OPINION
Featherston, Judge:
Respondent determined a deficiency in the amount of $1,149,015 in petitioner’s Federal income tax for 1972. The only issue remaining for decision is the amount of Canadian taxes paid by petitioner’s Canadian subsidiary which will be deemed to have been paid by petitioner pursuant to section 902(a)1 as a result of a dividend it received from the subsidiary in 1971.2 The resolution of this issue will require us to determine the effect of a net operating loss carryback properly claimed by the subsidiary under Canadian law on the "deemed paid” computation of section 902(a).
All of the facts have been stipulated.
Petitioner Champion International Corp. (Champion or petitioner), a New York corporation, had its principal place of business in Stamford, Conn., when the petition in this case was filed. Champion and its affiliates filed a calendar year consolidated Federal income tax return for 1972 with the Internal Revenue Service Center, Holtsville, N.Y.
At all times relevant to this case, petitioner owned stock which represented 74.9 percent of the total voting stock in Weldwood of Canada, Ltd. (Weldwood or the subsidiary), a Canadian corporation engaged in the forestry products industry in the Province of British Columbia. The remainder of Weldwood’s stock was publicly traded on Canadian stock exchanges.
Weldwood operated at a profit in 1968 and 1969, paying Canadian Federal and Province income taxes, but realized a loss from its operations in 1970. The loss was carried back to 1969 under Canadian law, and Weldwood received a refund of part of the Canadian taxes that it had paid for 1969. In 1971, Weldwood again realized a profit and, on November 23 of that year, it paid petitioner a dividend of C$7,831,8873 from which the dispute in this case arises. In all of the years 1968 through 1971, Weldwood paid dividends on the publicly traded stock.
The following table shows the amount of Weldwood’s "accumulated profits”4 (before giving any effect to the 1970 loss), the amount of the Canadian taxes paid (without regard to the 1969 refund), and the amount of dividends paid on its publicly held stock:
Year Accumulated profits Canadian taxes paid Dividends on publicly held stock
1968 C$13,230,006 C$7,062,784 C$688,750
1969 12,237,999 6,336,356 689,541
1970 0 0 868,216
1971 2,859,833 76,051 11,078,750
In 1970, as noted above, Weldwood realized a loss. The'amount of the loss, computed according to U.S. tax principles, was C$2,150,505. A loss also resulted under Canadian tax principles, and, under Canadian tax law, it was carried back to 1969. As a result of applying this 1970 loss as a carryback to 1969, Weldwood received a refund of C$1,393,108 of the Canadian taxes it had paid for 1969. The result is that its net Canadian tax payment for 1969 was C$4,943,248.
The issue here is the amount of the Canadian taxes paid by Weldwood which petitioner will be deemed to have paid as a result of the C$7,831,887 dividend it received from Weldwood on November 23, 1971. Section 901(a), as amplified by section 901(b)(1),5 permits domestic taxpayers to elect to receive a credit, within prescribed limits, for any foreign income tax which they pay, or are "deemed” to have paid under section 902, during the taxable year. Under section 902, which is discussed in more detail below, a qualifying domestic corporation which receives a dividend from a foreign corporation is deemed to have paid the foreign taxes actually paid by the foreign corporation on the amount of accumulated profits which were the source of the dividend. The domestic corporation electing the "deemed paid” credit must increase its own income by the amount of the foreign corporation’s tax that the domestic corporation is deemed to have paid, as well as the amount of the dividend it actually receives. Sec. 78.6 The effect is to treat the domestic corporation as though it had received a distribution out of the foreign corporation’s before-tax profits and then paid the foreign income tax thereon itself.
The terms on which the deemed paid or indirect credit may be allowed are prescribed mainly by section 902(a)(1).7 that section, insofar as it is here pertinent, provides that, if a. domestic corporation which owns at least 10 percent of the voting stock of a foreign corporation receives dividends in any taxable year that are paid "out of accumulated profits” of the foreign corporation, the domestic corporation shall be deemed to have paid a portion of the foreign taxes actually paid by the foreign corporation on those accumulated profits. The portion deemed to have been paid by the domestic corporation shall be the same proportion of any income taxes paid by the foreign corporation to the foreign country "on or with respect” to such accumulated profits which the amount of such dividends bears to the amount of such accumulated profits in excess of such taxes.
Thus, section 902(a)(1) contemplates that the creditable taxes deemed to have been paid by the domestic corporation will be computed by multiplying the foreign taxes paid by the foreign corporation "on or with respect to” the year’s accumulated profits by a fraction. The denominator of the fraction is equal to the foreign corporation’s "accumulated profits in excess of * * * [creditable foreign] taxes,” and the numerator is limited "to the extent such dividends are paid [to the taxpayer] * * * out of accumulated profits.”8
Multiplying this fraction by the multiplicand, the foreign taxes paid, in a single step provides for two allocations. First, the credit may need to be apportioned among two or more shareholders who have received dividends, and the share allocable to each must be determined. Second, less than the entire year’s net profits may be distributed, and the credit must be limited to the portion of the foreign taxes that was paid "on or with respect to” the accumulated profits that were distributed. The fraction, dividends over accumulated profits in excess of the foreign taxes, multiplied by the foreign taxes paid on the accumulated profits, makes both allocations.9
Expressed in the terms of an algebraic formula, section 902(a)(1) contemplates that the foreign taxes deemed to have been paid by the domestic corporation will be determined as follows:
Foreign taxes deemed paid on dividend distribution „ . 1 , oreign ax Dividend received by Domestic corporation2 X Accumulated profits of foreign corporation minus foreign taxes paid3
The parties have stipulated that the dividend of C$7,831,887 paid to petitioner on November 23, 1971, represented distributions out of Weldwood’s accumulated profits of 1971 and prior years, as follows:
1971 .C$2,019,720
1970 . 0
1969 . 3,586,489
1968 . 2,225,678
7,831,887
There is no dispute with respect to the "deemed paid” credit to which petitioner is entitled for the foreign taxes paid on Weldwood’s accumulated profits of 1971 or 1968.10 The only issue is the amount of the credit to which petitioner is entitled for the foreign taxes paid on Weldwood’s accumulated profits of 1969.
Applying the section 902(a)(1) formula, petitioner computes the foreign taxes deemed to have been paid by it on Weld-wood’s 1969 accumulated profits as follows:
Foreign taxes deemed paid 11C$4,943,248 12C$3,586,489 13C$5,144,246 C$3,446,356
Using the same section 902(a)(1) formula, respondent computes foreign taxes deemed to have been paid by Champion as follows:
Foreign taxes deemed paid C$4,943,248 C$3,586,489 14C$7,294,751 C$2,430,365
The disagreement between the parties, it becomes apparent, is with respect to the denominator of the section 902(a)(1) fraction, and that disagreement is limited to the amount of Weldwood’s "accumulated profits” for 1969. The parties agree that the 1969 accumulated profits "out of’ which the dividend was paid, reflected indirectly in the numerator (shown by note 12 supra), must be reduced to reflect the 1970 loss carryback and the resulting refund. Petitioner argues that the 1969 accumulated profits in excess of creditable foreign taxes, the denominator of the fraction, should also be reduced by the loss carryback from 1970. Such a reduction would have the effect of decreasing the denominator of the section 902(a)(1) fraction and increasing the amount of the taxes deemed to have been paid. Respondent contends that, notwithstanding the effect of the 1970 loss on the numerator of the section 902(a)(1) fraction, Weldwood’s 1969 accumulated profits, as reflected by the denominator of the fraction, may not be reduced by the 1970 net operating loss.
We agree with petitioner.
Crucial to the decision are the meaning and implications of the term "accumulated profits” as used in section 902(a)(1). Section 902(c)(1)(A)15 defines the term "accumulated profits” as "gains, profits, or income computed without reduction” for foreign taxes paid. Section 1.902-l(e), Income Tax Regs., amplifies the statute by providing that accumulated profits for any taxable year is the sum of the "earnings and profits” of the foreign corporation for such year plus the foreign income taxes imposed on or with respect to the income to which the earnings and profits are attributable. In general terms, the earnings and profits of a foreign corporation for a particular year may be computed substantially as if the foreign corporation were a domestic corporation. Sec. 1.902-l(g), sec. 1.964-1, Income Tax Regs.
As to the numerator of the fraction, the word "dividend,” as used in section 902(a)(1), has the same meaning as it has in other income tax provisions, including section 316. Sec. 1.902-1(a)(6), Income Tax Regs.; Central Aguirre Sugar Co. v. Commissioner, 24 T.C. 630, 636-637 (1955). Section 316 defines the term "dividend” to mean any distribution of property made by a corporation to its shareholders out of "earnings and profits.” In determining the source of a dividend paid by a domestic corporation, consideration is given first to earnings and profits for the taxable year (notwithstanding prior or subsequent losses) and next to earnings and profits accumulated in earlier years, which may extend back to February 28, 1913. The earnings and profits of a corporation thus "embodies primarily the concept of an aggregate of a corporation’s undistributed profits.” H. H. Robertson Co. v. Commissioner, 59 T.C. 53, 78 (1972), affd. per order (3d Cir. 1974).
A dividend paid by a foreign subsidiary that may support a deemed paid foreign tax credit must also be paid out of the foreign subsidiary’s earnings and profits, but section 902(a)(1) and (c)(1) requires a further computation. That section refers to dividends paid "out of accumulated profits” and contemplates that accumulated profits will be determined by separate yearly calculations. "To properly give effect to the language of [section 902(a)(1) and (c)(1)], * * * it is necessary to relate the tax credit to the particular year or years in which the accumulated profits (from which the dividends were paid) were earned and taxed.” General Foods Corp. v. Commissioner, 4 T.C. 209, 216 (1944). The rule was explained in H. H. Robertson Co. v. Commissioner, supra at 79, as follows:
It is of critical importance to determine the "accumulated profits” of each year, so that they can be matched against the foreign taxes paid for that year, and so that when dividends are paid "out of’ or "from” such "accumulated profits,” a foreign tax credit may properly be computed as a portion (in accordance with the statutory formula) of the foreign taxes paid in respect of the "accumulated profits” of that year.
Once a separate yearly determination of accumulated profits has been made, the dividend paid in a particular year may exceed the accumulated profits of that year, as did the November 23, 1971, dividend here. It is then necessary to allocate the dividend to another particular year or other years. Section 902(c)(1), quoted in note 15 supra, authorizes the Secretary or his delegate to determine from the accumulated profits of what year or years the dividend was paid, treating dividends as having been paid from the most recently accumulated profits. As explained in H. H. Robertson Co. v. Commissioner, supra at 79-80:
a dividend is allocated first to the full extent that it can be absorbed by the "accumulated profits” of the most recent year from which it was paid; the remainder is then allocated to the "accumulated profits” of the next most recent year to the extent that it can be absorbed by such "accumulated profits,” and so on down the years. A separate foreign tax credit is then computed for each of the years to which the dividend has been allocated, and the determination for each such year is made only in respect of that portion of the dividend that is allocated to such year. * * * The foreign tax credit under section 902 is then determined for the year in which the dividend was paid by adding together the separate credits thus computed for each of the years from the accumulated profits of which the dividend was paid. * * *
Consistent with these principles, the Commissioner here allocated part of the November 23, 1971, dividend to 1971 accumulated profits, and this allocation exhausted the accumulated profits of that year!16 Since 1970 was a loss year, 1969 was the most recent year for which accumulated profits were available for distribution. As reflected by the figures in note 12 supra, only accumulated profits of C$3,586,489 were available for distribution as a dividend in that year after taking the loss carryback from 1970 into account. The remainder of the 1971 dividend was allocated to 1968 accumulated profits. The parties have agreed to the correctness of these allocations. They also agree on the resultant numerator of the section 902(a)(1) fraction for 1969.
The dispute here, as we have indicated, is with respect to the denominator of the section 902(a)(1) fraction — accumulated profits less foreign taxes paid. The accumulated profits in the denominator represent the fund from which the dividends were paid and must be determined in accordance with U.S. law. Rev. Rul. 63-6, 1963-1 C.B. 126; secs. 1.9Q2-l(e) — 1.902-1(g), Income Tax Regs.
Under respondent’s position, the term "accumulated profits” means one thing for purposes of the numerator of the section 902(a)(1) fraction and another ás it is used in the denominator. We do not agree with this position. In addition to being algebraically indefensible, respondent’s treatment of the term "accumulated profits” would inevitably create a distortion in the amount of the "deemed paid” credit. As two commentators have noted:
After all the profits of a year have been distributed, the quantity used in the numerator of the fraction, "dividends,” must be the same as the amount used in the denominator, * * * [accumulated profits less foreign taxes]. Otherwise, the amount of indirect credit will be either more or less than the amount of the foreign tax actually paid. In short, the section 902 allocation fraction functions properly only when the definition of dividends (aggregate) and * * * [accumulated profits less foreign taxes] is exactly the same. * * * [1 E. Owens & G. Ball, The Indirect Credit 162 (1975).17]
Under respondent’s shifting definition of "accumulated profits,” the numerator and the denominator of the fraction would not be the same after all the year’s profits had been distributed. Because respondent would reduce the numerator, but not the denominator of the statutory fraction to reflect the loss carryback, the taxes deemed paid by petitioner would be less than its proportionate share of the taxes actually paid. That distortion can be prevented only by adjusting both the numerator and the denominator for the loss.
Furthermore, if, as respondent contends, the loss carryback does not reduce accumulated profits, but nevertheless limits the dividends that can be distributed "out of’ accumulated profits, then a portion of the accumulated profits of 1969 could never be distributed. In order to account for the source of the 1971 dividend, those inaccessible profits of 1969 would have to be bypassed, and the accumulated profits of 1968 would be considered the source of the remainder of the 1971 dividend. This treatment would violate the requirement of section 902(c)(1) that dividends be treated "as having been paid from the most recently accumulated gains, profits, or earnings.” This section clearly contemplates that all of the accumulated profits of 1969 shall be treated as having been distributed before looking to the accumulated profits of 1968 or earlier years. Under petitioner’s position, which we have adopted, the definition of "accumulated profits” is consistent in both the numerator and denominator of the section 902(a) fraction, and all of the accumulated profits of 1969 are considered distributed before those of 1968 are reached.
To support his position, respondent emphasizes the literal language of section 1.902-l(e)(l), Income Tax Regs., which defines accumulated profits as the sum of "(1) The earnings and profits of such corporation for such year, and (2) The foreign income taxes imposed on or with respect to the gains, profits, and income to which such earnings and profits are attributable.” (Emphasis added.) Respondent argues that the denominator of the section 902(a)(1) fraction for 1969 may not be reduced by the 1970 loss because this treatment of the loss would violate accepted principles of U.S. income tax law under which a loss reduces earnings and profits only in the year in which the loss is sustained. Rev. Rui. 64-146,1964-1 C.B. (Part 1) 129.
If this case were subject to decision solely on the basis of literal language, the words of section 902(a)(1) would be decisive. As pointed out above, section 902(a)(1) provides that, to the extent a qualifying domestic corporation receives dividends which are paid by a foreign corporation "out of accumulated profits,” the domestic corporation shall be deemed to have paid the same proportion of foreign taxes paid "on or with respect to such accumulated profits, which the amount of such dividends * * * bears to the amount of such accumulated profits in excess of such [foreign taxes].” (Emphasis added.) The repeated reference to "such accumulated profits [emphasis added]” is a clear indication that the meaning assigned to the term is to be the same throughout the computation.
In this area of the law, however, fraught with complex definitional concepts and requiring the meshing of U.S. tax laws with those of foreign countries, we look not only to the language of the statute, but also to its purpose. We observe the following teaching of Associated Telephone & Telegraph Co. v. United States, 306 F.2d 824, 832 (2d Cir. 1962):
The answer to the question before us is not to be found by relying upon terminology, but in a consideration of the policies behind the foreign tax credit allowance generally, and the policies behind section 902(a) in particular. The basic policy underlying all the foreign tax credit provisions is the elimination of double taxation!18] * * *
Petitioner points out, and respondent in his reply brief concedes, that in the instant case petitioner will not receive the full benefit of a credit of the foreign taxes Weldwood paid if the 1970 loss is subtracted in determining the numerator but not the denominator of the section 902(a)(1) fraction.19 To this extent, petitioner will pay a double tax with respect to a portion of its accumulated profits contrary to the basic policy of the statute. This is a direct result of reducing the numerator but not the denominator of the section 902(a)(1) fraction by the amount of the carried back loss.
As noted above, one purpose of the section 902(a)(1) formula is to allocate to each shareholder (in this case, petitioner and the public shareholders) its share of the credit resulting from the taxes paid by the foreign corporation on the share of the accumulated profits received by that shareholder. The numerator of each shareholder’s apportionment fraction is the amount of dividends paid "out of’ the accumulated profits of the year in question. The total dividends paid out of accumulated profits, however, are limited to the accumulated profits available for distribution, those remaining after foreign taxes. The sum of the numerators of all the shareholders’ respective apportionment fractions, therefore, should normally equal the accumulated profits less foreign taxes. That is, the sum of the numerators should equal the denominator, and the sum of the fractions should equal one. When this is the case, all of the foreign taxes may be apportioned to one shareholder or another,20 once all of the year’s accumulated profits available for distribution have been paid out.
This sensible result would not be achieved under respondent’s interpretation of the statute. Because, under his theory, the accumulated profits which are the source of the dividends in the numerator must be reduced by the loss carryback, while the accumulated profits in the denominator remain unreduced, the apportionment fractions of the several shareholders cannot, in his computation, add up to one. Accordingly, some of the foreign taxes would remain unapportioned, and, to that extent, the statutory policy of preventing double taxation would be defeated.21
In addition to conflicting with the general purpose of the foreign tax credit, the prevention of double taxation, respondent’s position conflicts with the narrower purpose of the deemed paid credit of section 902(a)(1). As stated by the court in Associated Telephone & Telegraph Co. v. United States, 306 F.2d at 832—
the purpose of Section 902(a) is to extend the same tax treatment to a domestic corporation with a foreign subsidiary as is accorded a domestic corporation with a foreign branch within the same corporate shell. [Fn. ref. omitted.]
Had Weldwood been a branch of Champion, rather than its subsidiary, Champion would have been entitled to claim direct credits, year by year, subject to the limitations of section 904(a)22 for the Canadian taxes paid on the profits of the branch’s operations. The credit for 1969 would have been recomputed pursuant to section 905(c)23 to adjust for the refund attributable to the loss carryback from 1970, but the total credits allowed to Champion, subject to section 904(a), would equal the Canadian taxes actually paid on the branch’s profits. Under respondent’s position, however, a different result would obtain because Weldwood was operated as a subsidiary. As shown above, respondent’s approach provides for a section 902(a)(1) credit which is less than the amount of the Canadian taxes, net of refund, that Weldwood actually paid.24
In developing his argument that reducing accumulated profits in the denominator of section 902(a)(1) would violate U.S. tax laws, respondent emphasizes a statement in H. H. Robertson Co. v. Commissioner, 59 T.C. at 78, that "like criteria are employed in determining 'earnings and profits’ and 'accumulated profits.’ ”25 Respondent then goes on to point out that a net operating loss affects earnings and profits of the year in which the loss arises, not the year to which it is carried, and that the tax refund generated by the loss is similarly reflected in the earnings and profits of the loss year.26 Respondent concludes that the reduction of accumulated profits by the amount of a loss carryback would violate the principle that accumulated profits and earnings and . profits are to be determined according to like criteria. We disagree.
The general statement in H. H. Robertson Co. that like criteria apply in the determination of both earnings and profits and accumulated profits is undeniably correct because of the relationship between the two concepts, but closer analysis is required in this case. A payment to a shareholder is a dividend under U.S. tax law if it is made out of current or accumulated earnings and profits. Sec. 316. Accumulated earnings and profits are reduced by losses in any prior years. Thus, under U.S. tax law, Weldwood’s 1970 loss would have reduced the accumulated earnings and profits available to cover the 1971 dividend by an amount equal to the loss. The particular year or years to which the accumulated earnings and profits as of 1971 were attributable would not need to be identified in order to establish that the November 23, 1971, payment was a dividend.
Under section 902(a)(1), however, it is not sufficient merely to establish that the source of the 1971 dividend was accumulated profits of earlier years. Annual computations are required. The purpose of U.S. tax principles in the context of section 902 is to establish a correlation between the foreign taxes paid on the foreign subsidiary’s accumulated profits of a particular year and the dividends paid from the same year’s accumulated profits. To ignore the fact that foreign law computes the tax liability by allowing for loss carrybacks is to assure that this correlation will be lost.
Petitioner’s resolution of this problem does not conflict in any fundamental way with U.S. earnings and profits criteria; it merely reconciles those criteria with the principle that "it is necessary to relate the tax credit to the particular year or years in which the accumulated profits (from which the dividends were paid) were earned and taxed.” General Foods Corp. v. Commissioner, 4 T.C. at 216. That reconciliation is then applied consistently throughout the section 902(a)(1) computation, in the denominator as well as in the numerator of the fraction. Only in this way can the purpose of the section 901 credit be achieved and the language of section 902(a)(1) followed.
Respondent relies heavily upon H. H. Robertson Co. v. Commissioner, supra, but his reliance is misplaced. In that case, the taxpayer, a domestic corporation, had liquidated its foreign subsidiary, and had been required by the Commissioner to include the subsidiary’s earnings and profits in income as a dividend. Because a previous distribution of appreciated property had reduced the subsidiary’s accumulated profits by the property’s full fair market value, but had reduced earnings and profits only by the property’s basis, earnings and profits exceeded accumulated profits available for distribution at the time of the liquidation.
In determining the deemed paid credit to which it was entitled as a result of the dividend, the taxpayer sought, in effect, to include the full amount of the dividend in the numerator of the section 902(a)(1) fraction. Had it been permitted to do so, the fraction would have been greater than one, and the taxpayer would have received a credit for taxes which its subsidiary never paid.
The Commissioner argued in H. H. Robertson Co. that the numerator of the fraction in the statutory formula was not necessarily the amount of dividends included in income, but the amount paid "out of’ the accumulated profits of the particular year. The numerator of the fraction was, therefore, limited to the accumulated profits of the year in question remaining available for distribution after foreign taxes thereon had been paid. Because the denominator of the fraction is also accumulated profits less foreign taxes thereon, the fraction in the Commissioner’s computation was equal to one, and, when multiplied by foreign taxes paid, yielded a deemed paid credit equal to the foreign taxes actually paid by the subsidiary. This Court held for the Commissioner, finding that the taxpayer’s position was not supported by the language of the statute or the policy underlying it. The limitation of the dividend in the numerator of the statutory fraction to the accumulated profits available for distribution followed directly from the words of the statute: "to the extent such dividends are paid * * * out of accumulated profits.” H. H. Robertson Co. v. Commissioner, 59 T.C. at 87. In addition, the taxpayer’s computation, in yielding a deemed paid credit greater than the foreign taxes actually paid by the subsidiary, went beyond the statute’s underlying purpose of preventing double taxation. H. H. Robertson Co. v. Commissioner, 59 T.C. at 86.
When these criteria, the language and purpose of the statute, are applied to the contending arguments in the instant case, it is respondent’s theory that is revealed as defective. As described above, respondent argues that the term "accumulated profits” means one thing for purposes of the numerator and another as it is used in the denominator. The statute contains no warrant for this algebraic anomaly. On the contrary, as discussed above, the repeated references in section 902(a) to "such accumulated profits” indicate that the meaning of the term is to be constant throughout the prescribed computation. Contrary to both the language and purpose of the statute, respondent’s method leads in this case to the imposition of double taxation. Therefore, when measured by the standards which were determinative in H. H. Robertson Co., respondent’s argument must fail.
Respondent also relies on Rev. Rui. 74-550; 1974-2 C.B. 209, which he describes as "the Service’s only published attempt to deal with foreign net operating losses in the context of the computation of a section 902 of the Code deemed-paid foreign tax credit.” That ruling deals specifically with "the determination of the taxable year out of which dividends are paid for the purpose of computing the foreign tax deemed paid by a domestic corporate shareholder under section 902.” It states that: "A deficit in earnings and profits for any taxable year shall reduce the most recently accumulated earnings and profits for a prior year.” Respondent argues here that:
under Rev. Rui. 74-550, a deficit in earnings and profits in one year reduces the earnings and profits available in earlier years for purposes of the section 902 of the Code dividend numerator, but the annualized accumulated profits in the denominator is not affected by deficits in other years.
We find nothing in the ruling that calls for that skewed result. The discussion and examples contained in the ruling deal with the question of which year’s accumulated profits are the source of dividends, i.e., with the numerator of the fraction in the section 902(a) formula. Nothing is said directly about the denominator. The ruling concludes, however, by stating that: "The holdings in the instant case are for computing the foreign tax deemed paid under section 902 of the Code for the years of dividend distribution.” This statement indicates that the reduction of accumulated profits by the amount of the subsequent year’s loss is to be effective for the entire section 902 computation, not just for part of it. Finally, the ruling specifically considers a case in which "the applicable foreign country’s income tax law has no loss carryback or carryover rules.” The ruling is thus distinguishable from the instant case, in which the effect of the Canadian carryback provisions on the section 902 computation is at the center of the controversy.27
In addition to his primary theory, which we have discussed and rejected above, respondent argues as follows:
As an alternative to permitting the loss carryback to affect the numerator but not the denominator of the section 902 of the Code equation, respondent contends that the loss carry back should not reduce the earnings and profits for the year to which the loss is carried for purposes of either the dividend numerator or the accumulated profits denominator of the section 902 of the Code equation. Under this alternative approach, a loss reduces earnings and profits in the year of the loss and any foreign tax refund resulting from a loss carryback or carryover increases the foreign corporation’s earnings and profits in the year of the loss (that is, the refund reduces the amount of the loss).* * *
Under this alternative theory, the amount of foreign taxes paid for the year to which the loss is carried would apparently be reduced by the amount of the refund, just as it was under both respondent’s primary theory and petitioner’s theory.28
This theory, like respondent’^ primary one, cannot be reconciled with the words of the statute. Section 902(a) calls for a fraction whose numerator is the "dividends * * * out of accumulated profits” and whose denominator is the excess of "such accumulated profits” over the foreign taxes thereon. The fraction is to be multiplied by the foreign taxes paid "on or with respect to such accumulated profits.”
In suggesting that the foreign taxes be reduced to reflect the refund occasioned by the carryback, while the dividends and accumulated profits remain unaffected by the carryback, respondent has proposed a formula whose terms do not stand in the coherent relationship provided for in the statute. The foreign taxes under that formula are not those which were paid "on or with respect to such accumulated profits,” because the taxes have been changed to reflect the carryback while the accumulated profits have not.29 In addition to conflicting with the terms of the statute, this treatment is contrary to our statement in H. H. Robertson Co. v. Commissioner, 59 T.C. at 79, that: "It is of critical importance to determine the 'accumulated profits’ of each year, so that they can be matched against the foreign taxes paid for that year.” The two terms cannot be said to have been "matched” when, as respondent proposes, foreign taxes are reduced to reflect the loss carry-back, while accumulated profits remain unchanged.
Finally, respondent argues that: "Under certain circumstances, petitioner’s position results in over crediting of the foreign taxes paid.” Respondent illustrates his point with an example in which a wholly owned foreign subsidiary corporation operating in a country that does not allow loss carrybacks or carryovers has the following accumulated profits, foreign taxes, and dividends paid to the U.S. parent:
Accumulated Foreign profits taxes Dividends
Year 1 100 50 0
Year 2 (49) 0 1
Respondent states:
Under petitioner’s approach, the deemed-paid credit computation for Year 1 would be as follows:
Year 1 50 X ^ ^ gQ = $50 foreign taxes deemed paid
The $51 accumulated profits in the denominator for Year 1 is reduced from $100 by the $49 Year 2 loss. This results in a potential foreign tax credit of $50 for a dividend payment of $1.
Respondent argues that:
This is an obvious overcrediting situation because the total amount of income subject to United States income tax is $1 plus the $50 section 78 of the Code gross-up.
We do not agree with this argument. By posing a situation in which the tax law of the foreign country involved has no loss carryback provisions, respondent has assumed away one of the critical elements of this case, and made his example irrelevant. If the "deemed paid” computation is performed using respondent’s figures, but assuming that thé country involved has a carryback provision like the Canadian one here applicable, the weakness of respondent’s argument becomes apparent. The loss of 49 in year 2 would be carried back to year 1, yielding a tax refund of 24.5. The net taxes paid for year .1 would, therefore, be 25.5, and the section 902(a) computation, under the interpretation which we have here adopted, would be as follows:
Foreign tax deemed paid = 25.5 x 51 - 25.5
= 25.5 x 1 25.5
= 1
The taxpayer would therefore recognize dividend income in the amount of 2 (the actual dividend of 1 plus the section 78 "gross-up” dividend of 1), and would be eligible to claim a foreign tax credit of 1. In other words, the amount of the foreign tax credit, subject to the limitation of section 904(a), would be the amount of dividend income recognized by the taxpayer times the assumed rate of foreign tax on the accumulated profits out of which the dividend was paid. Thus, respondent’s own figures fail to produce an "overcrediting situation” when a loss carryback provision like the one concerned in this case is assumed to be in effect.30
All of respondent’s arguments having been rejected, we hold that petitioner was correct in reducing the accumulated profits for 1969 by the amount of the. loss carried back from 1970 for purposes of both the numerator and denominator of the fraction in the computation of section 902(a).
Steel Improvement & Forge Co. v. Commissioner, 36 T.C. 265 (1961), revd. and remanded 314 F.2d 96 (6th Cir. 1963), is not inconsistent with our holding in the present case. The taxpayer in that case was a domestic corporation, and Canadian Steel was its wholly owned Canadian subsidiary. The Court held that, in 1954, Canadian Steel paid the taxpayer a dividend of C$116,000,31 which was paid out of Canadian Steel’s accumulated profits of 1954 to the extent of C$64,000, and out of its accumulated profits of 1953 to the extent of C$52,000. Although Canadian Steel had earnings and profits and accumulated profits as defined by U.S. law for both years, it reported a taxable loss for 1954 in its Canadian tax return. The loss was due to a special "capital cost allowance” under Canadian tax law which would not have been an allowable deduction under U.S. law. Steel Improvement & Forge Co. v. Commissioner, supra at 277. Canadian Steel carried the loss back to 1953, offsetting the taxable income previously reported for that year and causing the taxes previously paid to be refunded. This Court held that the taxpayer was not entitled to a deemed paid credit with respect to the dividend because Canadian Steel had paid no taxes, net of refund, on the accumulated profits out of which the dividend had been paid.
In the instant case, we have held that the accumulated profits of 1969 were partially offset by the carryback of the loss from 1970. In Steel Improvement & Forge Co., the carryback of the loss did not affect the accumulated profits of the carryback year or the determination of what year’s accumulated profits were the source of the dividend therein concerned. This is because the loss was caused by a deduction not recognized by U.S. law. Therefore, although the loss eliminated the Canadian tax of both the loss year and the carryback year, it left the accumulated profits of those years unaffected. Accumulated profits are determined under American law rather than foreign law. Steel Improvement & Forge Co. v. Commissioner, supra at 277. In the instant case, however, the 1970 loss was not occasioned by deductions peculiar to Canadian law. The loss was recognizable for U.S. as well as Canadian tax purposes, and it reduced the American accumulated profits as well as the Canadian tax. Steel Improvement & Forge Co. is, therefore, distinguishable from the instant case, and is not inconsistent with our holding here.
To reflect the foregoing,
Decision will be entered under Rule 155.