If all of that $64 of U.S. profits is distributed as a dividend, the taxpayer should receive the full credit of $36. That is precisely the result obtained under the Section 902 formula if accumulated profits are defined according to U.S. concepts: $36 x $64/($100 - $36) = $36. But under respondent's approach of defining accumulated profits in terms of foreign tax law concepts, the taxpayer receives only a partial credit even though the subsidiary has distributed 100% of the available earnings and profits as a dividend. If $180 is used as the accumulated profits in the Section 902 formula, the taxpayer will receive credit for less than half of the foreign taxes paid: $36 x $64/($180 - 36) = $16. /13/ Moreover, the foreign tax that is not credited here may be permanently blocked because any distribution that exceeds the profits defined under U.S. law, by definition, will not be a dividend and thus will not give rise to any Section 902 credit (unless the foreign tax base in subsequent years is lower than the U.S. tax base). If the U.S. tax base is greater than the foreign tax base, the opposite distortion will occur -- namely, the taxpayer will be able to take a credit for all of the foreign tax paid even if the subsidiary does not distribute all of its earnings. Using the numbers from the third example in our opening brief with a 20% foreign tax rate, i.e., $100 of U.S. income and $60 of foreign income, the foreign tax will be $12, leaving $88 in after-tax U.S. profits. By distributing only $48 of those profits, the taxpayer would be able to take a credit for all of the $12 of foreign tax paid if respondent's interpretation of Section 902 were correct. The formula would be as follows: $12 x $48/($60 - $12) = $12. This result would clearly defeat the basic statutory purpose of the Section 902 allocation formula to prevent the parent-subsidiary operation from taking the full credit when the subsidiary has retained some of the earnings (thereby insulating them from U.S. tax). /14/ Conversely, this purpose is properly served if accumulated profits are defined in terms of U.S. tax concepts; in this example, that approach will permit the taxpayer to take a credit for slightly more than half of the foreign tax paid, which it can apply against the U.S. taxation of a dividend that reflects a distribution of slightly more than half of the subsidiary's earnings. The formula would be as follows: $12 x $48/($100 - $12) = $6.55. 8. Respondent's contention (Br. 31-37, 41-43) that the decision below does not conflict with the longstanding administrative practice is unresponsive to our prior filings in this case and entirely without merit. The crabbed construction respondent gives to Treas. Reg. Section 1.902-3(c)(1) (see Br. 31-33) has no basis in its text. Moreover, respondent's argument ignores the acknowledged fact (see Resp. Br. 42) that Rev. Rul. 63-6, 1963-1 C.B. 126, published just two years before the regulation, squarely holds that "U.S. rules should govern the determination of 'accumulated profits.'" /15/ The regulation plainly carries forward the substance of the revenue ruling -- namely, that "accumulated profits," like "earnings and profits," denote the source from which dividends are paid and must be calculated in accordance with U.S. principles of taxation. See U.S. Br. 33-34 & n.14. /16/ For the foregoing reasons, and those stated in our opening brief, the judgment of the court of appeals should be reversed. Respectfully submitted. KENNETH W. STARR Solicitor General OCTOBER 1989 /1/ As we noted in our opening brief (at 25 n.11), the concepts of "accumulated profits" and "earnings and profits" are not identical because the former are computed on a year-by-year basis. Hence, there is no force to respondent's reliance (Br. 27-28) on the failure of Section 902 to use the term "earnings and profits" or on the fact that the statute uses the phrase "to the extent * * * dividends are paid * * * out of accumulated profits." See, e.g., H.H. Robertson Co. v. Commissioner, 59 T.C. at 78-79, 86-87. /2/ For example, suppose the United States does not allow a $40 deduction that is allowed by Britain and therefore the income of the branch is defined as $100 under United States law but as $60 under British law. The British tax will be $24 ($60 x 40%), and the owner of the branch will be entitled to a direct foreign tax credit in that amount. This credit is only 24% of the income of the branch that is subject to United States taxation. The converse is also true -- if the profits of the branch are lower under United States law than under British law, the foreign tax will be greater than 40% of the profits that constitute income to the domestic owner (although Section 904 prevents the owner from taking a credit that exceeds the U.S. income tax attributable to the income from the foreign country (see Resp. Br. 20)). /3/ This divergence of the "effective tax rate" from the stated foreign tax rate is a well recognized phenomenon. See, e.g., 1 E. Owens & G. Ball, The Indirect Credit 169-170) (1975). This effective rate is not a "synthetic tax rate in no way related to reality" (Resp. Br. 25); to the contrary, it reflects the true burden of the foreign tax on the earnings reported as income on the U.S. return. Indeed, one of the major reasons for the 1986 change in the computation of the Section 902 credit was to eliminate the ability of multinational corporations to obtain a tax advantage by manipulating their dividend distributions to take advantage of fluctuations in the effective rate. Congress found that many companies sought to maximize the credit by having their foreign subsidiaries accumulate earnings in years when the effective foreign tax rate was low and distribute the earnings in years when it was high, and Congress eliminated the annualization feature of the credit largely to end the use of this "rhythm method," noting that fluctuations in "effective foreign tax rate(s)" created opportunities for tax avoidance. See S. Rep. No. 313, 99th Cong., 2d Sess. 305-306 (1986). We note that Congress's recognition of the widespread use of this "rhythm method," which can be used only if "accumulated profits" are computed in accordance with U.S. concepts of taxation or if the actual foreign tax rate fluctuates, further demonstrates how settled was the interpretation of Section 902 that has been overturned by the court of appeals below. /4/ In 1970, a branch operation would have had to recognize 3,812,473 in income, computed under principles of U.S. law. The British tax for that year was 758,437, yielding an effective rate of 19.9%. In 1971, the branch would have recognized U.S. income of 3,224,486 and paid British tax of only 52,567. For that year, the effective tax rate -- and therefore the effective rate of the credit -- would have been only 1.6%. See J.A. 31. Thus, the figures computed by respondent on page 26 of its brief show that the Section 902 formula works properly when "accumulated profits" are defined in terms of U.S. concepts. Conversely, if the credit is computed on the basis of an effective British tax rate of 40%, it will bear no relation to the credit that would have been available to a branch operation. /5/ For example, suppose a subsidiary pays $80 in British tax on income computed as $200 under British law and as $1000 under U.S. law. If the subsidiary distributes 50% of its after-tax earnings and profits to the parent ($460), one would expect Section 902 to allow the parent to take 50% of the foreign tax ($40) as a credit on its return. But under respondent's erroneous definition of "accumulated profits," the numerator of the fraction -- the "dividends" paid out of "accumulated profits" -- would be only $120 (the after-tax earnings of the subsidiary according to foreign law principles). A ratio containing this figure cannot be expected to yield the correct proportional credit to be applied against the tax on the $460 dividend; in this example, respondent's approach would allow the parent to take the full $80 credit on this 50% distribution ($80 x $120/($200 - $80) = $80). /6/ Respondent seeks to blunt this argument by asserting that the amount of foreign taxes paid "is part of the numerator" (Br. 46), and therefore the denominator must also be determined in accordance with foreign tax principles. This assertion is incorrect. The Section 902 fraction measures the ratio of dividends to after-tax profits. The amount of foreign taxes paid simply represents the maximum credit, to which this ratio is applied to determine the proportional credit. In mathematical terms, the amount of tax paid is the "multiplicand" (American Chicle Co. v. United States, 316 U.S. at 452), not part of the fraction, and there is no logical reason why the derivation of the amounts in the ratio should be affected by the manner in which the tax is computed. /7/ There is similarly no merit to respondent's effort to analogize this case to one where there is a difference in U.S. and foreign rules regarding the timing of a loss carryback (see Br. 22-23). Example 3, and respondent's references to the timing of a loss, all involve carryback of a loss that is recognized under both foreign and U.S. law, but where the laws of the two countries differ with respect to the timing of the carryback. Here, however, the discrepancy is over the availability of the loss at all. If respondent's profits are computed under U.S. principles, the loss deduction allowed by Britain in 1973 is not available. Instead the subsidiary's loss for 1973 is converted into a profit, and there is no question of carryback at all. See J.A. 24. /8/ The "zero income" anomaly has, for the most part, been cured prospectively by the 1986 amendments that eliminate the annualized approach and create a pool of accumulated earnings. See U.S. Br. 38-39 & n.17. One of the reasons for this change identified in the legislative history was to eliminate the "zero income" anomaly that could occur through the application of U.S. carryback rules. See S. Rep. No. 313, supra, at 305-306. /9/ Respondent is correct in stating (Br. 30) that there is no affirmative indication that Congress in 1921 sought to change from a foreign definition of income to one based on U.S. tax concepts. But there is also no reason to believe that Congress understood the 1918 statute to refer to foreign law concepts. We would add that there is no indication that Congress has ever intended to make a change in Section 902 from foreign to U.S. concepts, even though it is clear that the statute today unequivocally contemplates the use of U.S. concepts (see U.S. Br. 39 n.17). If the statute contemplates the use of U.S. concepts today and Congress never changed it in this respect, it would follow from respondent's line of argument that the statute has always contemplated the use of U.S. concepts. /10/ Thus, Congress could not simply have used the term "net income," as Amicus Vulcan Materials Co. suggests (Br. 7). A foreign corporation has no "U.S. corporate tax base" (ibid.), and therefore, strictly speaking, it has no "net income." /11/ Respondent places undue reliance on the Second Circuit's per curiam opinion in United Dyewood Corp. v. Bowers, 56 F.2d 603 (1932), aff'g 44 F.2d 399 (S.D. N.Y. 1930). The court there sought to apply the 1918 version of the statute to a case where the foreign subsidiary had paid two creditable foreign taxes -- an income tax and an excess profits tax -- that were imposed on overlapping foreign tax bases. While the Second Circuit stated that the term "total taxable income" referred to the foreign tax base, it is not apparent that there was any difference there between the U.S. and foreign definitions of income or that it was even suggested to the court that U.S. concepts might control; the dispute centered on the effect upon the credit of the overlap between the two tax bases. Moreover, the case was litigated and decided prior to the Treasury's statement of its position in 1933 that U.S. principles govern the definition of profits. Thus, the Second Circuit's statement concerning the meaning of a phrase that disappeared from the statute more than 65 years ago cannot be viewed as a reasoned rejection of the proposition that, where profits would be computed differently under U.S. and foreign law principles, the Section 902 credit functions properly only if "accumulated profits" are determined in accordance with U.S. principles. /12/ See S. Rep. No. 313, 99th Cong., 2d Sess. 299 (1986); Staff of Joint Comm. on Taxation, 99th Cong., 2d Sess., General Explanation of the Tax Reform Act of 1986, at 858 (Comm. Print 1987) ("accumulated profits * * * were generally calculated in accordance with the principles governing the calculation of earnings and profits for U.S. tax purposes"); H.R. Conf. Rep. No. 861, 98th Cong., 2d Sess. 841 (1984) ("accumulated profits are essentially equivalent to earnings and profits"); S. Rep. No. 1039, 96th Cong., 2d Sess. 15 (1980). /13/ This point is cogently expressed in one detailed scholarly analysis of the foreign tax credit (E. Owens, The Foreign Tax Credit 122 (1961) (footnote omitted)): (I)f all the accumulated profits of a given year are distributed and if they are distributed to one stockholder this (Section 902) fraction should have no effect on the amount of foreign tax creditable. In other words, the fraction should be one over one and the entire foreign tax should be credit. It is clear that this result will be reached only if "accumulated profits" in the denominator of the fraction is defined in exactly the same way as is the source of the "dividends" in the numerator; that is, as "earnings and profits" of the foreign corporation as computed under United States tax law. /14/ Amicus National Foreign Trade Council itself uses this example in Table V of its brief (at 20). Elsewhere in its brief, however, it recognizes the unacceptability of the result engendered by the use of a foreign definition of accumulated profits in the example. Amicus correctly states that "if a wholly owned subsidiary in a particular year distributes only a part of what is available after payment of the foreign tax, the credit should obviously not be for the full amount of the foreign tax" (Br. 5). /15/ As we have explained (Br. 34 n.14; Pet. Reply Br. 5-6 & n.2), the administrative step of declaring this ruling "obsolete" does not indicate a departure from its substance. /16/ Respondent also repeats (Br. 31, 43) its contention that subparagraph (4) of the regulation supports its position. As we explained in our reply at the petition stage (at 6-7), this subparagraph does not define accumulated profits and is fully consistent with the explicit definition in subparagraph (1). We have also earlier refuted respondent's claim (Br. 10, 33, 46) that the statutory phrase "gains, profits, and income" necessarily refers to profits computed under foreign tax principles. See U.S. Pet. Reply Br. 7 n.3.