Anderson, Clayton & Co., Plaintiff-Appellee-Cross-Appellant v. United States of America, Defendant-Appellant-Cross-Appellee

GOLDBERG, Circuit Judge: Anderson, Clayton Co. (taxpayer) brought this refund action to recover federal income taxes paid for Í964* Two discrete tax matters are involved. The first matter involves determining the geographic source of a minimum distribution to taxpayer of a foreign subsidiary’s “subpart F income” for the purpose of computing the per-eountry limitation on the foreign tax credit allowed taxpayer under I.R.C. § 904(a)(1). 1 The second matter concerns promissory notes distributed to taxpayer as dividends by a foreign subsidiary. The question is whether the taxpayer realized a deductible loss from the decline in exchange value of the notes, which were payable in a foreign currency. The first matter will turn initially on the applicability, retroactivity, and validity of a treasury regulation governing • the “sourcing” of dividends that was promulgated after the district court’s 'decision in this case. 2 We find the regulation, applicable, retroactive, and valid and r'everse the district court’s judgment for the taxpayer on this issue. The second matter will turn on the putative collateral estoppel effect of a Court of Claims decision that, contrary to accepted tax practice, allowed the taxpayer a loss deduction for the wartime decline in exchange value of foreign currency holdings that had not been converted into dollars. 3 We hold that the government was not collaterally estopped and affirm the district court’s judgment for the government on this issue. SOURCE OF A MINIMUM DISTRIBUTION OF SUBPART F INCOME I. Taxpayer is a large, widely-held corporation engaged along with its subsidiaries in activities including the merchandising of cotton, coffee, and other commodities, financing various crops, and manufacturing and selling food products. In its income tax return for fiscal 1964, 4 taxpayer, as *975 authorized by § 963 of the Code, elected to report as income a minimum distribution of earnings and profits in the amount of $4,684,233.96 from its Swiss subsidiary, Anderson, Clayton & Co., S. A. (hereinafter Lausanne). By electing to report the minimum distribution as income in 1964, the taxpayer avoided the necessity of reporting as its own income all of Lausanne’s subpart F income, as otherwise required by § 951(a)(1)(A) of the Code. The bulk of Lausanne’s subpart F income for 1964 was foreign base company sales income derived from sales of commodities grown or produced within the countries of Argentina, Brazil, and Peru. Subsidiary corporations of the taxpayer domiciled in those countries sold the commodities to Lausanne, which then resold the goods in the countries of their origin to unaffiliated customers in arms-length transactions. Lausanne paid no tax on its earnings from these sales to any of the South American countries in which it transferred title to the merchandise. It paid tax on the accrued profits only to Switzerland. When taxpayer computed its fo.reign tax credits under the per-country limitation of § 904(a)(1) of the Code, it treated $13,659 of the minimum distribution as income sourced in Argentina, $3,233,293 of the minimum distribution as income sourced in Brazil, and $15,517.49 of the minimum distribution as income sourced in Peru. That Lausanne earned those amounts in those countries is undisputed. The Commissioner determined, however, that for the purpose of computing the per-country limitation on foreign tax credits, all of the distribution from Lausanne had its source in Switzerland, the country of Lausanne’s incorporation. On February 14, 1972, the taxpayer filed a refund action in district court. The sole issue is the source for foreign tax credit limitation purposes of those portions of the minimum distribution from Lausanne that Lausanne earned in Argentina, Brazil, and Peru, respectively. The taxpayer asserted below that the distribution should be sourced where the profits comprising it were earned. The government asserted that the distribution should be sourced where the subsidiary that earned the profits was incorporated. The district court determined that because the Secretary of the Treasury had withdrawn and not formally reenacted a regulation supporting the Commissioner’s position, the taxpayer was entitled to prevail. The lower court thus failed to decide whether the sourcing rule proposed by the taxpayer or that proposed by the government more faithfully carried out Congress’s purpose regarding the interplay between subpart F, which governs the tax treatment of the foreign source income of a controlled foreign corporation, and the Code provisions governing the computation of the foreign tax credit. Rather, the trial court appears to have taken the position that the government’s failure to issue a treasury regulation embodying its view left the field open for taxpayer’s proposed sourcing rule, the merits of which it left entirely unexamined. On October 2,1975, the Treasury adopted Treas.Reg. § 1.902-3(d)(l) (1975), which provides that for purposes of the per-country limitation of the foreign tax credit, the dividend received by a domestic shareholder from a first-tier subsidiary corporation shall be deemed to be derived from sources within the country in which the first-tier corporation is incorporated. 5 *976 If Treas.Reg. § 1.902-3(d)(l) does in fact speak to the point at issue, if it is retroactively applicable to the case at bar, and if the regulation is valid, then we would no longer be faced with the question whether as an original matter the taxpayer’s proposed sourcing rule or the government’s rule more faithfully carries out Congress’s purpose with respect to the interplay between subpart F and the foreign tax credit provisions. Insofar as the regulation may be characterized as a legislative rule, it is as binding on a court as a statute. See Kramertown, Inc. v. Commissioner of Internal Revenue, 488 F.2d 728 (5th Cir. 1974); K. Davis, Administrative Law § 5.03 (3d ed. 1972). 6 Before addressing the questions whether Treas.Reg. § 1.902-3(d)(l) is valid and whether it is retroactively applicable to the case at bar, we need to place the regulation in the context of the complex statutory scheme regarding the tax treatment of sub-part F income. 7 After attempting to gain an overview of that scheme and how it relates to the Code’s provisions regarding the foreign tax credit, we shall consider whether the regulation applies to the case at bar. II. Prior to 1962, the foreign source income of a foreign corporation was not subject to United States income tax until distributed as dividends to its United States shareholders. 8 If the domestic shareholder was a corporation, it was eligible for foreign tax credit. The result was that by possessing a foreign subsidiary in a low tax country and deferring the distribution of dividends, a domestic corporation could thus at least defer taxation of foreign source income to the extent its foreign taxes were less than those it would have paid in United States tax. When the subsidiary did distribute its income to the parent, United States tax was imposed only to the extent the United States tax rate was above that applicable in the foreign country. As the Congress observed: In the case of foreign subsidiaries, therefore, this means that foreign income taxes are paid currently to the extent of the applicable foreign income tax, and not until distributions are made will an additional U.S. tax be imposed, to the extent the U.S. rate is above that applicable in the foreign country. This latter tax effect has been referred to as “tax deferral.” Sen.Rep.No.1881, 87th Cong., 2d Sess. 78, [1962] U.S.Code Cong. & Admin.News, p. 3381. In his message to Congress in 1961, President Kennedy questioned the wisdom of such favorable tax treatment: “The undesirability of continuing deferral is underscored where deferral has served as a shel- *977 ter for tax escape through the unjustifiable use of tax havens such as Switzerland.” Id. In order to eliminate the perceived abuses of its foreign tax scheme, Congress enacted the provisions of subpart F. Revenue Act of 1962, § 12, Pub.L. 87-834, 87th Cong., 2d Sess. (October 16, 1962). See generally Beemer, Revenue Act of 1962 and United States Treaty Obligations, 20 Tax L.Rev. 125 (1964). The central provision of the new legislation may be summarized as follows: if the subpart F income of a controlled foreign corporation 9 exceeds certain limits, a United States shareholder 10 of that foreign corporation is required to include in his taxable income a pro rata share of the corporation’s subpart F income, as defined in § 952, whether or not distributed to him. 11 As an ameliorative measure for corporate shareholders of controlled foreign corporations, Congress also enacted § 963. That section provides that subpart F income is not to be taxed to the domestic corporate shareholder if the foreign corporation meets a schedule of minimum distributions. The purpose of this provision is to forego any tax on the domestic shareholders with respect to undistributed income of controlled foreign corporations in cases when the combined foreign and United States tax (to the extent the latter is paid on the distributed income) is not substantially below the United States corporate tax rate. Sen.Rep.No. 1881, 87th Cong., 2d Sess. 88, 1962 U.S. Code Cong. & Admin.News, p. 3391. Hence the lower the foreign tax rate, the greater the required minimum distribution. For example, if the effective foreign tax rate is under 10% a controlled foreign corporation must distribute 90% of its earnings and profits after foreign taxes; if the foreign tax rate is 40% the required minimum distribution is 50%. The lower the foreign tax rate, the higher the proportion of total earnings must be subject to U.S. tax if the aggregate rate is not to be substantially below the U.S. corporate rate. As a condition to using the relief provision offered by § 963, a taxpayer is required to consent to all regulations promulgated under that section that are applicable to the year in question. 12 Treas.Reg. § 1.963-4(c)(1) provides that the foreign tax credit of a United States shareholder with respect to a minimum distribution received for the taxable year shall be determined under the provisions of §§ 901 through 905, subject to certain conditions not relevant here. Hence, questions regarding the treatment for purposes of the foreign tax credit of *978 § 963 minimum distributions require resort to the interstices of §§ 901 through 905 and their accompanying regulations. The foreign tax credit provisions, §§ 901 through 905, were enacted in order to eliminate double taxation by ensuring that income subject to tax in the United States and a foreign country is taxed no more than the higher of either the United States or foreign country rate. See generally American Chicle Co. v. United States, 316 U.S. 450, 451, 62 S.Ct. 1144, 1145, 86 L.Ed. 1591 (1942) (construing predecessor statutes). Section 901 provides that a taxpayer is allowed a credit against his federal income tax for taxes paid or deemed paid to a foreign country. Under § 902(a), a domestic corporation with a foreign subsidiary is deemed to have paid a pro rata portion of any foreign income tax paid by the foreign subsidiary on distributed earnings. The portion of taxes deemed paid by the United States shareholder is an amount that bears the same percentage relationship to the total taxes paid by the foreign corporation as the amount of dividends received by the domestic corporation bears to the total income received by the foreign corporation. The amount of the foreign tax credit is limited, however, by the provisions of § 904. For the taxable years in question, the Code offered taxpayers the option of computing the maximum foreign tax credit by either a “per-country” limitation, I.R.C. § 904(a)(1), or an “overall” limitation, I.R.C. § 904(a)(2). 13 Under the per-country option, the credit for any country may not offset a greater share of the taxpayer’s United States taxes than the earnings sourced in that country comprise of the taxpayer’s total earnings. 14 Thus, the source of foreign earnings becomes crucial in calculating the foreign tax credit. The only question presented here is the source for this purpose of the minimum distribution taxpayer received from Lausanne. Taxpayer would source the income received in the three South American countries in which Lausanne earned that income. This would increase the foreign tax credit allowed taxpayer with respect to those countries and would enable it to take a larger tax credit for taxes paid by its subsidiaries in those countries. The government would source the Lausanne distribution in Switzerland, the country of that subsidiary’s incorporation, thereby increasing taxpayer’s maximum foreign tax credit with respect to Switzerland. III. Sections 901 through 905 of the Code are unhelpful in deciding which of these proposed sourcing rules should be applied. 15 By adopting Treas.Reg. § 1.902-3(d)(l), the Secretary of the Treasury has, we believe, *979 sought to provide an answer to this question. 16 Taxpayer first asserts that because the regulation refers only to “dividends” rather than “minimum distributions” and because a § 963 minimum distribution differs from the traditional dividend, Treas.Reg. § 1.902-3(d)(l) does not apply to the situation presented by the case at bar. Taxpayer’s argument is entirely unpersuasive. 17 It would make little sense to treat a § 963 distribution by a different sourcing rule than that applicable to traditional dividends merely because § 963 establishes a schedule of minimum amounts necessary to achieve the tax benefits promised by that section. The taxpayer has suggested no basis in logic or policy for distinguishing dividends from § 963 distributions, and we have found none. Section 316(a) defines “dividends” as corporate distributions made out of earnings and profits. Section 963 by its term specifies that “minimum distributions” are made out of the earnings and profits of the controlled foreign corporation. Thus apart from the minimum amounts specified in § 963 for minimum distributions, § 963 distributions are identical to dividends; they are in fact a class of dividends. Accordingly, we reject taxpayer’s argument that Treas.Reg. § 1.902-3(d)(l) does not address the question presented by the case at bar. A far more difficult matter, and one that will occupy us at some length, is whether § 1.902-3(d)(l) should be applied retroactively to the taxable year in question. Two issues arise through statute or regulation on the Secretary’s power to prescribe retroactive regulations. Second, we must determine whether, under the peculiar circumstances of this case, the government’s attempt to apply the regulation retroactively constitutes an improper exercise of the Secretary’s power to promulgate retroactive regulations. A. We begin by noting that although Treas.Reg. § 1.902-3(d)(l) does not expressly state whether it is to have retroactive effect, in the absence of limitations imposed by statute or regulation such Treasury Regulations are generally entitled to retroactive application. Section 7805(b) of the Code provides that the Secretary “may prescribe the extent, if any, to which any ruling or regulation, relating to the internal revenue laws, shall be applied without retroactive effect.” In undertaking to discern the intent of Congress as to the retroactivity of such regulations as were authorized under Section 7805(b) we are compelled to note that the authorization to deal with the matter of retroactivity is worded in a manner that indicates clearly that in a generality of cases such rulings or regulations are to be applied with retroactive effect. This results from the fact that the authorization to the Secretary is to prescribe the extent, if any, to which they shall be applied without such effect. We discern from this a clear implication that regulations adopted under the authority of that section will generally have retroactive effect. The Commissioner’s failure to limit regulations to prospective application is, to be sure, reviewable for abuse of discretion. The point here is simply that the absence of language specifically according Treas.Reg. § 1.902-3(d)(l) retroactive effect is no bar to its application to this case. See Dixon v. United States, 381 U.S. 68, 71-75, 85 S.Ct. 1301, 14 L.Ed.2d 223 (1965); Charbonnet v. United States, 455 F.2d 1195, 1200 (5th Cir. 1972); Pollack v. Commissioner of Internal Revenue, 392 F.2d 409 (5th Cir. 1968). Anderson, Clayton claims to have discovered an applicable statutory limit to this general power to issue retroactive regulations. It offers the following attack on Treas.Reg. § 1.902-3(d)(l): § 963(a) re *980 quires a taxpayer to consent to all regulations promulgated under § 963 “prior to the last day prescribed by law for filing its return” for the taxable year; Treas.Reg. § 1.902-3(d)(l) is an indirect amendment to regulations promulgated under § 963; the new regulation was adopted after the last day for taxpayer to file its 1964 return; consequently taxpayer cannot be bound by a regulation to which it did not consent. This argument presumes that taxpayer’s consenting to certain regulations obligated the government to refrain from attempting to apply all other regulations. That assumption is incorrect. Nothing in § 963 or any regulation adopted thereunder suggests that when taxpayer “consents” the government is thereby entering an obligation that limits its future conduct. Rather, the taxpayer’s consent plainly represents only his agreement not to challenge existing regulations. There is clearly no requirement that a taxpayer consent to the regulations in question as a condition of his being bound thereby. With respect to regulations to which he has not consented, the taxpayer stands in a position neither better nor worse than that of any taxpayer confronted with an applicable regulation. He is free to challenge the regulation, but bound unless his challenge succeeds. Indeed, it would be unfair to prevent a taxpayer from challenging regulations he has not seen. But if by his consent to some regulations he does not consent to all that may later be imposed, neither does he automatically shield himself from all subsequently adopted regulations. We conclude, therefore, that no statute or regulation limits the Secretary’s power to adopt retroactive regulations regarding the treatment of a minimum distribution for purposes of computing the per-country limitation on foreign tax credit. B. Taxpayer’s second line of attack on Treas.Reg. § 1.902-3(d)(l) is that retroactively applying that regulation to the instant case would constitute an improper exercise of the Secretary’s power to promulgate retroactive regulations. Taxpayer relies on dicta from Chock Full O’Nuts Corp. v. United States, 453 F.2d 300 (2d Cir. 1971). In that ease, the Commissioner adopted a regulation resolving the legal issue involved in litigation pending between the taxpayer and the government. The taxpayer argued that the regulation should not be given retroactive effect because it represented an effort to change the policy of existing regulations in order to support the government’s position in its pending litigation. The court said: While retroactivity [of] tax regulations is . presumptively permissible, it is in each case for the court to determine whether under all the circumstances retroactive application would be warranted. . A taxpayer, when acting in an area of unsettled law, has no “vested interest in a hypothetical decision in his favor prior to the advent of the regulations.” Helvering v. Reynolds, 313 U.S. 428, 433, 61 S.Ct. 971, 974, 85 L.Ed. 1438 (1941). On the other hand, the Commissioner may not take advantage of his power to promulgate retroactive regulations during the course of a litigation for the purpose of providing himself with a defense based on the presumption of validity accorded to such regulations. 453 F.2d 302-03 (footnotes omitted). The court termed “questionable” the validity of the regulation at issue as an exercise of the Commissioner’s power to promulgate retroactive regulations, but found it unnecessary to resolve that question. No case has held that the Secretary abused his discretion to promulgate retroactive regulations merely because the regulation at issue affected a legal matter pending before a court at the time the regulation was adopted. To be sure, the case at bar is distinctive in that the district court had already issued its decision by the time the Secretary adopted Treas.Reg. § 1.902- *981 3(d)(1). We do not think that fact crucial, however. 18 It is more important to consider how a regulation stands in respect to prior law than to focus on how it relates in time to the litigation in which the government seeks to invoke it. The court in Chock Full O’Nuts, for example, was addressing its quoted remarks to the taxpayer’s argument that the Commissioner was trying to change settled law at the eleventh hour in order to defend against taxpayer’s claim. Viewed in that light, the Second Circuit’s concern is well-taken. Courts have declined to give retroactive effect to regulations or rulings when retro-activity would work a change in settled law relied on by the taxpayer and implicitly approved by Congress, Helvering v. R. J. Reynolds Tobacco Co., 306 U.S. 110, 116, 59 S.Ct. 423, 83 L.Ed. 536 (1939), when it would lead to inequality of treatment between competitor taxpayers, International Business Machines Corp. v. United States, 343 F.2d 914, 170 Ct.Cl. 357 (1965), cert. denied, 382 U.S. 1028, 86 S.Ct. 647, 15 L.Ed.2d 540 (1966), or when, in general, the result of retroactivity in a particular case would be unduly harsh, see Woodward v. United States, 322 F.Supp. 332, 335 (W.D. Va.) (dicta), aff’d, 445 F.2d 1406 (4th Cir. 1971). From these cases we may distill a list of some of the considerations that are relevant to a court in reviewing the Secretary’s exercise of his discretionary power to adopt retroactive regulations. That list includes: (1) whether or to what extent the taxpayer justifiably relied on settled prior law or policy and whether or to what extent the putatively retroactive regulation alters that law; (2) the extent, if any, to which the prior law or policy has been implicitly approved by Congress, as by legislative reenactment of the pertinent Code provisions; (3) whether retroactivity would advance or frustrate the interest in equality of treatment among similarly situated taxpayers; and (4) whether according retroactive effect would produce an inordinately harsh result. 19 Far from altering prior law or policy regarding the sourcing of dividends received by a domestic corporation from its foreign subsidiaries, Treas.Reg. § 1.902-3(d)(1) expresses a position consistently held by the Treasury, though not continuously manifested in the regulations. Insofar as there was a vacuum in applicable regulations after 1964 but prior to the promulgation of Treas.Reg. § 1.902-3(d)(l), the tax *982 payer could not, we think, justifiably rely on a contrary sourcing rule. Prior to the enactment of the 1954 Code, the Commissioner provided in I.T. 4089, 1952-2 Cum.Bull. 142, that unless otherwise provided by law (as, for example, when the income is United States source income under the pertinent provisions of the statute now encompassed in §§ 861 and 862 of the 1954 Code) all dividends paid by a first tier foreign subsidiary would be deemed to have been derived from sources within the country in which the subsidiary was incorporated, and that all taxes qualifying for the foreign tax credit would be treated as having been paid to the country of the subsidiary’s incorporation. In 1957 the Secretary incorporated the rule of I.T. 4089 in Treas.Reg. § 1.902-l(c), which provided that dividends received from a foreign subsidiary “shall be deemed to have been derived from sources within the foreign country ... in which such foreign corporation is incorporated, to the extent that under section 862(a)(2) such dividends are treated as income from sources without the United States . . .” The new regulation also provided that all income taxes paid or deemed paid by the foreign corporation to a foreign country would be deemed paid to the country of its incorporation. 20 Had Treas.Reg. § 1.902-1(c) been continued in effect, it would have governed the sourcing issue in the case at bar. The enactment of the Revenue Act of 1962, Pub.L. 87-834, 76 Stat. 960, added subpart F to the Code and changed the formula for computing the § 902 credit. The latter change required certain revisions in the regulations promulgated under § 902. 21 Through Treas.Reg. § 1.902-5(a) (1965), the government declared that paragraphs (a) through (e) of Treas.Reg. § 1.902-1 were inapplicable (with certain exceptions not relevant here) to distributions received during taxable years beginning after December 31, 1962. Among the provisions withdrawn was Treas.Reg. § 1.902-l(c) (1957). Nothing in the new Act changes or appears to have required a change in the general source rule of dividends from foreign corporations embodied in Treas.Reg. § 902-l(c). Accordingly, the Commissioner proposed to continue this source rule in a new regulation, § 1.902-3(d)(l), Proposed Treasury Regulations on Income Tax (1954 Code), 29 Fed.Reg. 12838 (1964). When the amended Treasury Regulations incorporating the changes made by the Revenue Act of 1962 were adopted, however, it was an *983 nounced that proposed Treas.Reg. § 1.902-3(d)(1) and certain other proposed regulations would be “reissued with a new notice of proposed rulemaking.” T.D. 6805, 1965-1 Cum.Bull. 39 22 No dividend sourcing rule was in fact adopted until 1975. 23 Nevertheless, in the interregnum during which no treasury regulation governed the sourcing issue involved in the case at bar, neither by regulation nor ruling did the Commissioner intimate his intention to follow a foreign dividend sourcing rule other than that which had obtained prior to the Revenue Act of 1962. Moreover, the Commissioner did adopt significant analogues to this dividend sourcing rule in related areas. For example, new Treas.Reg. § 1.902-3(d)(2) (1965) provided that foreign taxes paid by a first tier subsidiary are deemed to have been paid to the country of its incorporation. Treas.Reg. § 1.960 — l(i) (1971), which deals with the computation of foreign tax credit when the foreign subsidiary’s income is taxed to the domestic parent under subpart F of the Code, provides that such income “shall be deemed to be derived from services within the foreign country or possession of the United States under the laws of which such first-tier corporation . is created or organized.” That is, if the subpart F income of Lausanne had been imputed to taxpayer under § 951 of the Code, rather than actually distributed pursuant to § 963, the sourcing rule in Treas.Reg. § 1.960 — l(i) would have controlled. Given the skein of regulation into which the Commissioner has woven the sourcing rule originally expressed in I.T. 4089, it should have come as little surprise to the taxpayer to find the same sourcing rule approved in Treas.Reg. § 1.902-3(d)(l) (1975). 24 In sum, Treas.Reg. § 1.902-3(d)(l) (1975) did not alter a settled prior law or policy. 25 Insofar as there was any uncertainty regarding the proper dividend sourcing rule during the interregnum between 1962 and 1975, we do not think taxpayer could justifiably have relied on a sourcing rule contrary to that which concededly obtained before and after this period. The best that taxpayer can claim is that the law *984 or policy was unsettled during this period. 26 Under the circumstances, that is not enough to preclude our according Treas.Reg. § 1.902-3(d)(l) retroactive application. 27 C. Taxpayer’s final objection to according retroactive effect to Treas.Reg. § 1.902-3(d)(1) is that as an amendment to regulations issued under § 963 the regulation in question must be viewed as issued pursuant to the specific delegation of power contained in § 963(f). This, taxpayer argues, means that Treas.Reg. § 1.902-3(d)(l) is legislative as opposed to merely interpretative in nature. Taxpayer argues that we should not accord retroactive effect to a rule legislative in nature when to do so would produce a harsh or unfair result. That legislative rules may be retroactive is settled law. See, e. g., Manhattan Equipment Co. v. Commissioner of Internal Revenue, 297 U.S. 129, 135, 56 S.Ct. 397, 400, 80 L.Ed. 528, 531-32 (1936); Charbonnet v. United States, 455 F.2d 1195, 1200 (5th Cir. 1972) 28 Taxpayer does not claim that retroactive operation of Treas.Reg. § 1.902-3(d)(l) would violate due process. We have already considered and rejected the argument that retroactive application of the regulation would be inconsistent with the “consent” language of § 963. Nothing in § 902 would preclude retroactive application of a retroactive regulation promulgated thereunder. Since retroactive application of the regulation is consistent with the Secretary’s statutory authority to prescribe regulations, and since it effects no change in settled law, even if Treas.Reg. § 1.902-3(d)(1) is deemed legislative in character there is no bar to its application to the instant case. Neither the courts nor Congress have drawn any distinctions between interpretative and legislative Treasury Regulations as regards their retroactivity. See Rogovin, The Four R’s: Regulations, Rulings, Reliance, and Retroactivity — A View From Within, [1976] Stand.Fed.Tax Rep. (CCH) If 5980A.0153, 15980A.016. 29 Treas.Reg. *985 § 1.902-3(d)(l) is the first regulation to provide a sourcing rule for § 902 as revised by the Revenue Act of 1962. It is, as we have seen, consistent with the only prior dividend sourcing rule that obtained prior to the 1962 Act. Under these circumstances at least, that a regulation is legislative in character does not affect the question whether it should be applied retroactively. 30 IV. Having demonstrated that Treas.Reg. § 1.902-3(d)(l) may be applied retroactively to the taxable year in question, we must consider next whether that regulation is valid. The taxpayer does not question the procedure by which the regulation was adopted, but only its substantive validity. Accordingly, taxpayer bears a heavy burden. The regulation “must be sustained unless unreasonable and plainly inconsistent with the revenue statutes.” Bingler v. Johnson, 394 U.S. 741, 749-51, 89 S.Ct. 1439, 1444, 22 L.Ed.2d 695 (1969), quoting Commissioner of Internal Revenue v. South Texas Lumber Co., 333 U.S. 496, 501, 68 *986 S.Ct. 695, 698, 92 L.Ed. 831 (1948); Fitzgerald Motor Co. v. Commissioner of Internal Revenue, 508 F.2d 1096 (5th Cir. 1975). Anderson, Clayton has not met this rigorous standard. Taxpayer argues that the country of incorporation rule contained in Treas.Reg. § 1.902-3(d)(l) is inconsistent with the basic concept of the foreign base company component of subpart F income. See note 1, supra. By definition, foreign base company income must be earned in a country other than that in which the controlled foreign corporation is incorporated. Treas.Reg. § 1.902-3(d)(l) required taxpayer to source a minimum distribution of Lausanne’s sub-part F income to the country of the subsidiary’s incorporation for purposes of determining the limitation on taxpayer’s foreign tax credit. Anderson, Clayton contends that it is inconsistent to source to the country of Lausanne’s incorporation a distribution of income that by definition must be earned outside that country. We find no such inconsistency between Treas.Reg. § 1.902-3(d)(l) and the provisions of subpart F. The purpose of those provisions was to eliminate tax havens “such as the Swiss corporation which often serves as merely an address for sales made in every country in Europe except Switzerland,” to prevent a corporation from “flout[ing] our tax laws by simply setting up an address company, say in Panama, to sell goods in Europe which did not originate in Panama . . . and which had nothing to do with Panama.” 108 Cong.Rec. 17750, 17752 (1962) (remarks of Sen. Kerr, floor manager of H.R. 10650). That is why subpart F is concerned with, for example, income earned by a Swiss subsidiary outside Switzerland. With respect to any income earned within Switzerland, or more generally, the country in which the subsidiary is incorporated, the foreign corporation is not acting as the sort of tax haven or “address company” against which Congress aimed subpart F. Congress’s reasoning on this point can hardly be said, however, to carry decisive implications for the separate question regarding how a distribution of sub-part F income is to be sourced for purposes of the foreign tax credit. Insofar as there are such implications they are entirely consistent with Treas.Reg. § 1.902-3(d)(l). We start with the rule embodied in Treas. Reg. § 1.902-3(d)(2) (1965) that deems taxes paid by a foreign subsidiary to have been paid to the country of the subsidiary’s incorporation. This regulation carried forward the rule of I.T. 4089 (1952-2 Cum. Bull. 142) and Treas.Reg. § 1.902-l(c) (1957). In 1960, Congress expressed its approval of the rule. See note 20 supra. As we have seen, nothing in the Revenue Act of 1962 was related to this rule (even assuming for the sake of argument that the 1962 Act might have left open the proper dividend sourcing rule), and the Commissioner promptly adopted it once again. Given legislative reenactment of § 902 and the contemporaneous construction afforded that section in Treas.Reg. § 1.902-3(d)(2), the rule deeming taxes paid by a foreign subsidiary to have been paid to the country of its incorporation has force of law. Taxpayer does not contest the validity of this regulation. Congress included in subpart F § 963, which was intended to offer the corporate taxpayer some relief from the effects of § 951. But, as we have seen, the amount of the required minimum distribution varies inversely with the foreign tax rate. This was thought necessary in order to ensure that the combined United States and foreign tax rates on the subsidiary’s income would not be substantially below the United States corporate rate. S.Rep.No.1881, 87th Cong., 2d Sess. 88, 1962 U.S.Code Cong. & Admin.News, p. 3391. In short, the country to which the foreign subsidiary pays taxes is integral to the concept of a § 963 minimum distribution. 31 *987 The taxpayer’s sourcing rule, which it claims is the only rule consistent with sub-part F, would allow the taxpayer to minimize the tax equalizing effects of the minimum distribution by using the distribution to claim increased tax credits for the relatively higher taxes that taxpayer’s South American subsidiaries paid to Argentina, Brazil and Peru and, thus, to reduce its overall tax burden. This would be true even though the earnings underlying the distribution (the taxes on which Lausanne paid only to Switzerland) had absolutely nothing to do with the actual taxes paid to Argentina, Brazil and Peru by the South American subsidiaries. In other words, application of taxpayer’s sourcing rule would sever the foreign country to which a foreign subsidiary paid creditable taxes on its earnings from any necessary connection to the foreign country to which the domestic parent attributes a distribution of those earnings for purposes of the foreign tax credit. We need not consider whether, in light of the legislative history of subpart F, the Code would permit such a curious result. It suffices for our purposes that the Code does not require that result. Nothing in subpart F forbids the Secretary to correlate the foreign country to which a foreign subsidiary is deemed to pay taxes on its earnings and profits with the source of a minimum distribution of those earnings and profits for purposes of the foreign tax credit. 32 We conclude that Treas.Reg. § 1.902-3(d)(1) is not “plainly inconsistent” with the Code and that, accordingly, the taxpayer’s challenge to the validity of the regulation cannot be sustained. Having found the regulation both valid and retroactively applicable to the case at bar, we must reverse the district court’s judgment on this issue. THE DEDUCTIBILITY OF AN UNREALIZED FOREIGN EXCHANGE LOSS The taxpayer has taken a cross-appeal from the district court’s determination that Anderson, Clayton is not entitled to a loss deduction with respect to the decline in foreign exchange value of promissory notes it received as dividends from its Argentine subsidiaries during fiscal year 1964. We uphold the district court’s conclusion. Consequently we need not reach the issue of the source of such a loss for purposes of the § 904(a)(1) limitation on foreign tax credit. *988 V. In fiscal 1964, the Argentine government issued an order “blocking” the country’s currency. That is, Argentine pesos could be neither expatriated from Argentina nor converted within that country into freely convertible currency of any other country. Subsequently, on December 20, 1963 and January 9,1964, Anderson, Clayton received dividend distributions from its Argentine subsidiaries in the form of negotiable promissory notes payable to taxpayer in Argentine pesos. The taxpayer determined the value of the notes by converting their principal amount into United States dollars at the rate of exchange prevailing on the date of distribution. It included that amount, $1,150,-320.87, in its gross income as a dividend. By the end of fiscal 1964, the value of Argentine pesos had declined relative to dollars. Taxpayer determined the United States dollar value of the promissory notes as of the end of fiscal 1964 and deducted $278,892.29 as an ordinary business loss. Taxpayer reported this loss as derived from sources within the United States. It premised this sourcing of the loss on the fact that it held the promissory notes within the United States at all times during fiscal 1964 subsequent to their receipt. At no time in fiscal 1964 did Anderson, Clayton actually exchange the notes for dollars. Upon an audit of taxpayer’s return for the taxable year in question, the Internal Revenue Service challenged the taxpayer’s sourcing that loss in the United States. At trial the government also argued that taxpayer realized no loss from the decline value of its notes. The district court agreed with the latter proposition and thus had no occasion to reach the sourcing issue. Moreover, the district court determined that the government was neither collaterally estopped to deny taxpayer a loss deduction by virtue of a prior Court of Claims decision, Anderson, Clayton & Co. v. United States, 168 F.Supp. 542, 144 Ct.Cl. 106 (1958), nor foreclosed from objecting to the deduction by an agreement executed in 1942 by taxpayer and the Commissioner relating to taxpayer’s foreign branch accounting. 33 VI. It should be clear at the outset that allowing Anderson, Clayton an ordinary‘business loss deduction for the decline in foreign exchange value of its promissory notes would violate a fundamental tenet of our income tax system!. That is the proposition that property must be sold, abandoned, injured by physical causes or demonstrated to be worthless before a taxpayer is entitled to a deductible loss. See, e. g., United States v. S. S. White Dental Mfg. Co., 274 U.S. 398, 401, 47 S.Ct. 598, 71 L.Ed. 1120 (1927). It is equally clear that mere diminution in the value of foreign money is not enough to permit a deduction for a loss sustained. Losses may be recognized only when the taxpayer converts foreign currency into dollars. S. J. Mertens, Law of Federal Income Taxation § 28.82 (rev. ed. 1975). The taxpayer does not deny that allowing it a deduction for the mere decline in value of the promissory notes would violate this basic rule of taxation. Anderson, Clayton nevertheless boldly insists that we are compelled to reach that result by a Court of ^=Claims~decision handed (down áliñost twenty ago, Anderson, Clayton & Co. v. United States, 168 F.Supp. 542, 144 Ct.Cl. 106 (1958), and an agreement struck thirty-five years ago between taxpayer and the Commissioner regarding taxpayer’s foreign branch accounting and, allegedly, taxpayer’s foreign subsidiary accounting. *989 A. For fiscal years 1933 and 1934 the taxpayer used a method of accounting whereby it determined the profits of its foreign branches by adjusting the current accounts on their books to dollar values at the close of each fiscal year. Anderson, Clayton carried this amount in terms of dollars to the home office account and reported it. as income for United States tax purposes. The IRS challenged taxpayer’s method of accounting for 1933 and 1934. Taxpayer filed suit in tax court. Prior to that court’s adjudication on the merits, the parties reached a settlement agreement. That agreement, reached in 1942, provided that taxpayer was to compute its income by applying consistently the accounting practice it had adopted: The income of the Havre Branch and other autonomous foreign offices keeping their accounts in a foreign currency is to be determined by the difference in dollar net worth at the beginning and end of the year adjusted for any profits transferred from such branch during the year. The first skirmish between the taxpayer and the government concerns the scope of this agreement. The taxpayer asserts that the agreement required it to recognize gains or losses from exchange fluctuations on payables to and receivables from its foreign branches and its foreign subsidiaries. The government argues that the agreement was limited to taxpayer’s foreign branch accounting practices. Taking merely the language of the agreement, the government’s view seems correct. Taxpayer argues that the “other autonomous foreign offices” language encompasses foreign subsidiaries. Since the phrase is preceded by the word “other”, which assumes that the Havre branch was an “autonomous” office, we do not think the phrase carries the freight taxpayer would load upon it. The phrase is entirely consistent with the notion that “branches” are autonomous and that the agreement included not only Le Havre but also other branches. Moreover, the phrase, “Havre and other autonomous foreign offices” is followed after an interval by the words, “such branch,” which obviously refer back to the former quoted phrase. It seems likely, therefore, that the agreement envisaged taxpayer’s applying its accounting method-only- to foreign ^branches. This inference is strongly reinforced upon considering the overall nature of the agreement. The agreement represents the resolution of a dispute concerning taxpayer’s income. It begins by addressing the question of properly calculating the income of the Havre Branch and other offices that kept their accounts in foreign currency. Taxpayer would be concerned with computing the income of a foreign entity only to the extent that its income was reflected in taxpayer’s income for purposes of United States tax. At' the time this agreement was written — before the advent of subpart F — only the income of its unincorporated branches was income taxable to Anderson, Clayton. Adjustments in the income of its branches constituted adjustments in taxpayer’s income. This was not true of subsidiaries. The income of a subsidiary mattered directly only insofar as taxpayer received from it a dividend. Hence, the most sensible construction of the 1942 agreement is that it addressed the question how taxpayer was to reflect the foreign currency income of foreign branches in its own income. Nothing in the agreement addresses the treatment for tax purposes of dividends received by the parent from a foreign subsidiary. That the 1942 agreement concerned only foreign branch accounting is enough to warrant our resolving this part of the dispute in the government’s favor. We note additionally, however, that Anderson, Clayton conceded at oral argument that even if the 1942 agreement did address the problem of dividend income, it could bind the government only with respect to the taxable years covered by that agreement. *990 B. Taxpayer presents a more serious claim with respect to the collateral estoppel effect of Anderson, Clayton & Co. v. United States, 168 F.Supp. 542, 144 Ct.Cl. 106 (1958). That case followed the government’s challenge to a claimed exchange loss determined under the same accounting practice countenanced by the 1942 settlement agreement. The court interpreted ' the settlement agreement in language that, taxpayer claims, compels us to accept taxpayer’s construction. Moreover, taxpayer asserts, the judgment in Anderson, Clayton, supra, awarded taxpayer a deduction for the diminution in foreign exchange value of a receivable from a foreign subsidiary. Taxpayer thus attempts to bootstrap its way tó a reversal. The argument is as follows: Although the settlement agreement alone could not bind the Commissioner for future tax years, and although it may be arguable whether the 1942 agreement even covered foreign subsidiaries, nonetheless the Court of Claims decision which interpreted the agreement to be binding and to cover foreign subsidiaries must control this case. We think both the 1942 agreement and the Court of Claims’ holding far more limited than taxpayer’s reading would suggest. In 1930 Anderson, Clayton established a branch office in Alexandria, Egypt. In 1939, shortly after the outbreak of the Second World_ War, the Egyptian government blocked the conversion of Egyptian pounds into United States dollars. Taxpayer could thus no longer remit tiie profits of its Egyptian branch to its home office. The branch office’s bopks included the account of Nile Ginning Cdmpany, a foreign subsidiary of Anderson, Clayton. The current Egyptian pound account of Nile Ginning on the books of thé branch and also, apparently, of the parent company was treated in the same manner as all other foreign currency balances. That is, it was valued in dollars at the current rate of exchange and at the end of each fiscal year the difference between that value and the value previously recorded on the books was carried into an “optional account” as a gain or loss in exchange. Taxpayer adjusted its accounts twice a year and reflected in its tax returns at the end of each fiscal year the resulting “profit” and “loss.” As of July 1945-"'taxpayer had net unremitted earnings of 128,119.71 Egyptian pounds that had been reported as United States income at the rate of $4.13 per pound. From 1945 until 1949, the taxpayer continued to take into its United States income for tax purposes the United States dollar value of its branch office profits. In 1949 taxpayer liquidated its branch office and the Nile Ginning Company took over the branch’s assets and liabilities. In July 1949 taxpayer had on its books a current account receivable of 250,665.745 Egyptian pounds, representing blocked funds at the rate of $4.13 per pound. In September 1949 Britain devalued the pound sterling. The Egyptian pound plummeted. By July 1950, the Egyptian pound was worth $2.50. The resulting decrease in taxpayer’s blocked Egyptian earnings amounted to $413,463.19, for which taxpayer claimed a loss deduction on its fiscal 1950 income tax return. For tax years 1946-1949, taxpayer had elected to file deferred income tax returns under Mimeograph 6475, 1950-1 Cum.Bull. 50, which specified particular tax consequences when foreign currency income was “blocked” by the foreign country. Under this Mimeograph, the taxpayer would not report such income until the foreign country lifted its currency restrictions. The Mimeograph provided specifically that a taxpayer that elected to defer reporting its blocked income waived the right to claim that the deferrable income was includible in its gross income for any year other than the year restrictions were lifted. Taxpayer claimed that notwithstanding its election to file deferred income tax returns for the years 1946-1949, the 1942 settlement agreement allowed it to ignore the provisions of the Mimeograph and include the 1946-1949 exchange fluctuations of its Egyptian account in determining its *991 1950 taxable income. That was the only question the Court of Claims decided in 1958. The court ruled that taxpayer could not take a deduction for foreign exchange losses for the taxable years 1946-1949. It reasoned that because the claimed 1946-1949 loss was based upon amounts unreported as income, taxpayer could not take a deduction. That is the holding of the case. In the case at bar, Anderson, Clayton relies not upon this holding but upon an additional part of the court’s judgment. In its 1950 tax return, taxpayer had also claimed loss deductions for exchange fluctuations in its pre-1946 blocked Egyptian income. The parties did not litigate the propriety of these deductions. The government conceded the point. The court wrote: After filing of suit in this court the Government has conceded that with respect to 47,487.426 Egyptian, pounds in the Egyptian account plaintiff is entitled to a loss deduction of $77,404.50 in 1950. The 47,487.426 pounds represent earnings for years prior to 1946 which were included in United States income at the exchange rate of $4.13 per pound and upon which United States taxes have been paid. Anderson, Clayton & Co. v. United States, supra, 168 F.Supp. at 545. The court stated expressly that this portion of the Egyptian account represented the Egyptian account of the Alexandria branch. Id. at 545 n. 2. No mention is made of the earnings of the subsidiary, Nile Ginning. Consequently, the district court’s characterization of Anderson, Clayton, supra, as holding “that plaintiff was entitled to a loss deduction for exchange losses determined under its accounting practice with respect to all payables from foreign branches, subsidiaries, or unrelated entities for which the plaintiff had a tax basis” is unnecessarily overbroad. First, the judgment of the Court of Claims allowed Anderson, Clayton a loss deduction with respect to unremitted earnings of a branch, not a subsidiary. Indeed, in light of that court’s finding of fact that the 1942 agreement was one “concerning . . . foreign branch accounting,” it would have been surprising had that court sanctioned the extension of the agreement to cover foreign subsidiaries. See Record at 23. 34 Second, even that portion of the judgment was based not on a fully litigated matter, but a concession of the government. As we have seen, the Court of Claims held only that taxpayer was not entitled to a loss deduction arising from exchange fluctuations for the years 1946-1949. Assuming for the sake of argument that a judgment based on the government’s concession could have collateral estoppel effect, that judgment is inapposite to the case at bar. The Havre branch of the 1942 settlement agreement and the Alexandria branch of the Court of Claims decision were not separate corporate entities, and the income earned by taxpayer through those offices was its own. Although the Alexandria branch was no longer in existence during taxpayer’s 1950 fiscal year, the government could reasonably have conceded that the intent of the agreement would serve to allow taxpayer to deduct the decline in dollar value of the Nile Ginning current account to the extent that it represented the Alexandria branch’s unremitted earnings upon which taxpayer had paid United States income tax. The government could have conceded this proposition without also conceding that a dividend received from a subsidiary should be accorded the same tax treatment. Insofar as there remains any ambiguity, since both the agreement and the result in the Court of Claims represent exceptions to the general rule that a mere *992 decline in value does not constitute a deductible loss, we construe them narrowly. Even assuming that the government’s concession in Anderson, Clayton, supra, were on point, we would not estop the Commissioner from seeking a result taxpayer concedes would otherwise be the correct treatment of the unrealized foreign exchange losses. Collateral estoppel applies only to an issue that was actually litigated and determined in a prior action, not to an issue that might have been litigated. This was not a case, moreover, in which the issue with respect to which taxpayer seeks to estop the government is necessarily implied by the prior judicial decision even though not expressly decided. Once before the Court of Claims, the government conceded^ the deductibility of some foreign exchange losses but denied it as to others. The court ordered judgment for taxpayer to the extent of the government’s concession and-judgment for the government on the only point it contested. We start from the proposition that strong policy considerations favor confining narrowly the scope of collateral estoppel in tax cases. See Griswold, Res Judicata in Federal Tax Cases, 46 Yale L.J. 1320 (1937). The most persuasive policy consideration in the present context is that perpetuation of an erroneous tax decision over a number of years would prejudice the losing party and violate the policy of tax uniformity among taxpayers. 18 Moore’s Federal Practice § 0.422(1) (2d ed. 1974). It is difficult to think of a case in which according collateral estoppel effect to a prior decision would more blatantly offend the policy of tax uniformity. Whatever the precise nature' of Anderson, Clayton, supra, it was decided under unique circumstances. Extending what taxpayer asserts to have been the principle of that case would exempt Anderson, Clayton from a requirement regarding the realization of foreign exchange losses that all other taxpayers must bear. The chances that a court may reach an unsound result that binds the taxpayer and the government and offends the policy of tax uniformity are greatly increased when the parties do not fully litigate the issue. 35 When one party to a tax case concedes or stipulates the issue upon which the court bases its judgment, the issue is not conclusively determined for purposes of collateral estoppel unless it is clear that the parties so intended. See 18 Moore’s Federal Practice § 0.444(4). In United States v. International Building Co., 345 U.S. 502, 73 S.Ct. 807, 97 L.Ed. 1182 (1953), the Court held that tax judgments based on consent agreements between taxpayers and the government do not collaterally estop litigation on the same issue for lgter tax years. • In that case decisions entered by the Tax Court were based on stipulations that taxpayer owed no federal tax deficiency. The Court said that unless the prior judgment was “an adjudication on the merits, the doctrine of estoppel by judgment would serve an unjust cause: it would become a device by which a decision not shown to be on the ’ merits would forever foreclose inquiry into the merits.” Id. at 506, 73 S.Ct. at 809. This principle has b§en uniformly applied. See United States v. California Portland Cement Co., 413 F.2d 161, 163 (9th Cir. 1969) (collecting cases); cf. Seaboard Air Line R. Co. v. George F. McCourt Trucking, Inc., 277 F.2d 593, 597 (5th Cir. 1960)' Although there may be cases in which á judgment entered with the consent of the parties involves a determination of questions of fact and law by the court, the party seeking collateral estoppel effect has the burden of proving this to be so. See United States v. International Building Co., supra, 345 U.S. at 506, 73 S.Ct. at 809. A fair reading of Anderson, Clayton, supra, demonstrates *993 that the court did not determine the question of law for which taxpayer claims it stands. Finally, the taxpayer argues that we should not apply the principle that a judgment based on stipulated or conceded issues lacks collateral estoppel effect unless the parties expressly indicate their intention not to be bound in the future by the judgment. Anderson, Clayton points out that in United States v. California Portland Cement Co., supra, the government had included such a statement in its stipulations. Similarly in United States v. International Building Co., supra, the government had made such a statement in withdrawing from the bankruptcy proceeding that preceded the Tax Court judgment, though no such disclaimer attached to its stipulations to the Tax Court. .We think the absence of •> such a disclaimer does not suffice to show that the parties intended the judgment to have collateral estoppel effect, a burden properly borne by the party seeking such effect in litigation concerning different taxable years. The presumption is that an issue resolved by stipulation or concession in one suit is not conclusively established in a subsequent suit on a different cause of action unless it is clear that the parties so intended. See 18 Moore’s Federal Practice § 0.444(4). VII. The judgment of the district court with respect to the sourcing of taxpayer’s Lausanne dividend for purposes of calculating the per-country limitation on Anderson, Clayton’s foreign tax credit is reversed and the cause remanded for entry of an order consistent with this opinion. The judgment of the district court regarding taxpayer’s claim for a loss deduction arising from the decline in foreign exchange value of promissory notes received as dividends from its Argentine subsidiaries is affirmed. AFFIRMED IN PART, REVERSED IN PART, AND REMANDED. Footnotes [1] Subpart F, part III, subchapter N, chapter 1 of the Code, §§ 951-964 concerns the income of a foreign corporation controlled by a domestic corporation. For purposes of this case, we may assume that subpart F income is foreign base company sales income, defined by § 954(d)(1) in pertinent part as income derived in connection with the purchase of personal property from a related person and its sale to any person . . . where — (A) the property which is purchased ... is manufactured, produced, grown, or extracted outside the country under the laws of which the controlled foreign corporation is created or organized, and (B) the property is sold for use, consumption, or disposition outside such foreign country . . As will be explained, infra, taxpayer found it tax beneficial to receive as a “minimum distribution” or dividend part of a foreign subsidiary’s subpart F income during fiscal 1964. [2] See note 5, infra. [3] Anderson, Clayton & Co. v. United States, 168 F.Supp. 542, 144 Ct.Cl. 106 (1958). [4] Taxpayer keeps its books and records and files its income tax returns on the accrual method of accounting as authorized by I.R.C. § 446(c)(2). For the year in question it used an accounting period ending July 31. [5] Treas.Reg. § 1.902-3(d)(l): For purposes of section 904(a)(1) (relating to the per-country limitation), in the case of a dividend received by a domestic shareholder from a first-tier corporation there shall be deemed to be derived from sources within the foreign country or possession of the United States under which the first-tier corporation is created or organized the sum of the amounts which under paragraph (a)(3)(ii) of § 1.861-3 are treated, with respect to such dividend, as income from sources without the United States. It is undisputed that taxpayer is a domestic shareholder, meaning for purposes of § 902 a domestic corporation owning at least 10 percent of the voting stock of a foreign corporation, see Treas.Reg. § 1.902-3(a)(l) (1975), or that Lausanne qualifies as a first-tier corporation. The sole point of contention regarding whether Treas.Reg. § 1.902-3(a)(l) addresses the problem at hand is whether a “minimum *976 distribution” for purposes of § 963 of the Code is a “dividend” for purposes of Treas.Reg. § 1.902 — 3(d)(1). This matter is discussed, infra. [6] Regulations issued pursuant to a specific statutory authorization are clearly legislative as opposed to interpretative rules. If consistent with the statutory authorization adopted pursuant to proper procedure, and reasonable, they have force of law. See Fitzgerald Motor Co., Inc. v. Commissioner of Internal Revenue, 508 F.2d 1096 (5th Cir. 1975); Posey v. United States, 449 F.2d 228 (5th Cir. 1971). Section 963(f) of the Code provides that the Secretary shall provide such regulations as he deems necessary regarding the receipt of minimum distributions by domestic corporations. As the taxpayer asserts, Treas.Reg. § 1.902-3(d)(l) amends regulations issued under § 963 of the Code and is therefore at least in part legislative in nature. See note 17, infra. [7] The recent repeal of the minimum distribution option of § 963 of the Code means that the issue of the geographic source of subpart F income will be unimportant to foreign corporations for taxable years beginning after December 31, 1975. Tax Reduction Act of 1945, Pub. L.No. 94-12, § 602(a)(1) (March 29, 1975). Similarly, the issue of the geographic source of income from sources without the United States for purposes of the per-country limitation upon foreign tax credit has become substantially moot for taxable years beginning after December 31, 1975, through repeal of § 904(a)(1) of the Code. Tax Reform Act of 1976, Pub.L.No. 94-455, §§ 1031(a) and (c) (October 4, 1976). [8] The rule given in text did not apply to income of a foreign personal holding company not engaged in a trade or business within the United States. [9] For purposes of subpart F, a controlled foreign corporation means a foreign corporation of which on any day within the taxable year United States shareholders (see note 9, infra) own more than 50% of the total combined voting power of all classes of stock entitled to vote. I.R.C. § 957(a). [10] Section 951(b) of the Code defines “United States shareholder” to mean, with respect to any foreign corporation, a United States person who owns actually or constructively 10% or more of the total combined voting power of all classes of stock of such corporation entitled to vote. [11] The general class of subpart F income that concerns us in the case at bar is foreign base company income, as defined in § 954(a). More particularly, we are concerned with foreign base company sales income as defined in § 954(d)(1). See note 1, supra. [12] Section 963 provides in part: (a) In the case of a United States shareholder which is a domestic corporation and which consents to all the regulations prescribed by the Secretary or his delegate under this section prior to the last day prescribed by law for filing its return of the tax imposed by this chapter for the taxable year, no amount shall be included in gross income under section 951(a)(l)(i) for the taxable year with respect to the subpart F income of a controlled foreign corporation if— (1) in the case of a controlled foreign corporation described in subsection (c)(1) the United States shareholder receives a minimum distribution of the earnings and profits for the taxable year of such controlled foreign corporation; (f) The Secretary or his delegate shall prescribe such regulations as he may deem necessary to carry out the provisions of this section, including regulations for the determination of the amount of foreign tax credit in the case of distributions with respect to the earnings and profits of two or more foreign corporations. [13] For taxable years after December 31, 1975, taxpayers are required to determine their foreign tax credit limitation on an overall basis, subject to certain carryover provisions for taxpayers previously on the per-country limitation basis. See note 7, supra. [14] This may also be expressed as follows: Maximum Credit for tax year for taxes paid or deemed paid to Country C Taxable Income Received from Country C_ World-Wide Taxable Income X United States Income Tax on World-Wide Income (before foreign tax credit) [15] This is by no means to suggest that even without the recent promulgation of Treas.Reg. § 1.902-3(d)(l) (1975), the taxpayer’s position would prevail. Indeed, as we shall see, in light of the consistency of the government’s rule on sourcing over the years, the constancy of regulations that deem taxes paid by a foreign subsidiary to have been paid in the country of its incorporation, and the facility with which the government’s sourcing rule meshes with the legitimate role of the foreign tax credit, the government’s sourcing rule appears to us far more consistent with the intended interplay between subpart F and the foreign tax credit provisions. Had the district court considered the relative merit of the two rules, we think it would properly have reached the same conclusion. [16] For text of Treas.Reg. § 1.902-3(d)(l), see note 5, supra. [17] There is an inconsistency between taxpayer’s arguing that Treas.Reg. § 1.902-3(d)(l) is a legislative rule because it amends regulations issued under § 963, which deals only with minimum distributions, and its arguing that Treas. Reg. § 1.902 — 3(d)(1) does not address the question of treating minimum distributions for tax credit limitation purposes. See note 6, supra. [18] Taxpayer calls our attention to the following language in Commissioner v. Goodwyn Crockery Co., 315 F.2d 110, 113 (6th Cir. 1963): Since the regulations were not in effect at the time the tax liability accrued, or at the time of the [Tax Court] hearing, they have no binding force here. They are authority, however, as a departmental ruling. This statement of law is incorrect. If it were true that regulations had no binding force merely because they were not in effect when the tax liability at issue accrued, the Secretary would be divested of his discretion to issue retroactive regulations. Not only does the Sixth Circuit decision in Goodwyn Crockery fail to consider § 7805(b) of the Code, it never once discusses the concept of retroactive regulations. We can only conclude that no one suggested to the Goodwyn Crockery court that the regulations in question might have had retroactive effect. Other courts have reached the same conclusion and have accordingly disregarded the quoted language. For example, in Exel Corp. v. United States, 451 F.2d 80 (8th Cir. 1971), the court held'that the district court had erred by relying on the rule set forth in Goodwyn Crockery to deny retroactive effect to a treasury regulation. The Eighth Circuit noted that this erroneous view of the law was contrary to § 7805(b) and Dixon v. United States, 381 U.S. 68, 74, 85 S.Ct. 1301, 14 L.Ed.2d 223 (1965). Because the Goodwyn Crockery court was so obviously operating under a mistaken view of the law regarding retroactivity, we can give no credit to its statement that regulations not adopted by the time of the lower court’s hearing have no binding force. [19] The relationship in time between the promulgation of a regulation and litigation in which the regulation would be controlling if applied may be a component of this last factor. We reject the taxpayer’s argument that the chronology should be dispositive, however. In addition, we note that the list of relevant considerations is not intended to be exhaustive. For example, in a particular case it might become relevant to consider the extent to which the new regulation responds to a policy inapplicable to earlier years. [20] When in 1960 Congress added the overall limitation on the foreign tax credit to the existing per country limitation, Act of September 14, 1960, § 1(a), Pub.L. 86-780, 74 Stat. 1010, the Senate Report expressed approval of the source rule regarding foreign taxes set forth in Treas.Reg. § 1.902-l(c). S.Rep.No.1393, 86th Cong., 2d Sess., 4-5 reprinted in [I960] U.S. Code Cong. & Admin.News, pp. 3773-74. [21] Section 9 of the Act changed the formula for computing the § 902 credit with respect to dividends received from other than “less developed country” corporations. The change was a response to a perception that an unjustified tax advantage resulted when a domestic corporation received income in the form of a dividend from a foreign corporation that paid tax at less than the United States rate. The inequity resulted when the amount paid in foreign taxes was not only allowed as a credit in computing the United States tax of the corporation receiving the dividend, but was also “in effect allowed as a deduction (since the dividends can only be paid out of income remaining after payment of the foreign tax).” S.Rep.No.1881, 87th Cong., 2d Sess., reprinted in [1962] U.S. Code Cong. & Admin.News, p. 3368. That is, as a result of including only dividend income in the tax base of the domestic corporation (instead of dividend income plus foreign taxes paid) and at the same time allowing a foreign tax credit, the total of foreign and United States taxes paid on earned income was less than would be paid by a domestic corporation operating in this country (to the extent the United States tax rate exceeded the foreign rate). Accordingly, the Revenue Act of 1962 added a “grossing-up” provision, now codified at § 902(d)(1) and § 78 of the Code, by which a domestic corporation electing to take a foreign tax credit must include in its gross income an amount equal to the taxes of its subsidiary that it is deemed to have paid for purposes of the foreign tax credit provision with respect to the dividend income received. Id. 3371-72, 3524. This matter is entirely independent of the question in which foreign country foreign source income should be sourced. [22] The regulations incorporating the changes made by the 1962 Act were designated as Treas.Reg. §§ 1.902-3 and 1.902-4 (1965). [23] The government asserts that the failure to approve Treas.Reg. § 1.902-3(d)(l) before 1975 arose from the problem of determining the amount of actual dividends and § 78 dividends (under § 78, a domestic corporation that elects to take foreign tax credits must treat as a dividend an amount equal to the foreign taxes that the domestic corporation is deemed to have paid under § 902(a) or § 960(a)(1)) to be treated as foreign source income for purposes of the tax credit when a portion of the actual dividends must be treated as United States source income under § 861. This problem does not arise in the case at bar because the entire dividend taxpayer received from Lausanne concededly constituted foreign source income. [24] Indeed, had the Secretary never promulgated Treas.Reg. § 1.902-3(d)(l) (1975) we should find the consistency with which the “country of incorporation” sourcing rule has been applied in related areas a powerful argument for applying it to the problem at hand. In short, the government’s sourcing rule seems to us more consistent with administrative policy than the taxpayer’s proposed dividend sourcing rule. [25] That a regulation or ruling does alter settled prior law or policy upon which a taxpayer has relied does not, of course, necessarily preclude its retroactive effect. For example, in Pollack v. Commissioner of Internal Revenue, 392 F.2d 409 (5th Cir. 1968), this court held that even assuming a regulation was inconsistent with a policy announced in a prior Technical Information Release, the regulation could still be given retroactive effect at least if the prior interpretation of law was erroneous: The Commissioner may indeed retroactively correct any prior erroneous interpretation of the law, even though a taxpayer may have relied to his detriment on the Commissioner’s mistake. 392 F.2d at 411. See also Dixon v. United States, 381 U.S. 68, 79-80, 85 S.Ct. 1301, 1308, 14 L.Ed.2d 223, 231 (1965) (“Insofar as petitioners’ arguments question the policy of empowering the Commissioner to correct mistakes of law retroactively when a taxpayer acts to his detriment in reliance upon the Commissioner’s acquiescence in an erroneous Tax Court decision, their arguments are more appropriately addressed to Congress.”) (footnotes omitted). [26] Taxpayer asserts that in the absence of existing regulations under § 902, it was required to look to Treas.Reg. § 1.901 — 2(d) (1957), which referred it to § 861 and succeeding sections of the Code for determining the sources of income received. Sections 861(a)(2) and 862(a)(2) pertain to the question whether income received was United States source income or foreign source income. Taxpayer argues that Treas. Reg. §§ 1.861-1 to 1.863-5, which govern the extent to which income is deemed from sources within or without the United States, are predicated not on a country of incorporation test but instead look to the place where the income was earned. Taxpayer concedes, however, that §§ 861 through 864 of the Code are devoid of specific provisions regarding the determination from which foreign country income has been derived once that income is determined to be foreign source income. Nothing in these sections of the Code or in any regulation promulgated thereunder suggests these provisions were intended to provide a general principle to be applied in locating a specific foreign country as the source of dividend income. [27] The additional considerations we have identified as relevant to whether the Commissioner abused his discretion by adopting a retroactive regulation do not militate against retroactive application of Treas.Reg. § 1.902-3(d)(l). According retroactive effect to the regulation would serve the interest of treating equally similarly situated taxpayers and would not have an unduly harsh result in this case. Finally, we find no evidence that Congress had implicitly approved a prior contrary rule. [28] See K. Davis, Administrative Law Text § 5.05, 133 (3d ed. 1972): Statutes may be retroactive without violating due process; the test is whether they are unreasonably retroactive. The same test applies to legislative rules, except that courts are less reluctant to upset administrative rules than to upset statutes, and except that rules must be within granted power as well as constitutional (footnote omitted). [29] A distinction similar to that between legislative and interpretative rules may become significant with regard to retroactivity when a new regulation alters an existing regulation promulgated under a statute that Congress has reenacted. The party seeking to escape retroactive application of the new regulation will argue that Helvering v. R. J. Reynolds Tobacco Co., supra, 306 U.S. 110, 59 S.Ct. 423, 83 L.Ed. 536, precludes such retroactive effect. It has been held, however, that the Reynolds Tobacco principle (that by repeated reenactment of a statute Congress gives its sanction to existing regulations) is applicable only when the regulations are legislative as opposed to “administrative” in character. Automobile Club of Michi *985 gan v. Commissioner of Internal Revenue, 20 T.C. 1033, 1041 (1953), aff'd, 230 F.2d 585, 589 (6th Cir. 1956), aff'd, 353 U.S. 180, 185-86, 77 S.Ct. 707, 1 L.Ed.2d 746, 751 (1957). See also Helvering v. Reynolds, 313 U.S. 428, 61 S.Ct. 971, 85 L.Ed. 1438 (1941), which confined Reynolds Tobacco to its facts, specifically the existence of a prior Regulation that negatived a tax liability that the new and putatively retroactive regulation prescribed, and announced that Congressional reenactment of a statute that had been given a settled construction was no more than an aid in statutory interpretation and, where no regulations embodied the prior construction, did not preclude retroactively applying a regulation contrary to that construction. [30] Strictly speaking, the question of retroactivity can arise only with respect to rules that are at least in part legislative in character. That is to say, to the extent a regulation merely interprets a statute, it in theory merely elucidates a meaning that has resided in the statute since its enactment. If an interpretative regulation merely clarifies what the language of the statute was intended to convey, it is ultimately misleading to term it retroactive. “It is no more retroactive in its operation than is a judicial determination construing and applying a statute to a case in hand.” Manhattan General Equipment Co. v. Commissioner of Internal Revenue, 297 U.S. 129, 135, 56 S.Ct. 397, 400, 80 L.Ed. 528, 532 (1936). On the other hand, it seems unrealistic to suppose that many interpretative regulations merely express the one correct and intended interpretation of the statute under which they were promulgated. Many interpretative regulations will make explicit the answers to questions that Congress did not anticipate. Others will offer an answer that never crystallized during the legislative process. Professor Davis writes: [A] significant portion of what is called “interpretation” is not interpretation at all but is in truth creative law making. Whenever interpretative rules do in fact make new law, retroactive law making should be dealt with as such, unprejudiced by the false notion that results never flow from the interpreter. Problems of retroactivity then will be solved on the basis of ideas of fairness and the necessities of practical administration. Retroactive clarification of uncertain law ordinarily involves no unfairness. It is retroactive change of settled law, not retroactive settling of unsettled law, which may produce unjust results. K. Davis, Administrative Law Text, § 5.05, 135 (3d ed. 1972). By the same token, however, even legislative regulations must be consistent with the statute under which they were promulgated. We term them “legislative” because they are made pursuant to a specific delegation of authority and often without the particular legislative guidance typically found in statutes that spawn only interpretative regulations. But in a real sense they still interpret or explain existing legislation. The ideal types of legislative and interpretative regulations thus quickly break down in practice. Although the distinction has considerable utility for some purposes, that one regulation is denominated legislative in character and another interpretative in character contributes little to an understanding of whether each ought to be applied retroactively. In any event, whatever sharpness the distinction between legislative and interpretative rules might otherwise have is dulled by § 7805(a) of the Code, which authorizes the Secretary generally to prescribe all rules that enforcement of the Code requires. See Continental Equities Inc. v. Commissioner of Internal Revenue, 551 F.2d 74, 82 (5th Cir. 1977) (containing language that would eliminate the distinction entirely in tax cases by characterizing as legislative all regulations issued pursuant to § 7805(a).) For purposes of determining retroactivity, at least, the emphasis should be not whether a regulation more closely resembles the legislative or interpretative ideal type, but how the new regulation stands in relation to prior law or policy. [31] When the foreign tax rate in the subsidiary’s country of incorporation is high, for example, it is unlikely that the subsidiary is operating from a tax haven country, and the amount of the distribution (and hence the percentage of United States tax that the domestic parent must pay presently) is correspondingly smaller. [32] In many, perhaps most situations, the sourcing rule embodied in Treas.Reg. § 1.902-3(d)(l) better effectuates the purpose of the § 902 tax credit to avoid double taxation. Suppose, for example, that the sourcing rule advocated by Anderson, Clayton were applied to determine the per country limitation on the foreign tax credits allowable to domestic Corporation X. Corporation X owns the voting stock of Corporation Y, which is incorporated in country Alpha. Y derives its income from sales of goods in Beta and Gamma. Y pays an income tax to Beta and Gamma with respect to income earned in each country and an income tax on all of its income to Alpha. Y distributes a § 963 dividend from its foreign base company sales income. X, which has received dividends from subsidiaries in Delta and Epsilon, elects the per country limitation. All Y’s income taxes will be deemed to have been paid to Alpha, as required by Treas.Reg. § 1.902-3(d)(2). If Treas.Reg. § 1.902-3(d)(l) governs, the dividend received by X from Y will be sourced in Alpha. Since for purposes of computing X’s foreign tax credit, its subsidiary Y’s taxes are deemed to have been paid to Alpha, and since for purposes of computing the limitation of X’s foreign tax credit, the dividend is deemed to be sourced in Alpha, X will receive the maximum foreign tax credit. On the other hand, if Anderson, Clayton’s sourcing rule governs, a very different result will obtain. X will be deemed to have no foreign source income from Alpha, but only from Beta and Gamma, where Y earned the income. But X (through its subsidiary, Y) would be deemed to have paid foreign taxes only to Alpha, not to Beta and Gamma. Hence, X would be entitled to no foreign tax credit at all. In its brief, the taxpayer demonstrates that the sourcing rule embodied in Treas.Reg. § 1.902-3(d)(l) will be subject to abuse. The hypotheticals taxpayer constructs correctly point out some dangers involved in sourcing dividends in the subsidiary’s country of incorporation and call for attention by the Commissioner. Nevertheless, we think that the disadvantages of the country of incorporation rule are counterbalanced by its advantages, as evidence by the hypothetical above. In any event, the proper dividend sourcing rule represents a policy choice that Congress has delegated to the Secretary, not the courts, and we decline to upset his considered choice. [33] With respect to the prior settlement agreement, the district court held that it was intended only to adjust operating income between the taxpayer parent and its foreign holdings, not to authorize deductions for foreign exchange losses in the value of dividends issued by a foreign subsidiary to its parent. Regarding the prior judicial decision, the court held that the Court of Claims “was attempting to rectify an unfortunate result of war and was not intending to set a precedent for future litigation.” [34] Against the Court of Claims’ flat statement that the settlement agreement concerned foreign branch accounting, taxpayer juxtaposes the court’s observation that “pursuant to [that] agreement, plaintiff thereafter consistently reflected gains and losses from its exchange fluctuations on accounts payable to or receivable from its foreign subsidiaries and unrelated concerns as well as from its foreign branches.” 168 F.Supp. at 543. This empirical account of taxpayer’s practices can hardly be taken as decisive on the question of its entitlements. [35] Adversary litigation is an important safeguard in the judicial process. Thus, “neither the taxpayer nor the government should be precluded from raising a relevant point of law unless it appears beyond doubt that the precise point was actually contested and decided (not merely assumed) in the prior litigation.” Pelham Hall Co. v. Hasset, 147 F.2d 63, 67 (1st Cir. 1945), quoted in 18 Moore’s Federal Practice § 0.422(2) (2d ed. 1974).