Marsh & McLennan Companies, Inc. ("Marsh") appeals the decision of the United States Court of Federal Claims granting summary judgment for the United States. The court held that Marsh is not entitled to interest on overpayments of federal income tax for 1985 and 1986 (which were applied to satisfy tax liabilities for 1987 and 1988) for the period after the due date of its 1987 and 1988 tax returns. Marsh & McLennan Cos. v. United States, 50 Fed. Cl. 140 (2001). Because the language of the pertinent statute, as authoritatively interpreted by a Treasury regulation, plainly denies such interest, we affirm.

The facts of this case are not in dispute. Marsh, a financial services company, overpaid its 1985 federal income taxes by $275,139 ("1985 overpayment") and overpaid its 1986 taxes by $3,412,083 ("1986 overpayment").

Marsh also overpaid the amount of 1987 federal income taxes shown on the return by $12,694,216, and at the time the 1987 return was filed, Marsh elected to apply this overpayment to its 1988 tax liability ("1987 credit elect overpayment"). The Internal Revenue Service ("IRS") permits a corporate taxpayer that has paid more than its liability in one year to claim a credit for the excess against its estimated taxes for the succeeding year. 26 U.S.C. § 6402(b) (2000).1 The credit is called a "credit elect overpayment." As discussed below, credit elect overpayments do not earn interest. The 1987 credit elect overpayment was transferred by the IRS from Marsh's 1987 tax account into its 1988 tax account on March 15, 1989, the due date of its 1988 tax return.

Marsh also overpaid the amount of 1988 taxes shown on the return by $28,336,308, and at the time the 1988 return was filed, Marsh elected to apply this overpayment to its 1989 tax liability ("1988 credit elect overpayment"). A portion of the 1988 credit elect overpayment was transferred by the IRS from Marsh's 1988 tax account into its 1989 tax account on September 15, 1989, to satisfy its estimated 1989 tax liability, and the remainder of the 1988 credit elect overpayment was transferred from Marsh's 1988 tax account into its 1989 tax account on March 15, 1990, to satisfy its final 1989 tax liability.

Thus, Marsh had made four separate overpayments (if the 1987 and 1988 tax liabilities as shown in the returns were accurate): (1) the 1985 overpayment; (2) the 1986 overpayment; (3) the 1987 credit elect overpayment; and (4) the 1988 credit elect overpayment.

In 1994, the IRS determined that Marsh had in fact underpaid its 1987 taxes by $978,135. This increase in tax was less than the amount of the 1987 credit elect overpayment. The IRS applied Marsh's 1985 overpayment and a portion of its 1986 overpayment to its 1987 tax account on March 15, 1989, the date the 1987 credit elect overpayment was applied to Marsh's 1988 tax account.

Similarly, in 1994, the IRS determined that Marsh had in fact underpaid its 1988 taxes by $2,709,087. This increase in tax was less than the amount of the 1988 credit elect overpayment. The IRS applied the remainder of Marsh's 1986 overpayment to its 1988 tax account. The actual transfers were made on September 15, 1989 (the date the 1988 credit elect overpayment was applied to the estimated tax payment for 1989), and March 15, 1990 (the date the remainder of Marsh's 1988 credit elect overpayment was applied to Marsh's 1989 tax account).

There is no dispute concerning the amount of Marsh's 1985 and 1986 overpayments, the dates that those payments were credited to later years' taxes, the amount of the 1987 and 1988 credit elect overpayments, the dates on which those credit elect overpayments were needed to satisfy later years' obligations, or the amount of Marsh's 1987 and 1988 tax liabilities. The ultimate issue concerns the amount of interest due on the 1985 and 1986 overpayments.

An "overpayment" occurs when a taxpayer "has paid as an installment of the tax more than the amount determined to be the correct amount of such installment." 26 U.S.C. § 6403 (2000). The IRS has authority to either refund or credit the overpayment to any unpaid installment. Id. Unlike credit elect overpayments, overpayments bear interest. See id. § 6611(a). For the 1985 overpayment amount and the portion of the 1986 overpayment amount credited against Marsh's 1987 tax liability, the IRS allowed interest through April 15, 1988, the due date for Marsh's 1987 tax return (plus one month).2 For the portion of the 1986 overpayment amount credited against Marsh's 1988 tax liability, the IRS allowed interest through April 15, 1989, the due date for Marsh's 1988 tax return (plus one month). The government stopped the running of interest on those dates because it reasoned that overpayment interest runs only until the "due date" of a taxpayer's federal income tax return for the year to which the overpayment was applied.

The governing statute, 26 U.S.C. § 6611 (2000), entitled "Interest on overpayments," provides in relevant part that "[s]uch interest shall be allowed and paid as follows: ... In the case of a credit, from the date of the overpayment to the due date of the amount against which the credit is taken." Id. § 6611(b)(1). Treasury Regulation § 301.6611-1(h)(1) provides: "General rule. If an overpayment of tax is credited, interest shall be allowed from the date of the overpayment to the due date ... of the amount against which such overpayment is credited." 26 C.F.R. § 301.6611-1(h)(1) (2001). "Due date" is defined in subsection 301.6611-1(h)(2)(i): "In general. The term `due date,' as used in this section, means the last day fixed by law or regulations for the payment of the tax (determined without regard to any extension of time)...." Id. § 301.6611-1(h)(2)(i). For taxes reported on a return, this payment date is the last unextended date fixed for filing the return. 26 U.S.C. § 6151(a) (2000).

The parties apparently agree that, absent the credit elect overpayments, the two due dates involved here were March 15, 1988, and March 15, 1989 (i.e., the due dates for the 1987 and 1988 tax returns respectively) and that interest on the 1985 and 1986 overpayments would have ceased on those dates. Thus, the issue on appeal is the effect of the credit elect overpayments on the running of interest. It is well established that the taxpayer earns no interest on credit elect overpayments. If a taxpayer elects to credit against its estimated liability for one taxable year an overpayment shown on its return for the preceding year, it collects no interest even though at one point it has paid the IRS more than was due. 26 C.F.R. § 301.6611-1(h)(2)(vii) (2001); Martin Marietta Corp. v. United States, 216 Ct. Cl. 47, 572 F.2d 839, 841 (1978). The question is whether credit elect overpayments must be taken into account in calculating the interest due on other overpayments.

Marsh filed a complaint in the Court of Federal Claims seeking $833,522 in overpayment interest for the period between April 15, 1988, and March 15, 1989, for its 1985 overpayment and a portion of its 1986 overpayment; for the period between April 15, 1989, and September 15, 1989, for a portion of its 1986 overpayment; and between April 15, 1989, and March 15, 1990, for the remainder of its 1986 overpayment, the ending dates representing the dates on which the credit elect overpayments were applied to subsequent tax year liabilities. The government filed a motion for summary judgment, and Marsh filed a cross motion for summary judgment.

Marsh argued that 26 U.S.C. § 6611, which provides for the accrual of interest on overpayments "to the due date of the amount against which the credit is taken," should be construed to mean that interest runs until an account actually becomes deficient, that is, "due and unpaid." Marsh argued that its 1987 tax account was not actually deficient until March 15, 1989, when the 1987 credit elect overpayment was applied to its 1988 tax account. Similarly, Marsh argued that its 1988 tax account was not actually deficient until September 15, 1989, when a portion of its 1988 credit elect was applied to its 1989 estimated taxes, and on March 15, 1990, when the remainder of its 1988 credit elect was applied to its 1989 tax account.

The government argued that, under the plain meaning of section 6611 and Treasury Regulation § 301.6611-1(h)(2), the "due date of the amount against which the credit is taken" is the due date of a taxpayer's federal income tax return for the year to which the overpayment is applied. Marsh's 1985 overpayment and a portion of its 1986 overpayment were applied to its 1987 tax account, and the remainder of the 1986 overpayment was applied to Marsh's 1988 tax account. Marsh was required to file its 1987 tax return on March 15, 1988, and its 1988 tax return on March 15, 1989. Therefore, the government argued, Marsh was entitled to interest on its 1985 overpayment and a portion of its 1986 overpayment until March 15, 1988, and on the remainder of its 1986 overpayment until March 15, 1989.3

The Court of Federal Claims recognized that the case turned on the interpretation of section 6611, and determined that, read in isolation, the statutory language, "which is, at best, not a model of clarity," does not "yield[] a satisfactory result." Marsh, 50 Fed. Cl. at 143. Similarly, the court concluded that neither the "use of money" principle, under which overpayments and underpayments are generally netted in calculating interest, nor IRS revenue rulings provided the answer. Id. at 143-45. The court then assessed the relative "predictability" of the parties' competing interpretations. Id. at 146. The court noted that Marsh's interpretation would make interest accrual cease on the date an overpayment was actually credited against another tax liability, and, under Marsh's interpretation, "the `due date' becomes wholly dependent on the IRS's accounting procedures, which have not been shown to be governed by the statutory language of I.R.C. § 6611(b)(1)." Id. at 142. The court rejected Marsh's interpretation in large part because the "due date" would be impossible to predict in advance in that it depended on the "vagaries" of the IRS internal accounting procedures, and accepted the government's interpretation because it "harmonizes with the principle of predictability and comports with the statutory language." Id. at 146. Accordingly, the court granted summary judgment in favor of the government.

Marsh timely appealed. We have jurisdiction under 28 U.S.C. § 1295(a)(3).

We review a trial court's grant of a motion for summary judgment without deference. Impresa Construzioni Geom. Domenico Garufi v. United States, 238 F.3d 1324, 1330 (Fed.Cir.2001). Summary judgment is appropriate when "there is no genuine issue as to any material fact and ... the moving party is entitled to a judgment as a matter of law." Fed.R.Civ.P. 56(c).

This case involves the question when interest on tax overpayments ceases to run.

As noted above, the parties apparently agree that, if there had been no credit elect overpayment in 1987 or 1988, the "due date" for the additional tax due for 1987 and 1988 would have been March 15 of the year following the tax year in question, here March 15, 1988, and March 15, 1989. Under either party's interpretation of section 6611, interest on the 1985 and 1986 overpayments would have ceased to run on those dates.

The parties' disagreement arises from the existence of the 1987 and 1988 credit elect overpayments. Marsh argues that because of the credit elect overpayments, there was no outstanding liability in its 1987 account until the 1985 overpayment and a portion of the 1986 overpayment were applied to its 1988 tax account on March 15, 1989, and there was no outstanding liability in its 1988 account until the remainder of the 1986 overpayment was applied to its estimated 1989 tax on September 15, 1989, and its final 1989 tax on March 15, 1990.

Marsh urges that, as a result of its 1987 credit elect overpayment, it overpaid the government between the period of the credit elect overpayment and the due date of the tax return of the year to which the credit elect overpayment was applied, that is, from March 15, 1988, to March 15, 1989. Similarly, Marsh urges that, as a result of its 1988 credit elect overpayment, it overpaid the government from March 15, 1989, to September 15, 1989, and March 15, 1990.

Marsh contends that there was no payment "due" until its 1987 tax account actually became negative on March 15, 1989, i.e., the date that the 1987 credit elect overpayment was applied to its 1988 tax account. Similarly, Marsh contends that there was no payment "due" until its 1988 tax account became negative on September 15, 1989, the date when a portion of its 1988 credit elect overpayment was applied to its 1989 tax account, and again on March 15, 1990, when the remainder of its 1988 credit elect overpayment was applied to its 1989 tax account.

The government urges that the "due date" for purposes of section 6611 is the due date of the returns, and that this date does not change as a result of the 1987 and 1988 credit elect overpayments.

We agree with the government and disagree with Marsh for several reasons.

First, the language of section 6611(b)(1), as interpreted by Treasury Regulation § 301.6611-1(h)(2), requires this result.

Both parties argue from their view of the supposed plain language of section 6611(b)(1). In fact, the statutory language is ambiguous. The "due date" could refer either to the due date of the return (the government's position) or the date on which there is a net deficit in the amount due in the taxpayer's account for the particular year (the taxpayer's position). Under Chevron U.S.A., Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 843-44, 104 S. Ct. 2778, 81 L. Ed. 2d 694 (1984), we look to the Treasury regulations to resolve the ambiguity. See United States v. Mead Corp., 533 U.S. 218, 227-28, 121 S. Ct. 2164, 150 L. Ed. 2d 292 (2001); United States v. Haggar Apparel Co., 526 U.S. 380, 392, 119 S. Ct. 1392, 143 L. Ed. 2d 480 (1999). The Treasury regulations are, we think, quite clear. They provide: "General rule. If an overpayment of tax is credited, interest shall be allowed from the date of the overpayment to the due date ... of the amount against which such overpayment is credited," 26 C.F.R. § 301.6611-1(h)(1) (2001), and define "due date" as: "In general. The term `due date,' as used in this section, means the last day fixed by law or regulations for the payment of the tax (determined without regard to any extension of time)," id. § 301.6611-1(h)(2)(i). Marsh notes that because the regulations state that this definition of "due date" will be used "[i]n general," they do not purport to interpret the entire phrase "due date of the amount against which the credit is taken" in section 6611(b)(1). As the government points out, however, Treasury Regulation § 301.6611 is actually keyed to the relevant section of the code, section 6611, indicating that it does in fact relate to section 6611. We agree with the government that the regulation does apply.

Marsh also contends that the regulations are "unreasonable." We cannot agree. For reasons that we discuss below, there is nothing unreasonable in disregarding credit elect overpayments in calculating overpayment interest. And, while an IRS interpretation of its regulation might be entitled to deference where there is an ambiguity, Am. Express Co. v. United States, 262 F.3d 1376, 1382 (Fed.Cir. 2001), Marsh cites no interpretation of the regulations by the IRS in a context similar to this one that would contradict their plain meaning.4

Contrary to Marsh's argument, the legislative history of section 6611 does not support Marsh. Section 6611 as it reads today was enacted in the Technical Amendments Act of 1958, Pub.L. No. 85-866, § 83, 72 Stat. 1606, 1663-64 (1958), reprinted in 1958 U.S.C.C.A.N.1925, 1997, and is identical in relevant part to section 1019 of the Internal Revenue Act of 1924.5 Marsh urges that the legislative history establishes that the "due date against which the credit is taken" is the date on which the overpayment is economically offset against a liability. Marsh argues that the legislative history reveals that "Congress intended that interest accrue on a credited overpayment until the date on which an overpayment and underpayment simultaneously existed (i.e., on the date on which `mutuality of indebtedness' arose)," citing the Senate Finance Committee Report, S.Rep. No.1983, 85th Cong., 2d Sess., 234-35 (1958), reprinted in 1958-3 C.B. 922, 1156. But the legislative history on which Marsh relies does not address the precise issue here or suggest that a credit elect overpayment must be considered in calculating interest. In enacting and amending section 6611 and its predecessor, section 1019, Congress sought to address specific issues, such as: to eliminate the distinction between the running of interest in the situation where an overpayment is credited against a later-arising deficiency and where it is credited against an underpayment of original tax; to eliminate the running of interest where overpayments and underpayments offset each other; to address the anomaly in the transitional situation in which a taxpayer (which owed money to the government on taxes imposed by a prior revenue act) paid no interest but collected interest upon money the government owed it; and to eliminate the requirement of a protest as a condition precedent for a taxpayer to receive overpayment interest.6 Thus, the legislative history does not support Marsh's interpretation of the statute.

Second, contrary to Marsh's argument, section 6402(a) of the code does not require a different result. That section provides:

In the case of any overpayment, the Secretary, within the applicable period of limitations, may credit the amount of such overpayment, including any interest allowed thereon, against any liability in respect of an internal revenue tax on the part of the person who made the overpayment and shall, subject to subsections (c), (d), and (e) refund any balance to such person.

26 U.S.C. § 6402(a) (2000) (emphasis added). Marsh argues that there was no "liability" within the meaning of section 6402(a) until the credit elect overpayments were applied to subsequent years' taxes. Marsh points out that the IRS in fact did not apply the 1985 and 1986 overpayments to the 1987 tax account until the 1987 tax account actually became negative, that is, when the 1987 credit elect overpayment was transferred to the 1988 tax account on March 15, 1989. Similarly, the IRS did not apply the 1986 overpayment to the 1988 tax account until the 1988 tax account in fact became negative, that is, when the 1988 credit elect overpayment was transferred to pay 1989 estimated taxes on September 15, 1989, and the final 1989 taxes on March 15, 1990.

Section 6611 makes no reference to section 6402, and whatever limitations section 6402 may impose on the IRS's authority to credit, section 6402 does not define the term "due date" for purposes of section 6611. There is no indication in section 6402 that it was designed to affect the running of overpayment interest.

Third, Marsh argues that we should follow the Second Circuit's decision in Avon Products, Inc. v. United States, 588 F.2d 342 (2d Cir.1978), which Marsh interprets as holding that there is no tax liability until the credit elect overpayments were in fact credited to later years' taxes. Avon is distinguishable and does not stand for the broad proposition attributed to it by Marsh. In Avon the taxpayer initially reported an overpayment on its 1967 return, and made a credit elect. 588 F.2d at 343. Thereafter, it filed an amended return showing a 1967 deficiency, and the IRS later assessed an additional 1967 deficiency. Id. The IRS assessed interest on the deficiencies without regard to the credit elect overpayment. Id. The Second Circuit held that this was impermissible, and that the credit elect overpayment could not be disregarded; interest on the deficiencies did not run until the date the credit elect overpayment was applied to the next year's tax. Id. at 344. Marsh argues that the same should be true here, and that the amount of the deficiencies should be offset by the amount of the credit elect overpayments.

However, Avon was an underpayment case, interpreting the section of the statute governing underpayment interest, section 6601(a). That section provides "If any amount of tax ... is not paid on or before the last date prescribed for payment, interest on such amount ... shall be paid for the period from such last date to the date paid." 26 U.S.C. § 6601(a) (2000) (emphasis added). Unlike the situation here, the language of the statute contradicted the government's position since the statute focused on the question whether the tax was "paid," and not on the "due date" for the tax. As the Second Circuit noted, "if we were to construe § 6601(a) literally, it would not even be apposite to this case. Avon's full tax was in fact paid `on or before the last date prescribed for payment,' June 15, and so the premise of the provision is undercut." Avon, 588 F.2d at 344. To avoid a result clearly contrary to the statutory language, the court interpreted section 6601(a) as providing that underpayment interest shall begin running when a tax becomes "both due and unpaid." Id.

As the Avon court noted, the tax code differentiates between underpayment and overpayment interest situations. Id. at 345. It does not, as Marsh argues, require identical treatment.7 Underpayment interest ceases to run on "the date paid," 26 U.S.C. § 6601(a) (2000), while overpayment interest ceases to run on "the due date of the amount against which the credit is taken," id. § 6611(b). Thus, cases like Avon interpreting the section of the statute governing underpayment interest situations are irrelevant. See also May Dep't Stores Co. v. United States, 36 Fed. Cl. 680, 688 (1996) (holding that the government was not entitled to underpayment interest between the date of payment of the estimated tax and the date of filing of the return where the taxpayer overestimated its overpayment and elected to apply the overpayment to the following year's estimated tax as a credit elect, thereby creating a deficiency); Revenue Ruling 99-40, reprinted in 1999-2 C.B. 441 (analyzing the calculation of underpayment interest in the context of a credit elect overpayment and subsequently determined deficiency). Moreover, the court in Avon expressly distinguished the case before it from cases involving overpayment interest, and emphasized that the statutory scheme governing overpayment interest, as well as the policy considerations underlying it, were very different:

[T]his case is not covered by the cited regulation [Treasury Regulation § 301.6611-1(h)(2)(vii), governing overpayment interest]. Nor do we believe that it provides a helpful analogy. Indeed, the fact that there is a Treasury Regulation explicitly denying the taxpayer interest in the [credit elect overpayment] situation ... but nothing comparable suggesting that Avon must pay interest for its later-created deficiency would indicate that the contrary result is compelled.

Avon, 588 F.2d at 345.

To support its argument that there is symmetry in the payment of overpayment and underpayment interest, Marsh cites a section of the code that equalizes underpayment and overpayment interest rates, 26 U.S.C. § 6621(d) (2000). But the issue here is the running of interest, not what interest rate applies. Marsh also cites Brown & Williamson, Ltd. v. United States, 231 Ct. Cl. 413, 688 F.2d 747, 750 (1982), and United States v. Koppers Co., 348 U.S. 254, 267, 75 S. Ct. 268, 99 L. Ed. 302 (1955), for the same proposition. Brown & Williamson involved the amount of overpayment interest that should be paid on a refund resulting from a retroactive reduction of taxes by an income tax treaty. Brown & Williamson, 688 F.2d at 750. Koppers involved the calculation of interest under a provision allowing abatement of wartime excess profits taxes. Koppers, 348 U.S. at 267, 75 S. Ct. 268. Neither Brown & Williamson nor Koppers is helpful here.

Fourth, Marsh argues that the government had the use of its money and should pay interest. We agree that the IRS had the use of Marsh's money represented by the credit elect overpayments, and that Marsh will not receive interest on those amounts. However, this is not an anomaly. Once the taxpayer designates an overpayment as a credit elect overpayment, the taxpayer loses control over that money. Once the election is made, that election is irrevocable and binding on both the IRS and the taxpayer. Martin Marietta, 572 F.2d at 842. The regulations expressly provide that interest shall not be allowed on credit elect overpayments under section 6611(a). 26 C.F.R. § 301.6611-1(h)(2)(vii) (2001);8 Martin Marietta, 572 F.2d at 841. Marsh does not contest this principle, and admits that "[t]his exception is set forth in Treas. Reg. §§ 301.6402-3(a)(5) and 301.6611-1(h)(2)(vii), and was ratified by Congress."9 (Appellant's Reply Br. at 27 n. 14.) Just as the taxpayer cannot get interest on the credit elect overpayment, the taxpayer cannot use the credit elect overpayment to increase the amount of interest that otherwise would be due on the 1985 and 1986 overpayments.

In any event, the "use of money" cases, on which Marsh so heavily relies, cannot contradict the statutory language as interpreted by the regulation. In general, the "use of money" principle is a tool of statutory construction that supports the payment of interest to compensate one party for the time the other party had the use of its money. See, e.g., Manning v. Seeley Tube & Box Co., 338 U.S. 561, 565-66, 70 S. Ct. 386, 94 L. Ed. 346 (1950). The use of money principle has been recognized by the IRS generally,10 and has been applied by the statute and the regulations in specific situations.11 But Marsh has not called our attention to any case in which the use of money principle has been held to override statutory language requiring a contrary result. See MNOPF Trustees, Ltd. v. United States, 123 F.3d 1460, 1465 (Fed. Cir.1997) (determining the starting date for interest on an overpayment attributable to a refund to be the unextended date by which the banks that withheld the money were required to file returns reporting the withholding, based on the relevant Treasury regulation, rather than from the date that the government actually had "use of" the funds); May Dep't Stores, 36 Fed. Cl. at 688 (reaffirming the Avon principle that underpayment interest may be imposed only for the period during which a tax is both due and unpaid, based on "the plain language" of section 6601); Avon, 588 F.2d at 343-44 (recognizing the use of money principle as a background principle, then interpreting section 6601 as required by the statutory language and legislative history). The use of money principle does not trump the language of the relevant statute and Treasury regulations.

Finally, Marsh argues that if it had sought a refund instead of making the credit elect in 1987, it would have had the use of those funds. This is true enough, but we fail to see how that assists Marsh. There is no principle that requires that the taxpayer be treated the same whether it seeks a refund or a credit elect. Indeed, the statute as interpreted by the regulations provides that, for a refund, interest will be allowed from the date of the overpayment until "a date (to be determined by the Secretary) preceding the date of the refund check by not more than 30 days," 26 U.S.C. § 6611(a)(2) (2000), but denies interest on credit elect overpayments, 26 C.F.R. § 301.6611-1(h)(2)(vii) (2001). The taxpayer could have sought a refund for the excess funds, or left the excess funds as an interest-bearing overpayment. A taxpayer that makes a credit elect has no one to blame but itself for the non-payment of interest on that amount. See Avon, 588 F.2d at 345. See also Mellon Bank, N.A. v. United States, 265 F.3d 1275, 1281-82 (Fed.Cir.2001) ("The Supreme Court has `observed repeatedly that, while a taxpayer is free to organize his affairs as he chooses, nevertheless, once having done so, he must accept the tax consequences of his choice, whether contemplated or not....'" (quoting Comm'r v. Nat'l Alfalfa Dehydrating & Milling Co., 417 U.S. 134, 149, 94 S. Ct. 2129, 40 L. Ed. 2d 717 (1974))).

In essence, Marsh urges that we interpret the statute to achieve what it regards as a more just result. But we cannot revise the language of the statute as interpreted by the Treasury to achieve what might be perceived to be better tax policy, just as we cannot approve the Court of Federal Claims' decision below interpreting the code to achieve greater predictability. The tax code is complex, see generally Cheek v. United States, 498 U.S. 192, 199-200, 111 S. Ct. 604, 112 L. Ed. 2d 617 (1991), and we must be careful to enforce the statute as written and interpreted, see Philadelphia Reading Corp. v. United States, 944 F.2d 1063, 1074 (3d Cir.1991) (The precision and complexity of our tax laws ... bespeak Congress's intent to keep the Article III judiciary out of the administration of the tax code except to enforce and monitor the precise rules Congress has written....). It is noteworthy that some of the very cases on which Marsh relies recognize that the tax code interest provisions are not to be interpreted, contrary to their language, to require the favorable netting of overpayments and underpayments that Marsh urges is essential to tax equity. See Seeley Tube, 338 U.S. at 565-66, 70 S. Ct. 386 (declining to offset deficiency by loss carry-back for interest purposes); N. States Power Co. v. United States, 73 F.3d 764, 768 (8th Cir.1996), cert. denied, 519 U.S. 862, 117 S. Ct. 168, 136 L. Ed. 2d 110 (1996) (holding that, in calculating interest, overpayments and underpayments need not be netted to avoid interest rate differential).

We have considered the taxpayer's other arguments, and find them to be without merit.

AFFIRMED.