Opinion
No. 98-5033
DECIDED: February 9, 1999
Appealed from: United States Court of Federal Claims, Judge Robert H. Hodges, Jr.
Matthew J. Zinn, Steptoe Johnson, of Washington, DC, argued for plaintiff-appellant. With him on the brief wereJames Walker Johnson and Theodore R. Groom, Groom Norlberg, of Washington, DC.
Edward T. Perelmuter, Attorney, Tax Division, U.S. Department of Justice, of Washington, DC. With him on the brief were Loretta C. Argrett, Assistant Attorney General andDavid I. Pincus, Attorney.
William B. Harman, Jr. Davis Harman, of Washington, DC, amici curiae for The Stock Company Information Group. On counsel on the brief were John T. Adney andKirk Van Brunt.
Before RADER, Circuit Judge, FRIEDMAN, Senior Circuit Judge, and BRYSON, Circuit Judge.
A provision of the Internal Revenue Code requires a mutual life insurance company to reduce a particular deduction from its income (and thereby increase its taxable income) by "the excess of" the "imputed earnings rate [of stock life insurance companies] over . . . the average mutual [life insurance company] earnings rate." The appellant mutual life insurance company contends that if "the average mutual earnings rate" is greater rather than less than "the imputed earnings rate," it may increase its deduction (and thereby decrease its taxable income) by that amount. The Court of Federal Claims rejected the company's theory, and we affirm.
I
A. Life insurance companies traditionally rebate to their policy holders, as excessive charges, part of the premiums paid and deduct these payments from their income. In addition to these so-called policyholder dividends, stock life insurance companies also pay dividends out of earnings to their shareholders, which the company cannot deduct. Mutual companies, however, have no stockholders; the policyholders in effect own the company. Accordingly, the policyholder dividends that mutual companies pay to their policyholders also may include a distribution of earnings that cannot be identified as such because the mutual company distributes only the single category of policyholder dividends. Therefore, a deduction of policyholder dividends that does not distinguish between mutual and stock companies may give mutual companies an advantage, because they can effectively deduct otherwise non-deductible distributions of earnings.
B. In 1984 Congress revised the Code provisions governing the taxation of life insurance companies and set out new rules distinguishing between stock and mutual company policyholder dividend deductions. While allowing stock companies to deduct all their policyholder dividends, the new rules provided a series of complex calculations designed to determine what portion of a mutual life insurance company's policyholder dividends constitute a distribution of earnings and reduce the mutual insurance company's policyholder dividend deduction accordingly. I.R.C. §§ 808(c), 809 (Supp. IV 1986).
The Internal Revenue Service first compares the average earnings of the two segments of the industry — (1) an "imputed" rate for stock companies, set at 16.5 percent for 1984 and thereafter annually recomputed on a three year basis pursuant to a formula, §§ 809(c)(1)(A), (d), and (2) the actual average earnings rate for the mutual segment of the industry, albeit from "the second calendar year preceding the calendar year in which the taxable year begins," § 809(c)(1)(B), because the current average rate is unavailable before the filing deadline. The "excess" of the "imputed earnings rate" for the taxable year "over" the "average mutual earnings rate" for the earlier year, the "differential earnings rate," § 809(c)(1), is then multiplied by the particular mutual company's "average equity base," § 809(a)(3). The resulting dollar "differential earnings amount" is then deducted from the amount paid or accrued as policyholder dividends in the taxable year. § 809(a)(1).
The House Committee Report on the legislation explained the theory of these calculations as follows:
Because mutual companies' policyholders are also the owners of the enterprise, policyholder dividends paid to them are distributions from the company that are a combination of price rebates, policyholder benefits and returns of company profits. Although there is no precise way to segregate a policyholder dividend or other payment into these various components, the committee believes that it is valid to conclude that profit oriented enterprises tend to distribute earnings to their owners in amounts that are proportional to the owners' equity in the business. Thus, the committee believes that the portion of a policyholder dividend that is a distribution of earnings can be measured as a percentage of the mutual company's equity ("the average equity base"). To determine the appropriate percentage of the equity base, the after [policyholder] dividend rates of return on equity for both stock and mutual companies were examined, and it was determined that the pre-tax return on equity of mutual companies falls below that for a comparable group of stock companies. The committee believes that this difference is attributable to distribution by mutual companies of earnings to their owners.
Under the bill, this theoretical approach to identifying ownership distributions is given effect by means of a reduction in the policyholder dividends deduction by a "differential earnings amount."
H.R. Rep. No. 98-432, pt. 2, at 1422 (1984), reprinted in 1984 U.S.C.C.A.N. 697, 1067.
Congress recognized that the use of the older average mutual earnings rate would lead to inaccuracy. Because it could not directly correct the problem it required mutual companies to base their tax calculations upon the older rate and then, in the next year, after the average mutual rate for the previous year is published, determine a recomputed differential earnings amount by making the same calculations but using the updated average mutual industry rate. § 809(f)(3). According to § 809(f)(2), where
the differential earnings amount . . . exceeds . . . the recomputed differential earnings amount . . . such excess shall be allowed as a life insurance deduction for the succeeding taxable year and § 809(f)(1) states that where "the recomputed differential earnings amount . . . exceeds . . . the differential earnings amount" that "excess" should be included in income.
A hypothetical example may clarify the procedure. Assume that in 1997 a mutual company is preparing its return for 1996. The company determines its taxable income based upon the determination of its policyholder dividends deduction. Because final data on the average earnings of the mutual industry is not available when the company prepares its 1996 return in 1997, it uses the average earnings figure for 1994 to determine the differential earnings rate.
In the next year, 1998, the company prepares its 1997 return in the same way. In addition, it redetermines its 1996 tax on the basis of the final average mutual earnings rate for 1996 that the Internal Revenue Service has since announced, recomputing the "excess" for the earnings rate under § 809(c)(1) and the dollar differential earnings amount under § 809(a)(3). Instead of filing an amended return for 1996, however, § 809(f)(2) requires the company to adjust its 1997 income either up or down, depending on whether the recomputed differential earnings amount is greater or lesser than the original differential earnings amount.
C. In its taxable year 1986, the appellant CUNA Mutual Life Insurance Company paid or accrued policyholder dividends of $36,365,926 and its differential earnings amount was $11,937,382, which produced a policyholder dividend deduction of $24,428,544. When CUNA recomputed its 1986 policyholder dividend deduction the following year, however, the newly calculated 1986 average mutual earnings rate was greater than the imputed rate. CUNA took the difference between the rates, multiplied it by its 1986 average equity base, and labeled the resulting figure, $1,919,903, as a negative recomputed differential earnings amount. In its 1987 return CUNA calculated the excess of the original differential earnings amount, $11,937,382, over its recomputed amount, negative $1,919,903, at $13,857,285 and deducted that amount from its 1987 income. The calculation produced a loss for CUNA for 1987, which it carried back to 1984.
On audit, the Service disallowed the loss carryback to the extent that it relied on the recomputed "negative excess," and assessed an additional tax for 1984. After CUNA paid the deficiency the Service denied its refund claim.
D. In 1994, the Treasury issued Treas. Reg. § 1.809-9, which provides in part that "[n]either the differential earnings rate under section 809(c) nor the recomputed differential earnings rate that is used in computing the recomputed differential earnings amount under section 809(f)(3) may be less than zero." Treas. Reg. § 1.809-9(a). The Treasury explained this regulation as follows:
both the differential earnings rate and the recomputed differential earnings rate are determined based on "the excess of" the imputed earnings rate over the average mutual earnings rate. Whenever the average mutual earnings rate is greater than the imputed earnings rate, there is no "excess of" the imputed rate over the average mutual earnings rate. By referring to the excess of the imputed earnings rate over the average mutual earnings rate, the statutory formula does not permit the differential earnings rate to be negative.
T.D. 8499, 1993-2 C.B. 246. The regulation is effective for all taxable years beginning after December 31, 1986. Treas. Reg. § 1.809-9(c).
E. CUNA filed the present refund suit in the Court of Federal Claims in 1996. On cross-motions for summary judgment, the court granted the government's motion. The court held that Treas. Reg. § 1.809-9 was not "unreasonable or plainly inconsistent with the Internal Revenue Code," CUNA Mutual Life Insurance Company v. United States, 39 Fed. Cl. 660, 664 (1997) and precluded the use of any "negative excess."
II
A. As in any case of statutory interpretation, the starting point is the language of the statute. See Bailey v. United States, 516 U.S. 137, 144 (1995). Section 809 is captioned "Reduction in certain deductions of mutual life insurance companies." Section 809(c)(1) is the critical provision in issue here — the recomputation provision, while separate, utilizes the machinery of § 809(c)(1). It states that:
the differential earnings rate for any taxable year is the excess of —
the imputed earnings rate for the taxable year, over the average mutual earnings rate for the second calendar year preceding the calendar year in which the taxable year begins.
We read the words "in tax statutes . . . in their ordinary and natural sense," Helvering v. San Joaquin Fruit Inv. Co., 297 U.S. 496, 499 (1936). In ordinary usage, "excess" means "the amount or degree by which one thing or number exceeds another," and "exceed" means "to be greater than or superior to."Webster's Third Int'l Dictionary 791-92 (1968). There is nothing unclear or ambiguous about the statutory words. If the imputed earnings rate is less, rather than greater, than the average mutual earnings rate, there simply is no "excess" of the former over the latter. Indeed, the notion of a "negative excess," upon which CUNA's entire case rests, is a contradiction in terms. A "negative excess" is not an "excess" at all, but is a deficit or a deficiency.
As the court stated in American Mutual Life Insurance Co. v. United States, 43 F.3d 1172, 1175 (8th Cir. 1994),cert. denied, 516 U.S. 930 (1995), in rejecting the same contention by another mutual company,
The statute states that the desired figure is the "excess" of A "over" B. This is the same as asking for the amount by which A exceeds B. If B is greater than A, there is no excess, and the amount by which A exceeds B is zero.
CUNA argues that "excess" has a technical meaning within the insurance industry that recognizes positive and negative differentials. Before we disregard the plain meaning of an ordinary word for a technical one, however, we need some indication in the statute that Congress intended to use the word in that technical sense. See San Joaquin Fruit Inv. Co., 297 U.S. at 499 (where the meaning of the word "acquired" was at issue and the choices were (1) at the time a lease containing an option to purchase was made and (2) when the option was exercised, the Court concluded that "[t]he plain import of the word is `obtained as one's own' [and] [i]n th[is] common and usual meaning of the term the land was acquired when it was conveyed [pursuant to the exercise of the option]."). Here, there is nothing to indicate that Congress intended "excess" to have that technical meaning. Although it is used in a complicated context, the word itself simply refers to a calculation — the amount by which one item is greater than another.
When Congress intended that either affirmative or negative amounts — excesses or deficiencies — were to be given effect, it explicitly so provided. Section 809(f), which governs the recalculation of income in the subsequent year, provides:
Inclusion in income where recomputed amount greater
In the case of any mutual life insurance company, if — the recomputed differential earnings amount for any taxable year, exceeds
the differential earnings amount determined under this section for such taxable year,
such excess shall be included in life insurance gross income for the succeeding taxable year.Deduction where recomputed amount smaller
In the case of any mutual life insurance company, if —
the differential earnings amount determined under this section for any taxable year, exceeds
the recomputed differential earnings amount for such taxable year,
such excess shall be allowed as a life insurance deduction for the succeeding taxable year.
If Congress had intended to recognize a similar reciprocal treatment for excesses and deficiencies in determining the differential earnings rate under § 809(c)(1), presumably it would have similarly structured the latter section. Instead, however, it merely provided that the "excess" of one earnings rate over another would be used to reduce the deduction for policyholder dividends.
Furthermore, while there is no question that in § 809(f) the word "excess" refers solely to an affirmative amount, CUNA would give that same word a broader meaning in § 809(c)(1) as also covering negative amounts. However, "[i]t is `a normal rule of statutory construction' . . . that `identical words used in different parts of the same statute are intended to have the same meaning.'" Commissioner v. Keystone Consol. Indus., Inc., 508 U.S. 152, 159 (1993) (citations omitted). This rule has even greater force where, as here, the identical words are in the same statutory section. There is no indication that Congress intended to depart from that principle here.
CUNA argues that § 802(c)(2)(B)(ii) of the old Life Insurance Company Tax Act of 1955, Pub.L. No. 429,reprinted in 1956 U.S.C.C.A.N. 46, 48, used "excess" in a similar way to that in the current § 809(c), and refers to a statement in the legislative history of the 1955 Act that "the `excess' described in clause (ii) of subsection (c)(2)(B) could conceivably be a negative figure which would then reduce the amount to be determined under such subsection," H.R. Rep. No. 84-1098, at 960 (1955), reprinted in 1956 U.S.C.C.A.N. 2272, 2281. That statement, however, related to a different provision that dealt with the "excess of the amount by which net premiums . . . exceed the dividends to policyholders on such contracts." CUNA points to nothing in the history of the 1984 amendments indicating that Congress intended to apply the principle indicated in the legislative history of the earlier act to the different situation that § 809(c) addresses.
CUNA's position has another fatal flaw. Section 808 is captioned "Policy Dividends Deduction," and § 808(c) states:
(1) In general
Except as limited by paragraph (2), the deduction for policyholder dividends for any taxable year shall be an amount equal to the policyholder dividends paid or accrued during the taxable year.
(2) Reduction in case of mutual companies
In the case of a mutual life insurance company, the deduction for policyholder dividends for any taxable year shall be reduced by the amount determined under section 809.
"[T]he amount determined" under § 809, by which the policyholder dividend deduction is to be "reduced," is the "excess" specified in § 809(c)(1). Like the word "excess," the word "reduced" is a common, unambiguous, non-technical term that is given its ordinary meaning. See San Joaquin Fruit Inv. Co., 297 U.S. at 499. "Reduce" means "to diminish in size, amount, extent, or number." Webster's Third International Dictionary 1905. Under CUNA's interpretation of "excess" in § 809(c), however, the result of the "amount determin[ation]" under § 809 would be not to reduce the policyholder dividends deduction, but to increase it. This would directly contradict the explicit instruction in § 808(c)(2) that the deduction "be reduced." The word "reduce" cannot be interpreted, as CUNA would treat it, to mean "increase."
The fact that § 808(c)(2) permits only the reduction of the policyholder dividend deduction reinforces the conclusion that § 809(c)(1) does not cover a "negative excess." Since Congress authorized only a "reduction," it is unlikely that it intended to permit a "negative excess," the effect of which would be not to reduce, but to increase the deduction.
In essence, CUNA asks us to interpret these two statutory provisions that jointly require the reduction of the deduction by the excess of the imputed stock company earnings rate over the average mutual earnings rate to permit an increase in the deduction by the deficit between these two rates. We decline to engage in such legal legerdemain.
C. Without any support in the statutory language, CUNA argues that because the imputed rate is the cumulative average of three years' rates while the average mutual rate is not, Congress would have — if it had considered the issue — wanted to obtain an overall cumulative average of both rates, something that could be done only by taking into account both negative and positive differentials. CUNA asserts that its position would accomplish what it characterizes as the congressional purpose behind the 1984 revisions, taxing stock and mutual companies comparably.
We cannot say what Congress would have done if it had specifically focused on the problem. If there is no excess, the mutual company suffers no reduction of its policyholder dividends deduction. Congress might have concluded that in that situation there was no occasion to give the company the additional benefit of increasing its deduction by the amount of the deficiency. Furthermore, the comparability of the taxation of stock and mutual companies to which CUNA refers may have been nothing more than correcting the prior situation in which mutual companies were able to shelter from taxation a portion of their earnings distributed as part of policyholder dividends, which stock companies could not do.
The argument is subject to a more basic flaw, however. The statute Congress enacted did not take the form that CUNA urges. We cannot rewrite the statute to accomplish what we might believe Congress may have intended to do. Such changes in the statute are for Congress, and not this court, to make. See Commissioner v. Lundy, 516 U.S. 235, 252-53 (1996);Beck v. Secretary of the Dep't of Health and Human Servs., 924 F.2d 1029, 1034 (Fed. Cir. 1991) ("Regardless of their merits, these policy arguments may be implemented only by Congress. Our duty is limited to interpreting the statute as it was enacted, not as it arguably should have been enacted.").
III
Since we conclude that § 809(c)(1) unambiguously precludes the use of a "negative excess" amount to increase the policyholder dividend deduction, it is unnecessary to invoke Treas. Reg. § 1.809-9 which, as noted, provides that neither the differential rate nor the recomputed differential earnings rate "may be less than zero." In any event, if we had any doubt about the outcome, the regulation would dispel it, since, as the Eighth Circuit stated in American Mutual, 43 F.3d at 1176, discussed below, "[f]or the reasons already outlined, this regulation is a reasonable interpretation of the statute. Treasury regulations are entitled to great deference, and must be sustained unless unreasonable and plainly inconsistent with the revenue statutes."
IV
Our conclusion that § 809(c)(1) does not permit reducing the policyholder dividend deduction by the amount of a negative excess is in accord with, and supported by, the decisions of the two other courts of appeal that have decided the question.
In American Mutual, the Eighth Circuit, stating that the statute was designed to reduce, not increase, the mutuals' policyholder dividend deductions, ruled that the statutory language and its legislative history indicated that the mutual company is not entitled to deduct an amount greater than its policyholder dividends. Id. at 1174.
In Indianapolis Life Insurance Co. v. United States, 115 F.3d 430 (7th Cir. 1997), the court read the statutory provisions as permitting a mutual company to deduct only the amount of policyholder dividends paid or accrued in the first year — and ruled that this limitation capped the total deduction (the initial deduction in year one plus the recomputation adjustment in year two). Id. at 435-36. The court further held that any remaining ambiguity was dispelled by Treas. Reg. § 1.809-9, which it was required to respect as not unreasonable. Id. at 436.
At oral argument CUNA stated that it was asking us to create a conflict with the foregoing two cases. "As a general matter, we `do not create conflicts among the circuits without strong cause.' . . . Moreover, as the courts of appeals have long recognized, the need for uniformity of decision applies with special force in tax matters. In recognition of the principle that `uniformity among the circuits is particularly desirable in tax cases to ensure equal application of the tax system to our citizenry,' we have held that we should not `reach a result in conflict with' a sister circuit `unless the statute [at issue] or precedent of this court gives us, in our view, no alternative.'"Washington Energy Co. v. United States, 94 F.3d 1557, 1561 (Fed. Cir. 1996) (citations and footnote omitted). In the present case, we agree with the decisions of these two other circuits, and have no reason to create a conflict with them.
CONCLUSION
The judgment of the Court of Federal Claims is
AFFIRMED.