The "Ins" "Outs" "Overs" and "Unders" of the New Global Netting Rules
This article reviews the background of "global netting," summarizes the applicable statutory provisions and provisions of Rev. Proc. 99-19, and discusses some of the issues that may arise in the administration of the new rules and any legislative revisitation of the area.
The Internal Revenue Service Restructuring and Reform Act (IRS reform act) enacted in 1998 contained a mandate that the same interest rate apply to under- and overpayments of tax that are outstanding for the same taxpayer at the same time. Congress was responding to pressure to alleviate the whipsaws facing individual and corporate taxpayers that simultaneously owed, and were owed, money in different tax accounts. These whipsaws were exacerbated by various statutory changes made since 1986. The new provision generalized the principles, if not the precise rules, that have long applied when particular over- and underpayments are offset ("netted") against one another -- this netting is frequently referred to as the "global netting" of interest.
As usual, however, the devil lies in the details of implementation. Congress evidently intended to simplify administration of the new provision both by providing the Internal Revenue Service with some flexibility to design substantive rules within the limits of its existing management structures and computer systems and by requiring taxpayers seeking retroactive application to notify the Service within a limited "window period." Both efforts seem to have backfired badly, compounding, rather than reducing, confusion.
The deceptively simple tax code provision involves three key concepts. First, a "net interest rate of zero" is prescribed for the future on "overlapping" under- and overpayments. Second, a transition rule provides that this "global netting" may be applied retroactively, if the relevant statutes of limitation are open. Finally, this retroactivity is further conditioned on the taxpayer filing a request by December 31, 1999, that "reasonably identifies" the over- and underpayments that it wants netted. Left open were basic questions about all of these issues: how a "net interest rate of zero" will translate into interest to be assessed, allowed, or abated; what exactly are the applicable statute(s) of limitations; and whether some taxpayers that would benefit from retroactive relief will be able to make the necessary "reasonable identification" before December 31, 1999.
A hurried "technical correction" in the extenders legislation passed in late 1998 served only to spawn further confusion about the statute of limitations without providing further guidance on the other issues. On March 16, the Service released Revenue Procedure 99-19 providing some guidance and instructions for submitting requests for the retroactive application of global netting. The revenue procedure reflects a considered effort on the part of the Treasury Department and the Service to provide a practical solution to common problems (some of which are unavoidable in light of the rather clumsy statutory language) and resolves some close questions in favor of taxpayers, but it remains silent on a number of critical questions and raises some fresh ones.
The problems are hardly remarkable, given that both the original provision and the "technical correction" began as last minute legislation drafted under enormous time pressure by staffers who could not reasonably be expected to immediately grasp the finer points about how the new rules might play out in the highly technical process of resolving multiyear settlement computations. But it left the Service struggling to issue guidance that would be both workable and consistent with existing authorities concerning matters such as the statute of limitations, while taxpayers were (and still are, to an extent) left guessing about whether and how they can file a claim and the likely dollar impact of the new provision. There seems to be a fair possibility of a further legislative fix down the road, to clarify the "clarification" of the statute of limitations (which, as it stands, is likely to spawn more confusion and, eventually, litigation), and, possibly, to extend the December 31, 1999, deadline, provide expressly for "protective" identifications, or both. However, the somewhat muddled prospects for any tax legislation this season, combined with the variety of concerns that have been raised by different taxpayers and commentators, make any prediction of future congressional action hazardous.
This article reviews the background of "global netting," summarizes the statutory provisions and the revenue procedure, and discusses some of the issues that may arise in the administration of the new provision and any legislative revisitation of the area.
I. The Problem
When a taxpayer both owes money to, and is owed money by, the IRS, it is generally better -- sometimes a lot better -- not to have to pay or receive interest at all than to have to pay interest on the deficiency and receive interest on the overpayment. There are several reasons for this predicament.
A. Different Rates
Since the Tax Reform Act of 1986 (TRA 86), the basic rate of interest that the IRS pays on overpayments has been 1 percent less than it charges on underpayments. This differential will be eliminated for individuals -- but not corporations -- starting in 2000. Two other provisions -- applicable only to corporations -- can further widen this differential.
Since 1991, a "hot interest rate" -- 2 percent above the normal underpayment rate -- has applied to "large" (more than $100,000) underpayments by corporations, beginning 30 days after the taxpayer receives a "30-day letter" or an equivalent notice. Since 1995, a special rate of 1.5 percent below the normal rate applicable to overpayments (sometimes referred to as "GATT interest," named after the legislation in which it made its appearance) has applied to overpayments by corporate taxpayers exceeding $10,000.  Thus, corporate taxpayers commonly face an interest rate differential of up to 4.5 percent.
B. Other Tax Consequences
For individuals, interest received on a tax refund is naturally taxable income, while the income tax regulations prohibit the deduction of interest paid on a deficiency. Corporations do not usually face a deductibility issue, but are subject to a possible sourcing whipsaw. Refund interest will be U.S.-source income, whereas interest deductions normally must be apportioned between U.S.- and foreign-source income. If a corporate taxpayer is in a chronic excess foreign tax credit position, the portion of the deduction apportioned to foreign-source income may produce only a deferred benefit, or none at all, if it only serves to "save" foreign tax credits that are fated to expire unused. This problem is not new, but has been dramatically exacerbated because TRA '86 brought lower corporate tax rates, stricter "basket" rules under section 904(d), and a requirement to apportion interest deductions according to assets rather than gross income.
II. Offsets and Interest Tolling
The code and Service practice offer some relief for the problems identified above in the common case when income tax years involving both over- and underpayments are resolved as part of a single "audit cycle."
A. Interest Tolling Provisions
Section 6402 provides that "the Secretary...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." When an overpayment is offset against an underpayment, interest is tolled for the "overlap" period during which interest would have run on the over- and underpayment separately if they had not been offset.
Example: In 1998, the IRS determines that a corporate taxpayer has an underpayment of $1 million for its 1993 taxable year, an overpayment of $2 million for 1994, and an underpayment of $3 million for 1995. Assume that in each case, interest runs from the due date (March 15 of the following year) with no "hot interest" or retroactive global netting, and that the taxpayer pays the net amount due on December 31, 1998. The $1 million underpayment for 1993 accrues about $85,000 in interest through March 15, 1995, and $1,085,000 of the 1994 overpayment is credited to the 1993 account as of that date, leaving a net credit balance of $915,000 to accrue $79,000 in refund interest over the following year. That balance is credited against the $3 million underpayment for 1995, and thereafter only the net debit balance of $2,006,000 accrues interest. The taxpayer ultimately pays $633,000 in interest and receives $79,000, for a net cost (before tax effects) of $554,000. Without offsets, the separate balances accrue $1,335,000 in underpayment interest and $690,000 in overpayment interest, for a net out-of-pocket cost of $645,000.
The savings are far more dramatic if "hot interest" is involved or (on a tax-effected basis) if the deduction for interest paid does not translate into a current reduction in tax for one reason or another.
Of course, this example is greatly simplified. In the "real world," credits or debits to the account may not be in the same amount as the underlying deficiency or overpayment of tax (because of the need to take into account interest previously paid or allowed), and a single over- or underpayment of tax and/or interest that is determined to exist as of a particular time may comprise multiple "slices" that are taken into account for interest computation purposes as of a variety of different dates, depending on previous activity in the account and whether carrybacks are involved.
The date as of which a debit or credit "arises" or is first taken into account for interest purposes -- sometimes referred to regarding credits as the "availability date" -- is central to any interest computation. A full discussion of the associated issues is outside the scope of this article, but the basic principle is that the taxpayer owes interest while a tax liability is due and unpaid, whereas the Service owes interest while it has the use of the taxpayer's money and has not applied it against a tax or other liability, except when a code provision provides otherwise. Credits for a payment of tax and interest "arise" on payment, debits resulting from a refund are given effect as of the date of the refund, debits arising from a credit to another year are given effect as of the due date of the liability against which the credit is applied, and changes in tax liability that are the direct or indirect result of a carryback of losses or credits are given effect as of the due date for the return for the year in which the loss or credit originates. However, while these refinements might complicate its practical application, the basic principle remains, and interest is suspended on the underpayment for the period for which it would otherwise be running on the overpayment. While this means that deficiency interest will continue to run over some periods during which the running of refund interest is suspended, such as under the "45-day rule" of section 6611(e), an offset generally is fairly effective at avoiding whipsaw.
Much obviously depends on exactly what overpayment is offset against exactly what underpayment. Offsets under section 6402 are discretionary with the Service,[16 ] but in practice they are routinely made in income tax cases when multiple years are resolved at the same time. However, taxpayers may only designate the application of "voluntary" payments of tax, and the weight of authority is that offsets under section 6402 are "involuntary." This means that the precise pattern of offsets will be determined under IRS rules. For example, overpayments will ordinarily be applied to the earliest deficiency first, and will not be applied to a "transient" or "potential" deficiency that has since been eliminated by a carryback or a subsequent payment.
Example: It is determined in 1998 that a corporate taxpayer has a $1 million "general" deficiency for the taxable year 1991. A carryback is allowed from 1994 that would reduce 1991 tax by $3 million. In the simplest case, "restricted interest" will run on the resulting "potential" deficiency from March 15, 1992, until March 15, 1995. If the same taxpayer is allowed a $500,000 overpayment for 1992, the Service will not offset that overpayment against the "potential deficiency" for 1991 because as of 1998 both accounts show credit balances.
Putting these specific fact patterns aside, existing offset procedures are, in general, efficient in eliminating the running of interest on "overlapping" over- and underpayments in a typical multiple-year income tax settlement.
C. Sequential Settlements: Northern States Problem
The big problems show up when income tax liabilities for different years or for different types of tax are not part of the same audit cycle, or for one reason or another are processed separately. For example, if a taxpayer resolves its 1990-92 audit cycle in 1997, receiving an overall refund and its 1993-95 cycle in 1998, with a payment of tax plus interest, the tolling provisions are of no help, because the overpayments have not been credited against the underpayments, even though they both were outstanding and separately accruing interest over several years.
Northern States Power Co. v. United States , 73 F.3d 764 (8th Cir. 1996), illustrates the problem. The taxpayer paid asserted deficiencies for its taxable years 1980, 1981, 1983, and 1984, with interest, in 1990. Outstanding refund claims were settled in 1994, producing overpayments for 1980-84. The overpayments for 1980, 1983, and 1984 were in amounts less than the amounts of the previously determined deficiencies for the same years and, thus, they "arose" for purposes of interest computation only when the payments were made in 1990. However, the overpayments for the taxable years 1981 and 1982 that were determined in 1994 exceeded the amounts paid in 1990, meaning that a portion of these overpayments arose earlier, possibly as far back as the return due date. The taxpayer requested that these credits be transferred to the accounts for 1980, 1983, and 1984 so as to reduce the deficiency interest in those accounts, entitling the taxpayer to a refund of more of the interest that it had previously paid. The Service refused to do this because, as of 1994, all five accounts were overpaid, and the court upheld its authority. Thus, the taxpayer lost $460,000 of net interest.
Over the years, well-advised taxpayers who have a hint of the likely outcome of subsequent audit cycles have resorted to various methods to deal with this problem. One tactic is to request that a credit resulting from one audit cycle be transferred to another year as a "payment on account" under Revenue Procedure 84-58, 1984-2 C.B. 501. This procedure permits "payments" of tax and interest to be made on account of taxable years for which the Service has made a written assertion of liability for additional tax (for example, in an outstanding "30-day letter"). However, allowing a credit in this manner, rather than making a refund, remains discretionary with the Service, and in any event this tactic cannot be used when there has not yet been a written assertion of liability for any other year as of the time the overpayment is processed, because Revenue Procedure 84-58 will not recognize a remittance as a "payment on account" in these circumstances.
Taxpayers may also try to combine separate administrative cases so that they can be resolved at once, maximizing potential offsets. Again, however, the success of this tactic depends on the existence of the right facts, careful calculations, and the willingness of the Service officials with jurisdiction over the relevant cases to work with the taxpayer.
III. Global Netting: The Concept
The netting problem seems to have first hit Congress's radar screen when the interest differential was first established in TRA 86. The relevant passage in the conference report stated:
The IRS can at present net many...offsetting overpayments and underpayments. Nevertheless, the IRS will require a transition period during which to coordinate differential interest rates. The Senate amendment, therefore, provides that the Secretary of the Treasury may prescribe regulations providing for netting of tax underpayments and overpayments through the period ending three years after the date of enactment of the bill. By that date, the IRS should have implemented the most comprehensive netting procedures that are consistent with sound administrative practice.
This schedule proved a bit optimistic,  and when Congress successively widened the differential in 1990 and 1994, it was reduced to reiterating that the Secretary "should implement the most comprehensive crediting procedures under section 6402 that are consistent with sound administrative practice," adding with a trace of impatience in 1994 that it "should do so as rapidly as practicable." However, at this stage, Congress evidently envisioned more systematic and consistent offsets rather than an expansion of the netting principle to equalize interest rates even when there was no actual offset.
Strictly speaking, "global netting" -- taking into account contemporaneously existing offsetting balances regardless of whether there is any actual offset -- surfaced during consideration of the Taxpayers Bill of Rights 2 (TBOR2) provisions. The House-passed version of the Balanced Budget Act of 1995 required that Treasury conduct a study "of the manner in which the Internal Revenue Service has implemented the netting of interest on overpayments and underpayments and of the policy and administrative implications of global netting," to be completed within six months of enactment. The provision was dropped from the stripped-down conference bill, which was in any event vetoed by the president.
Nonetheless, the abortive provision served to spark action at the Service and Treasury. Following the veto, the Service issued a laundry list of "administrative initiatives to enhance taxpayer rights," including the launching of a "formal study of issues relating to the IRS's current and future interest netting procedures," followed by a notice inviting comment specifically on the legal, policy, and administrative issues implicated in global netting.
The Eighth Circuit's decision in Northern States further raised the profile of the issue. In July 1996, Congress passed TBOR2 as free-standing legislation, including the mandate for a study, which Treasury had announced its intention to complete by October 1. The report actually made its appearance in April 1997 and concluded that, practical problems aside, the Service lacked authority to put procedures in place to achieve global netting because the statutory provisions providing for interest were specific, and the Service could not perform an offset under section 6402 when there were no balances to offset. Thus, the stage was set for congressional consideration.
The earliest introduced IRS reform bills simply provided for future equalization of the rates applicable to over- and underpayments at a single rate to be determined by the Secretary. However, a revised version, introduced by the original House sponsors in collaboration with House Ways and Means Committee Chair Bill Archer, R-Texas, substituted a provision providing for prospective application of a "net interest rate of zero" to apply to overlapping underpayments and overpayments of income and self-employment taxes. The provision, which applied only to interest accruing in calendar quarters beginning after the date of enactment, was included in the bills reported out of both the House Ways and Means and Senate Finance committees.
A Senate floor amendment offered by Senate Finance Committee Chair William V. Roth Jr., R-Del., extended the application of the new provision to all types of taxes, and added an uncodified transition rule under which, "subject to any applicable statute of limitation not having expired with regard to either a tax underpayment or a tax overpayment," the net zero interest rate would be applied retroactively if the taxpayer "reasonably identifies and establishes periods of such tax overpayments and underpayments" in a request filed before December 31, 1999. The Roth amendment was adopted in conference, but either because of an oversight or because it was deemed redundant, the language concerning "any applicable statute of limitation" was dropped.
However, the language was then restored as a "technical correction" in the Tax and Trade Relief Extension Act of 1998 (TTREA), the revenue component of the omnibus appropriations legislation passed in October. The restoration first surfaced in a package of technical corrections that was folded into the Republican tax cut bill considered by the House Ways and Means Committee in September. The Joint Committee's description of the chairman's mark stated that the intent was to "clarif[y] that the statute of limitations must not have expired with regard to either a tax underpayment or overpayment for the net zero rate to apply."
Practitioners complained that this description did not comport with the statutory language referring to any applicable statute of limitations because, as explained below, under the approach that was expected to be adopted by the Service (and in fact was later adopted in Revenue Procedure 99-19), global netting directly impacts the computation of interest for only one year, and therefore only the statute of limitations for that year (whether for the refund of previously charged deficiency interest, or the allowance of additional refund interest, as the case might be) would be "applicable." The actual Ways and Means report was changed to simply reiterate the statutory language, and both the bill and report language appeared unchanged in the Senate counterpart legislation and in the stripped-down House extenders bill that ultimately became TTREA. New section 6621(d) and the effective date provision (the TTREA amendment italicized) now read as follows:
Elimination of Interest on Overlapping Periods of Tax Overpayments and Underpayments. -- To the extent that, for any period, interest is payable under subchapter A and allowable under subchapter B on equivalent underpayments and overpayments by the same taxpayer of tax imposed by this title, the net rate of interest under this section on such amounts shall be zero for such period.
Effective Dates . --
(1) In General. -- Except as provided in paragraph (2), the amendments made by this section shall apply to interest for periods beginning after the date of enactment of this Act.
(2) Special Rule. -- Subject to any applicable statute of limitation not having expired with regard to either a tax underpayment or a tax overpayment , the amendments made by this section shall apply for periods beginning before the date of the enactment of this Act if the taxpayer--
(A) reasonably identifies and establishes periods of such tax overpayments and underpayments for which the zero rate applies, and
(B) not later than December 31, 1999, requests the Secretary of the Treasury to apply section 6621(d) of the Internal Revenue Code of 1986, as added by subsection (a), to such periods.
The remainder of this article addresses the three major issues identified above concerning the implementation of this new statute.
V. 'Net Interest Rate of Zero'
The Treasury's Report to the Congress on Netting of Interest on Tax Overpayments and Underpayments (April 1997) (Treasury netting report) identified two basic approaches to the equalization of interest rates on underpayments and overpayments. The first approach, which the report referred to as the "credit/offset approach," would essentially apply the interest tolling rules to mutual indebtedness in different accounts, regardless of whether there was an actual offset under section 6402. The result would be that no interest would be either charged or paid during periods of mutual indebtedness.
Example: A corporate taxpayer was determined to have an underpayment of $5 million for 1989, which bore interest from March 15, 1990. The taxpayer paid the $5 million tax plus $3,020,000 of interest as of July 31, 1995. No "hot interest" was involved. On December 31, 1998, an overpayment of $3 million is determined for 1991, and would bear interest from March 15, 1992.
Under the "credit/offset" approach, the interest in the 1989 account would be recomputed as if the overpayment for 1991 had been offset against the underpayment for 1989 as of March 15, 1992. The taxpayer would be entitled to a refund of approximately $905,000 of the $3,020,000 interest it paid on July 31, 1995.
The $3 million overpayment in the 1991 account would not bear interest for the period from March 15, 1992, through July 31, 1995, while it was treated as though it was offset against the 1989 underpayment. Thereafter, interest would resume.
The $905,000 refund of previously paid interest for 1989 and the 1991 overpayment of $3 million would accrue a total of approximately $920,000 in refund interest after July 31, 1995. Consequently, the taxpayer would receive (in addition to its $3 million overpayment), a return of $905,000 in interest previously paid, plus $920,000 in refund interest, for a total of $1,825,000, compared to the roughly $1,675,000 in refund interest that would have accrued on the 1991 overpayment in the absence of global netting. The difference is attributable to the opportunity to earn a return at the deficiency rate, rather than the refund rate, during the period of overlapping indebtedness (March 15, 1992, to July 31, 1995).
The second possible approach, referred to as the "interest equalization approach," is based on the methodology in Revenue Procedure 94-60, which addresses the situation in which a taxpayer has received a refund with interest and an underpayment for the same year is subsequently determined to have existed in the same account for a portion of the period during which interest ran on the refund. In these cases, the Service charges interest for the period of overlap only at the refund rate on the portion of the deficiency that does not exceed the refund.
Example: On April 30, 1993, a taxpayer receives a refund of $500,000 for its tax year 1989, with $150,000 in interest from March 15, 1990. It is later determined that the taxpayer underpaid tax on its original return by $300,000; thus, there is a $800,000 underpayment after allowing for the previous refund. Interest for the period from March 15, 1990, through April 30, 1993, on the first $500,000 of the underpayment is charged only at the rate at which interest was previously allowed on the refund.
Global netting under the "interest equalization approach" would apply the same principle when the overlapping overpayment and underpayment are in different tax accounts. To the extent, and for the duration, of the overlapping balances, either the rate charged on the underpayment would be limited to the rate paid on the overpayment or the rate payable on the overpayment would be increased to reflect the rate charged on the underpayment.
Example: Assume the same facts as in the example of the credit/offset approach above. On July 31, 1995, the taxpayer pays $5 million in tax and $3,020,000 in interest computed from March 15, 1990. On December 31, 1998, an overpayment of $3 million is determined for its tax year 1991, with interest running from March 15, 1992.
Under the "interest equalization" approach, there would be no recomputation of interest for tax year 1989. Interest would accrue on the 1991 overpayment at the underpayment rate for the period between March 15, 1992, and July 31, 1995, and thereafter at the normal overpayment rate. The total refund interest accrued as of December 31, 1998, would be approximately $1,825,000. With minor breakage aside,  this method produces exactly the same result as the "credit/offset approach" on a pretax basis; however, the tax effects could differ, because in the one instance the taxpayer is receiving a refund of $905,000 in deficiency interest that it previously paid, whereas in the other it is receiving the same $905,000 as statutory (refund) interest.
The "credit/offset approach" has the advantage of replicating most closely the effects of an actual offset and, thus, largely eliminating the incentives for taxpayers to play procedural games to ensure that under- and overpayment years are resolved at the same time. The "interest equalization approach," while clearly better than nothing, will not be as beneficial as the "credit/offset approach" for taxpayers for whom the receipt of refund interest and the payment of deficiency interest in the same amount do not add up to a "wash" because of deductibility, sourcing, or other considerations. Therefore, if the "interest equalization approach" is adopted, sophisticated taxpayers will continue to strive for "real" offsets under section 6402, thus ensuring that they would neither receive nor pay any interest at all.
On the other hand, the "interest equalization approach" has the signal advantage of affecting only the interest computation for the particular year under consideration (sometimes referred to as the "current year"), whereas the "credit/offset approach" could require recomputing interest for all other years ("previously determined years") involved in the computation. Thus, the choice between approaches has important implications for the statute of limitations, as discussed below.
The expectation of the drafters of the reform act seems to have been that the "interest equalization" approach would be adopted, although legislative history makes it clear that "the Secretary may use other procedures or methodologies that he deems appropriate, so long as a net zero interest rate is achieved." 
The word circulating on Capitol Hill during consideration of TTREA in September and October was that Treasury and the Service had indeed tentatively decided in favor of an "interest equalization" approach, with the interest rate actually applied being determined on the basis of which account (over- or underpayment) had been resolved first. In other words, if a taxpayer paid tax and interest first, the rate on a subsequently determined overpayment would be increased to match the rate of interest that the taxpayer had paid. If, on the other hand, a refund with interest was allowed first, the rate on a later determined underpayment would be reduced to correspond to the rate the Service had paid on the refund.
Revenue Procedure 99-19 provides for an "interest equalization" approach, but it is evidently confined to abatement and the refund of interest charged on the deficiency years.  The Service has reportedly suggested informally that additional overpayment interest might be allowed instead if the taxpayer can demonstrate a reason, but there are no procedures in place to request or allow this treatment.
VI. Statute of Limitations
A. Basic Approaches
Congress's desire to allow flexibility in crafting the mechanics and the uncertainty surrounding the precise approach to be adopted complicate consideration of statute of limitations issues, and the supposed technical correction has only further muddied the waters. A threshold issue is that the Delphic language ("[s]ubject to any applicable statute of limitation not having expired with regard to either a tax underpayment or a tax overpayment") can be read to support any of three possible basic approaches to the statute of limitations.
It could be read to mean simply that any refund claims relating to global netting must be timely filed under the normal rules applicable to an underpayment or overpayment, as the case may be, for the year in question (a "single statute of limitations"). If so, however, it is arguably surplus, as the normal rule is that retroactive legislation will not reopen an otherwise barred year, unless it expressly so provides. The language can also be read to mean that the statute of limitations must be "open" (whatever that means) for all underpayments and overpayments taken into account in the global netting computation (a "dual statute of limitations"), although as discussed below that interpretation opens a Pandora's box of further questions. This view was evidently that taken by Treasury during drafting, and it is now enshrined in Revenue Procedure 99-19.
A close perusal of the origins of the troublesome phrase suggests one possible interpretation that lies somewhere in between the two possibilities identified above. The discussion of the statute of limitations in the Treasury netting report appears to assume the possible application of the "credit/offset approach." The report gives an example of the impact of the statute of limitations in which a taxpayer pays a deficiency for 1990 and then, after the statute of limitations for refund claims has run on 1990, is determined to have an overpayment for 1991; the report also notes that the Service is legally prohibited from refunding any excess interest paid for 1990. Citing concern for finality and administrability, the report recommends that the Service not be required to reopen years that would not be open for any other purpose and specifically recommends against a proposal that the legislation should reopen statutes of limitation to permit global netting to apply to any underpayments and overpayments outstanding after the advent of differential rates in 1986.
If the writers were assuming application of the credit/offset approach, the reference may have reflected a lurking concern about possible whipsaws. For instance, a taxpayer might request a refund of interest for an open underpayment year, while the Service was barred from recapturing refund interest it had already paid for the overpayment year(s). This type of case would not fall under the mitigation provisions, and while certain judicial doctrines (such as equitable recoupment might be invoked to fill in the gap, the contours of these doctrines are somewhat vague and disputes could be expected. In this context, requiring that the applicable statute of limitations remain open for any additional assessment or refund -- for any year -- that is required as a result of a global netting computation would be workable and make perfect sense, and in the context of a "credit/offset" approach would mean a "dual" statute of limitations.
However, to the extent the "interest equalization approach" is employed, the proper computation of interest for only one year -- the "current year" -- is at stake. There can be no refund or additional assessment of interest for any of the "previously determined years," and the whipsaw potential disappears. Thus, it would follow that the only "applicable statute of limitations" would be that for the year for which interest was actually going to be recomputed.
Further compounding the confusion is the unexplained tweaking of the statutory language in the posthumous explanation found in the congressional joint committee's "blue book." The relevant passage reads: "A statute of limitations must not have expired as of the date of enactment with respect to both the underpayment and overpayment for this provision to apply." Dropping the word "applicable" and substituting "the" for "a," arguably shades the statutory language in favor of a "dual statute" reading. Also, the basis for concluding that the critical date is the date of enactment, rather than the date a taxpayer files a claim, is unclear.
Treasury and the Service were evidently convinced that congressional intent from the outset had been to impose a "dual" statute of limitations, and Revenue Procedure 99-19 explicitly provides as much. On the other hand, however, the revenue procedure follows the blue book in only reading the statute to require that both statutes have been open as of July 22, 1998, rather than as of the date the claim was filed.
B. Choice Among Statutes of Limitations
Having opted in favor of a "dual statute of limitations," the drafters of Revenue Procedure 99-19 faced another problem. Accepting arguendo that Congress meant to superimpose on the normal statute of limitations a requirement that any other year taken into account in a global netting computation be "open" in some sense, even if the actual computation of tax and interest for that year did not change one penny, the provision as written is fundamentally unworkable because the term "applicable statute of limitations" is meaningless without reference to a specific claim or assessment. If the previously determined year must be "open," open for what type of claim? And against whom, the taxpayer or the Service?
A smorgasbord of limitations provisions can potentially apply to different types of claims, or to a given type of claim on different facts. The basic statutory provisions are familiar. In general, the Service has three years (extendable to six years, or indefinitely, in some circumstances) from the filing of an income tax return to assess additional tax. Assuming a timely filed return, the taxpayer may claim a refund within three years of filing, but may file a claim at any time thereafter to the extent of any payment of tax or interest made within two years. A series of special rules apply to claims relating to particular issues and to situations involving carrybacks. In general, claims relating to the carryback of losses and credits (other than foreign tax credits) are timely so long as the source year of the loss or credit that ultimately produced the overpayment remains open. A 10-year statute applies to refund claims relating to foreign tax credits, /61/ while various different rules apply to the Service.
Because the tax liability for a given year is unitary, the taxpayer can offset a proposed assessment, or the Service a refund claim, by raising any new issue that goes to the proper amount of tax due for the year, and the other party may counter with "offsets of offsets," even if the statute of limitations has otherwise expired. Moreover, just as the taxpayer may file a claim, even if the general statute has expired, to the extent of any amount paid within two years, if the Service has allowed a refund, it can proceed for a "wrongful refund" within two years, even if the statute would not otherwise be open for a fresh assessment. Consequently, in many circumstances a tax year may be "open," but recovery may be limited to a given amount.
Two additional considerations are peculiar to interest claims. First, while the provisions cited above control the Service's ability to assess additional tax, if the Service has timely assessed tax, it can assess additional interest relating to that tax for as long as the 10-year statute of limitations on collection remains open. Second, while a taxpayer's claim that it overpaid deficiency interest is a claim for refund or credit of an overpayment subject to the ordinary three-year and two-year rules (unless some specialized provision applies), a claim that insufficient interest has been allowed on a refund (such as a taxpayer's claim that a higher rate should have been applied under the netting rules) is a general claim against the United States under the Tucker Act as to which the taxpayer has six years to file suit after the overpayment is allowed.
It is easier to appreciate the problems in implementing a dual statute of limitations by thinking about a real-life case. Suppose the "previously determined year" is an underpayment year for which the taxpayer paid interest computed at the deficiency rate, and it now seeks to increase refund interest payable in the "current year" to reflect that rate. The applicable statute for the "current year" is presumably the six-year statute generally applicable to claims for the allowance of additional refund interest. But what about the previously determined year? Suppose the general statute of limitations on refund claims has run, but a refund claim would be timely to the extent of a payment made within two years? How is the taxpayer's hypothetical "claim" to be measured? Is it based on the refund claim that the taxpayer would have for overpaid deficiency interest for the "previously determined year" if the credit/offset approach were employed instead of the interest equalization approach? What if the general statute of limitations for a fresh refund claim is closed, but the previous assessment of tax and interest reflected a recapture of a carryback and the statute of limitations would still be open for claims relating to the carryback?
On the other hand, suppose that a refund was allowed with interest in the previously determined year, and the taxpayer now seeks to reduce interest it would otherwise have to pay (or has paid) on a "current year" deficiency. If a refund claim for the "current year" is involved, its timeliness is presumably determined under the ordinary rules, but what determines whether the previously determined year is "open"? What if the Service is barred from assessing additional tax, but because the 10-year statute of limitations on collection is open, it can assess additional interest? What if the refund was allowed within the past two years, so that if the credit/offset approach were applied, the Service could recapture any theoretical overpayment of interest as a wrongful refund? What if the general statute of limitations is barred but the overpayment resulted from a carryback, and the source year of the carryback is still open?
It is tempting to say that these issues could be resolved by insisting that the general statute of limitations (for additional assessments, or a claim for refund, as the case might be) must not have run for either year; but this would lead to absurd (and grossly unfair) results. Suppose an overpayment was allowed, and a refund paid, after the general statute of limitations for the year had run, because a refund claim was lodged while the statute was still open, or because it stemmed from a foreign tax credit, or a carryback from a still open year? Adoption of this rule would mean that this type of overpayment could never be taken into account in any global netting computation. The same would apply to an underpayment that, because of one or another special rule, was assessed and paid after the general statute of limitations had expired.
Administrative guidance cannot conclusively resolve all these problems. As the Treasury netting report concluded, the Service does not have the authority to waive the statute of limitations provisions in the code. On the other hand, if the IRS refuses to process a claim filed within an arguably applicable statute of limitations, sooner or later some court will be presented with the task of construing a garbled statute with an ambiguous legislative history. The statutes of limitations applicable to interest claims, especially when there have been offsets between years, are complex enough without piling on consideration of hypothetical claims that will never be filed, and how a court might construe the phrase "any applicable statute of limitations" in a given set of circumstances is a matter for speculation.
Revenue Procedure 99-19, in this writer's view, made the best out of a bad job by providing that the applicable statute of limitations for overpayments was the Tucker Act statute applicable to claims for additional refund interest (six years from the date of allowance of the overpayment) and that for underpayments was the normal three years' or two years' statute that would apply to ordinary claims for overpaid interest. Some fuzziness remains, however.
For example, it is a little unclear how the Service proposes to handle the statute of limitations for the underpayment year in light of the "unitary" nature of the liability for a given taxable year. An overpayment is "the payment of more than is rightfully due" for a given year, whatever the reason. Any tax or interest paid is, thus, potentially an "overpayment" if in excess of the total correct liability for the year. Normally, as discussed above, even if the general statute of limitations for refund claims has run, a claim for refund within two years of the date of a payment (of tax or interest) is, in general, timely to the extent of that payment. Such a claim need not relate to the Service's grounds for assessing and collecting that particular payment, but (estoppel considerations and closing agreements, etc., aside) may be based on any grounds relating to the liability for the taxable year. On the other hand, the taxpayer has the burden of establishing that the tax for the year has been overpaid, so its recovery on a refund claim filed on one ground may be reduced or eliminated because of unrelated issues raised by the Service.
Revenue Procedure 99-19 states the alternative statute of limitations for overpayments of deficiency interest as "two years from the time the interest was paid." It is unclear whether, by this language, the Service is trying to imply that (if the general statute of limitations has run) a claim is only timely to the extent that amounts designated as interest for that particular underpayment have been paid within two years. The question can arise in any year in which there are two or more underpayments that were assessed and paid (with appropriate interest) for a given year, and some fall within the two-year window and some outside of it.
Example: In 1995, a taxpayer is assessed and pays a deficiency of $3 million in tax, plus $2 million of interest for a particular year. In 1998, it is assessed and pays a further $1 million in tax and $1 million in interest for the same year.  By 1999, the general statute of limitations for further assessments for that year has expired. Can the taxpayer invoke "global netting" only for the 1998 assessment, because that is the only one made within two years? Or will it be able to use all underpayments determined for that year in the computation, even though its total recovery naturally would be limited to the $2 million paid within two years?
The Service might be hard-pressed to justify the former interpretation in court, given 60 or so years of case law that consistently regards the ultimate issue as being the determination of whether there is an "overall" underpayment or overpayment for the year. Even if it were to succeed, the IRS would face a mind-boggling array of subsidiary issues in any case when it interposed unrelated "offset" issues to reduce a recovery.
C. Potential Legislative Solutions
Congress may yet see fit to enact a technical correction to the technical correction, maybe even under somewhat less time pressure. Ideally, this sort of a correction would avert the need to consider the parade of imponderables described above by clarifying that for a global-netting claim to be timely, any refund or reassessment of interest must be timely under the ordinary limitations provisions for the year(s) concerned -- in short, a single statute of limitations. If the goal were to avoid any possibility of whipsaw, a single statute of limitations could be backed up by a special rule, akin to the present mitigation provisions, providing that if the taxpayer requests a recomputation of interest for an open year under global netting, the statute of limitations for any other year would be reopened for the limited purpose of making any additional assessments required by the same computation. This rule could, like the mitigation provisions, prohibit unrelated offsets, so as to preserve finality for otherwise closed years.
An alternative would be to provide some "bright-line" rule, for example, that if the general statutes of limitation (for additional assessments or refund claims) were not open for a given year, only assessments made, or refunds allowed, within a fixed period (say, six years) of the taxpayer's lodging of its claim would be taken into account. This sort of rule could not avoid the additional complexity inherent in a specialized statute of limitations applicable to particular types of claims, but at least it would be fairly easy to understand and apply. However, at this point, the prospects for any action remain uncertain.
VII. 'Retroactive' Claims
An independent cloud of uncertainty attends the question of precisely what a taxpayer must file by December 31, 1999, to protect its rights to retroactive application of global netting. This "cliff problem" arises because of the strange manner in which the transition rule was drafted. Absent the transition rule, the provision is effective only for interest accruing after the date of enactment, regardless of the taxable year concerned. However, the transition rule permits unlimited retroactive application -- as long as the relevant year(s) are "open," whatever that is eventually determined to mean -- provided that a request is filed before December 31, 1999.
Example: Taxpayer A has an underpayment for 1985 and an overpayment for 1986 determined in 1999, which are not credited against one another, and applies for global netting. The interest rates on the under- and overpayment will be equalized from the advent of differential rates as of January 1, 1987. Taxpayer B finds itself in the same position in 2000, but it cannot apply for global netting under the transition rule. The interest rates will be equalized only as of July 22, 1998, the date of enactment of the IRS reform act; until then, interest will accrue on the overpayment and the underpayment at rates that may differ by up to 4.5 percent.
It is unclear why this additional hurdle was added, except perhaps for ease in revenue estimation, but the stakes are potentially large. If the underpayments and overpayments in the above example were each for $10 million, the retroactive application of global netting could mean the difference between a net liability (before tax effects) of $12.3 million versus $3.4 million as of December 31, 1998. The problem with the deadline has not been the subject of as much public comment as the statute of limitations issue, but it can be expected to come into sharper focus as the year progresses.
The potential retroactive application of global netting will remain a factor in the resolution of all tax years through at least 1997 (and through some taxpayers' 1998 fiscal years). As of 1999, taxpayers will not necessarily even know whether they will be audited for their later tax years, much less which of those years may be resolved together or, given carrybacks and the host of timing issues that tend to crop up in large audits, whether the end result of a given year or multiple-year audit "cycle" will be an underpayment or an overpayment. The Service frequently lags far behind in complex audits, and a significant number of taxpayers in the coordinated examination program are at the administrative appeals level, or even still under audit, for taxable years in the 1980s. Court proceedings can drag the process out even longer. It is not uncommon to find taxpayers with (allowing for carrybacks) 15 or even 20 years' tax liability "in play." How are these taxpayers to ensure the application of "global netting" to under- and overpayments that may be assessed or allowed five or more years from now? Even if the taxpayer has received a refund of an overpayment, or paid an underpayment, which would be eligible for a global-netting computation, it is currently unclear whether the taxpayer can identify a year, or a range of years, in which it expects an offsetting under- or overpayment so as to preserve its rights to the eventual application of global netting.
Revenue Procedure 99-19 basically punts the question, requesting written comments "regarding the level of specificity necessary to reasonably identify and establish on or before December 31, 1999, the period(s) for which an equivalent amount of overpayment and underpayment of tax overlap when the taxpayer cannot provide by December 31, 1999, a final computation of how the net interest rate of zero applies."
The transition rules can be expected to affect taxpayers' strategies for managing audits if a substantial amount of interest is potentially at stake. If a taxpayer has eligible overpayments but not underpayments, one possibility might be to make payments on account of proposed deficiencies for years still under audit or at Appeals. In the reverse case, a taxpayer is largely at the mercy of the Service's audit process (and, in appropriate cases, the congressional joint tax committee) insofar as whether an overpayment will be allowed before December 31, 1999. Some taxpayers might try to force a resolution of their administrative cases by refusing extensions of the statute of limitations, or bypassing the appeals level; others might try to enter into partial settlements with the examining district or Appeals -- if necessary reserving the right to file refund claims later -- or to make use of the accelerated issue resolution program if they are eligible. And it is probably a safe bet that as 1999 draws to a close, the Service will be bombarded with a spate of "protective" applications under the transition rule, some by taxpayers that will never have occasion to make use of global netting.
While Congress may revisit some of these issues next year -- if comprehensive tax legislation passes that is not vetoed by the president -- this is hardly a sure thing, and the legislative picture will likely not become clear until near the end of the year. As most possible strategies for maximizing the potential from global netting will take time to develop and implement, taxpayers should be reviewing their options now.
 Internal Revenue Service Restructuring and Reform Act of 1998, Pub. L. No. 105-206.
 IRS reform act section 3301, codified at section 6621(d). Except as specified, all section references are to the Internal Revenue Code of 1986, as amended , and the Treasury regulations thereunder.
 IRS reform act section 3301(a), codified at section 6621(d); id. section 3301(c)(2), as amended by Tax and Trade Relief Extension Act of 1998 (TTREA), Pub. L. No. 105-277, section 4002(d).
 TTREA section 4002(d), supra .
 Rev. Proc. 99-19, 1999-13 I.R.B. 10.
 Section 6621(a).
 IRS reform act section 3302(a), amending section 6621(a)(1)(B).
 Section 6621(c), added by Omnibus Reconciliation Act of 1990, Pub. L. No. 101-508, section 11341(a).
 Section 6621(a)(1), as amended by Uruguay Round Agreements Act, Pub. L. No. 103-465, section 713(a); and the Taxpayer Relief Act of 1997 (TRA 97), Pub. L. No. 105-34, section 1604(b)(1). Unlike "hot interest," this lower "GATT" rate applies for the entire life of the overpayment, but only to the portion exceeding US $10,000.
 Reg. section 1.163-9T(b)(2)(i)(A). The Tax Court in Redlark v. Commissioner, 106 T.C. 31, Doc 96-1488 (71 pages), (1996) rev'd and rem'd 141 F.3d 936 Doc 98-12692 (11 pages), (9th Cir. 1998), invalidated the regulation insofar as it applies to interest on deficiencies attributable to an individual's trade or business, but the Ninth Circuit on appeal, and the Eighth Circuit in Miller v. United States , 65 F.3d 687 (8th Cir. 1995), upheld it. Full treatment of this controversy is outside the scope of this article.
 Section 6601(f) provides, "If any portion of a tax is satisfied by credit of an overpayment, then no interest shall be imposed...for any period during which, if the credit had not been made, interest would have been allowable with respect to such overpayment," while section 6611(b)(1) provides that a credit shall cease to bear interest as of "the due date of the amount against which the credit is taken."
 "Debit" is used as shorthand for the creation or increase of an underpayment, or a reduction of an overpayment, whether due to an increase in tax liability or a refund or credit to another account, while "credit" refers to the creation or increase of an overpayment, or a reduction in an underpayment, due to a decrease in liability or a payment or credit for another year.
 Thus, if an overpayment per return or "quickie" carryback under section 6411 is refunded without interest, the corresponding portion of a subsequently determined underpayment bears interest only from the date of the actual refund. Rev. Rul. 88-98, 1988-2 C.B. 356 (situation 3). If the refund is with interest, the total amount is, in effect, taken into account on the date from which interest started running. See Rev. Proc. 94-60, 1994-2 C.B. 774 (technically, the Service "charges" the same interest previously allowed for the period(s) and balance(s) involved).
 E.g., an estimated tax payment. Avon Products, Inc. v. United States, 588 F.2d 342 (2d Cir 1978), and Rev. Rul. 88-98, supra (situations 1 and 2). Disputes persist about what payment the credit is applied to -- compare Rev. Rul. 88-98 with, e.g., Sequa Corporation v. United States (S.D.N.Y. 1998).
 Sections 6601(d) and 6611(f), as amended by TRA 97, section 1055(b). Before TRA 97, there was no express provision for foreign tax credits, which created controversy. Compare Fluor Corp. v. United States, 126 F.3d 1397, Doc 97-28980 (23 pages) (Fed. Cir. 1997) with Intel Corp. v. Commissioner, 111 T.C. 90) Doc 98-24174 (22 pages) (1998).
 See, e.g., Northern States Power Co. v. United States, 73 F.3d 764, 767 Doc 96-905 (12 pages) (8th Cir. 1996) (discussed below); Acker v. United States, 519 F. Supp. 178, 182 (N.D. Ohio 1981); Estate of Bender v. Commissioner, 86 T.C. 770, 778 (1986), aff'd and rev'd on other issues, 827 F.2d 884 (3d Cir. 1987); and Mounts v. United States (S.D. W.Va. 1995).
 See, e.g., Bender, 827 F.2d at 887-88; for a case in which the Service overlooked an offset, to the taxpayer's distress, see Winn-Dixie Stores, Inc. v. Commissioner, 110 T.C. No. 23, Doc 98- 13423 (10 pages) (1998).
 Rev. Rul. 73-305, 1973-2 C.B. 43, modified by Rev. Rul. 79- 284, 1979-2 C.B. 83; see generally In re Energy Resources Co., Inc., 871 F.2d 223, 227-30 (1st Cir. 1989), aff'd sub nom. In re U.S. v. Energy Resources Co., Inc. , 495 U.S. 545 (1990), and authorities cited.
 Ryan v. United States , 64 F.3d 1516, 1522-24, (Doc 95-9152 (9 pages)) (11th Cir. 1995); Kalb v. United States, 505 F.2d 506, 509-10 (2d Cir. 1974); but cf. Jung v. United States , 701 F. Supp. 175 (E.D. Wis. 1988) (suggesting taxpayer might be able to direct the application of an unrefunded overpayment before the Service had determined to make an offset under section 6402).
 See generally Estate of Bender, supra note 16, 827 F.2d at 887 n.7, citing IRS Internal Revenue Manual (IRM) 30(55)9, and Treasury's netting report, App. 4, excerpting IRM 31(59)(31). The IRS's internal procedures governing the computation of interest were to be collected at IRM(22)000 et seq., see Treasury netting report at 37 n.8, but this appears not to be publicly available.
 A "general" deficiency is a deficiency of tax before considering the effect of carrybacks (or prior allowances thereof). See IRS Form 2285.
 See LTR 9739003, in which this principle benefited the taxpayer because the Service was barred from assessing the additional restricted interest. However, in normal circumstances, the rule can prevent taxpayers from benefiting from an earlier offset.
 H.R. Rep. No. 841, 99th Cong., 2d Sess. (vol. II) 785 (1986). The Senate report includes similar language -- see S. Rep. No. 313, 99th Cong., 2d Sess. 185 (1986).
 The referenced transition rule provided, "The Secretary...may issue regulations to coordinate section 6621...with section 6601(f) [the interest tolling rule]. Such regulations shall not apply to any period after the date 3 years after the date of the enactment of this Act." TRA '86, Pub. L. No. 99-514, section 1511(b), 100 Stat. 2085, 2744 (1986).
 H.R. Rep. No. 964, 101st Cong., 1st Sess. 1101 (1990) (conference report on OBRA 1990); 136 Cong. Rec. S15629, 15712 (Oct. 18, 1990) (Senate report); H.R. Rep. No. 826, 103d Cong., 2d Sess. (vol. I) 178 (1994) (House Ways & Means report on Uruguay Round Agreements Act); S. Rep. No. 412 (103d Cong., 2d Sess.) (joint report of the Senate committees with jurisdiction).
 See Treasury netting report supra at 27-28; Northern States, note 6 supra , 73 F.3d at 767-68.
 H.R. 2491 (104th Cong., 1st Sess.) section 13364(a)(1), reprinted in H.R. Rep. No. 280, 104th Cong., 1st Sess. (vol. I) 28 (1995) (Budget Committee report).
 See H.R. Rep. No. 350 (104th Cong., 1st Sess.) 1398-99 (1995).
 Ann. 96-5, 1996-4 I.R.B. 99, 101.
 Notice 96-18, 1996-1 C.B. 370.
 TBOR2, Pub. L. No. 104-168, section 1208.
 See H.R. Rep. No. 506 (104th Cong., 2d Sess.) 50.
 H.R. 2292, 105th Cong., 1st Sess. section 309 (July 30, 1997); S. 1096, 105th Cong., 1st Sess. section 309 (July 31, 1997).
 H.R. 2676, 105th Cong., 1st Sess. section 331 (Oct. 21, 1997).
 H.R. 2676, 105th Cong., 1st Sess. section 331 (Oct. 31, 1997); H.R. Rep. No. 364, 105th Cong., 1st Sess. (pt. 1) 21, 63-64 (1997); H.R. 2676, 105th Cong., 2d Sess. section 3301 (Apr. 22, 1998); and S. Rep. No. 174, 105th Cong., 2d Sess. 61-62 (1998).
 Amendment No. 2383, offered by Roth (for Finance Committee members Bob Graham, D-Fla., Don Nickles, R-Okla., and Carol Moseley- Braun, D-Ill. The text appears at 144 Cong. Rec. S4546 (May 7, 1998).
 Pub. L. No. 105-277, Omnibus Consolidated and Emergency Supplemental Appropriations Act, 1999, 112 Stat. 685 (1998).
 JCX-61-98, Description of Tax Technical Corrections in the "Taxpayer Relief Act of 1998" 3 (Jt. Comm., Sept. 15, 1998).
 H.R. Rep. No. 739, 105th Cong., 2d Sess. 94 (1998).
 S. 2622, 105th Cong., 2d Sess. section 402(d) (referred to the Senate Finance Committee but never reported); JCX-70R-98, Description of Provisions in S. 2622, The Tax Relief Extension Act of 1998 28 (Jt. Comm., Oct. 10, 1998).
 H.R. 4738, 105th Cong., 2d Sess. section 402(d); JCX-67-98, Description of Tax Technical Corrections Contained in H.R. 4738 3 (Jt. Comm., Oct. 9, 1998); H.R. Rep. No. 817, 105th Cong., 2d Sess. 64 (1998) (Ways & Means report).
 H.R. 4328, 105th Cong., 2d Sess. section 4002(d),
 IRS reform act section 3301(a), codified at section 6621(d).
 IRS reform act section 3301(c), as amended by TTREA section 4002(d).
 See generally Treasury netting report, supra , at 28-33.
 The breakage is attributable to the "first $10,000" exclusion from the lower GATT overpayment rate.
 H.R. Rep. No. 364 (pt. 1), 105th Cong., 1st Sess. 64 and n.41 (1997).
 Rev. Proc. 99-19, section 4.04(5)(f) requires that the taxpayer submit a computation arrived at "by applying section 6621(d) to an underpayment year to reduce the taxpayer's liability for underpayment interest."
 See, e.g., U.S. v. Zacks , 375 U.S. 59 (1963).
 See "Conflict Looms Over Open-Year Netting Issue as Treasury Works to Implement New Law," DTR , Nov. 4, 1998, at G-2, quoting then Attorney-Advisor Chris Rizek to the effect that Congress was merely "restating the rule" that both underpayment and overpayment years had to be open.
 Rev. Proc. 99-19, section 3.01(1).
 Treasury netting report 34, 42.
 Section 1311 et seq .
 See, e.g., United States v. Dalm, 494 U.S. 596 (1990); and Bull v. United States , 295 U.S. 247 (1935).
 JCS-6-98, General Explanation of Tax Legislation Enacted in 1998 74 (Joint Comm., Nov. 24, 1998).
 See note 50, supra .
 Rev. Proc. 99-19, section 3.01(1).
 Section 6501(a).
 Section 6511(a) and (b).
 Section 6511(d)(2) and (4).
 Section 6511(d)(3). A 1997 amendment clarifies that the 10- year period runs from the return due date for the source year of the credit; the U.S. Court of Federal Claims had previously held the relevant year was the one in which the credit was used. Ampex Corp. v. United States, 620 F.2d 853 (Cl. Ct. 1980); compare Rev. Rul. 84- 125, 1984-2 C.B. 125. The rules applicable to "junior" credits displaced by foreign tax credits remain murky. In Ampex, the parties and the court alike assumed claims for displaced investment tax credit would be timely if the foreign tax credit claims were timely, 620 F.2d at 857 n.8, but see LTR 8727006, holding on similar facts that the source year of the investment tax credit would have to be open to permit such a claim.
 The Service has an extra year beyond the (ordinary) statute period to make assessments relating to carrybacks. Section 6501(i). There is no statute of limitations at all for assessments under section 905(c) stemming from a refund of foreign tax. See Rev. Rul. 72-525, 1972-2 C.B. 443, clarified by Rev. Rul. 83-80, 1983-1 C.B. 130, and authorities cited.
 Lewis v. Reynolds , 284 U.S. 281 (1932).
 E.g., Charter Co. v. United States, 971 F.2d 1576 (11th Cir. 1992); Union Pacific R.R. Co. v. United States , 389 F.2d 437, 447-48 (Cl. Ct. 1968).
 Section 6511(a), (b).
 Section 7405.
 Sections 6502, 6601(g); see, e.g., Marathon Oil v. United States , 1998 WL 790663 (Fed. Cir. 1998).
 E.g., Alexander Proudfoot Co. v. United States , 454 F.2d 1379 (Cl. Ct. 1972).
 28 U.S.C. section 1491. E.g., Colgate-Palmolive-Peet Co. v. United States, 58 F.2d 499 (Cl. Ct. 1932); Rev. Rul. 56-506, 1956-2 C.B. 959.
 28 U.S.C. sections 2401, 2501.
 Treasury netting report, supra , at 34.
 The Service is arguably entitled to reverse inter-year offsets at any point while the statute of limitations for collection remains open on an underpayment against which the offset is made. See Commissioner v. Newport Industries, Inc., 121 F.2d 655 (7th Cir. 1941); see also discussion in Fine v. Commissioner , 70 T.C. 684, 689- 90 (1978). This situation can leave a taxpayer who was disadvantaged by an error in an interest computation involving complex inter-year offsets in some doubt as to whether it has a claim for overpaid deficiency interest, or underpaid refund interest, and if so for what year. On a hypertechnical reading of the statute, the reversal of a credit and its application to another year might even leave the taxpayer with a claim that is barred as soon as it arises. See sections 6407 and 7422(d). Thankfully, such a case does not ever seem to have been presented to a court.
 Jones v. Liberty Glass Co. , 332 U.S. 524, 531 (1947).
 Deficiency interest charged and paid forms part of the principal amount of the overpayment. Reg. section 301.6611-1(c) (Ex. 2) (last sentence); see also Estate of Baumgardner v. Commissioner , 85 T.C. 445 (1985) (reviewed) (interest part of the overpayment for Tax Court jurisdictional purposes). Cf. section 6601(e).
 E.g., Lewis v. Reynolds , 284 U.S. 281 (1932).
 Rev. Proc. 99-19, section 2.04(1).
 The example assumes the existence of overpayments for other years for overlapping periods, as to which the "applicable" (six-year) statute of limitations remains open.
 Sections 1311-1314.
 See section 1314(c).
 Rev. Proc. 99-19, section 6.
 Taxpayers can, in general, make payments on an account once a deficiency has been proposed in writing for a given year. See Rev. Proc. 84-58, 1984-2 C.B. 501.
 See Rev. Proc. 94-67, 1994-2 C.B. 800.
This article first appeared in Tax Notes, April 1999, p.543, and is reprinted with permission.