A Question of Identity – Interest Netting, Part 1

Why interest netting?

With some exceptions, the government charges taxpayers interest on unpaid tax liabilities and pays taxpayers interest on refunds. There were some rate differences in the earliest years of the modern income tax, but for a long period of time the government used the same interest rate for both underpayments and overpayments. That’s still the case for non-corporate taxpayers but changed for corporate taxpayers over 30 years ago.

In the Tax Reform Act of 1986, Congress created a 1% difference in Section 6621(a) between the base rate it charges corporations on underpayments of tax (the federal short-term rate + 3%) and the base rate it pays corporations on overpayment of tax (the federal short-term rate + 2%). As a result of subsequent changes, the gap can be much larger. The Omnibus Budget Reconciliation Act of 1990 added Section 6621(c), establishing the “large corporate underpayment” or “hot interest” rate (the federal short-term rate + 5%) for underpayments exceeding $100,000. The Uruguay Round Agreements Act of 1994 added the flush language at the end of Section 6621(a)(1), establishing the “GATT interest” rate (the federal short-term rate 0.5%) for overpayments exceeding $10,000. The proper application of hot interest and GATT interest is itself a complex issue beyond the scope of this post, but many large corporations will be subject to rates that differ by as much as 4.5%.

What happens when a corporation has both an outstanding overpayment of tax and an outstanding underpayment of tax, accruing interest during the same period of time? Might a corporation find itself paying more interest than it receives on equivalent amounts? Prior to the enactment of interest netting, taxpayers had two protections against that.

“Annual interest netting” addresses situations where a taxpayer has both an overpayment and an underpayment with respect to the same tax return outstanding during a given period. For example, Big Corp’s income tax return for 2012, filed on 3/15/2013, shows an overpayment of $3,000,000, which the IRS does not refund until 1/5/2015. After an audit, the IRS assesses additional tax of $5,000,000 for 2012. The IRS might try to assess interest on the $5,000,000 underpayment, from 3/15/2013 until paid, while paying Big Corp interest (at a lower rate) on the $3,000,000 original overpayment, from 3/15/2013 until 1/5/2015. In Avon Products Co. v. United States, 588 F.2d 342 (2d Cir. 1978), the Second Circuit concluded that individual transactions must be netted into a single balance before computing interest. Thus, the IRS would assess interest only on the net $2,000,000 underpayment from 3/15/2013 until 1/5/2015, and on $5,000,000 thereafter until paid. The IRS acquiesced not only in the result but also in the reasoning of Avon Products, although it has occasionally argued for a different result in particular cases. Revenue Procedure 94-60 and Revenue Ruling 99-40 are interpretations of the Avon Products doctrine; in recent years the IRS changed interest computation software and (somewhat) its methodology. (Revenue Ruling 99-40 actually – and likely unintentionally – mischaracterizes Avon Products in a subtle manner that obscures another procedural problem concerning recovery of excessive overpayment interest. The problem is potentially significant, although few people seem to be aware of it. But that is beyond the scope of this post.) However, “annual interest netting” only applies when the overpayment and underpayment are for not only the same type of tax but also the same tax return.

In other circumstances, the IRS exercise of its authority under Section 6402(a) to credit overpayments against outstanding tax liabilities, rather than issuing a refund, results in the elimination of the rate differential problem. A Section 6402(a) credit, of course, is not intended to and does not by itself accomplish netting. It is merely a collection method, almost always applied automatically to credit an overpayment under one TIN against an underpayment for exactly the same TIN. When the IRS makes such a credit, however, any interest rate differential is eliminated because of specific interest computation provisions. If an overpayment for 2010 is applied to an underpayment for 2011, for example, Section 6611(b)(1) provides that interest on the overpayment runs only to the due date of the 2011 underpayment – the same date that interest on 2011 underpayment starts accruing. If an overpayment for 2011 is applied to an underpayment for 2010, Section 6601(f) provides that there is no interest on that portion of the underpayment “for any period during which, if the credit had not been made, interest would have been allowable with respect to such overpayment.” This avoids an overlap period with interest accruing both on an overpayment and an underpayment, but a Section 6402(a) offset is only available if both balances are still outstanding. If the overpayment was refunded or the underpayment was paid, Section 6402(a) won’t help.

Those two solutions still left unresolved some instances of overlapping overpayments and underpayments. When Congress created the 1% rate differential in 1986, it asked the IRS to implement “the most comprehensive netting procedures that are consistent with sound administrative practice.” It continued to request IRS action as the rate differential widened to 3% and then 4.5%, and again when it enacted the Taxpayer Bill of Rights, but Treasury pushed back because it had no statutory authority for such netting. Congress eventually enacted Section 6621(d) in the Internal Revenue Service Restructuring and Reform Act of 1998. That provision requires that for any period when there are “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.” I’ve always referred to this as “global interest netting,” to differentiate it from “annual interest netting,” but I’ve dealt with some DOJ attorneys who prefer “net interest rate of zero.” For the remainder of this post, I’ll use “netting” to refer only to Section 6621(d).

Netting is available, under its terms, for taxes other than income tax. However, it only applies if underpayment interest is payable and overpayment interest is allowable during that period. Various “restricted interest” provisions of the Code state that under appropriate circumstances interest is not payable/allowable. Thus, if during the relevant period the taxpayer’s overpayment balance does not earn interest, there is no elimination or reduction of the interest rate on the equivalent underpayment balance. (Annual interest netting, however, does effectively net balances even if interest is not payable/allowable for one of the balances. In fact, that was the situation in Avon Products.)

The IRS computers cannot readily identify situations in which netting would be available. As a result, the IRS requires taxpayers to specifically request such netting on a refund claim. Revenue Procedure 2000-26 sets forth the requirements.

The evolution of “same taxpayer”

The legislative history for Section 6621(d) didn’t really explain what “same taxpayer” means. The meaning would be clear if we were talking about individuals. But corporations can change their corporate identity, through mergers, and join affiliated groups that file consolidated tax returns. What effect does that have? The IRS issued guidance as it began dealing with how to apply netting, and courts decided at least four major cases, including Ford, to clarify the boundaries of the term.

One of the IRS early attempts at guidance, Field Service Advice 2002-12028, required that one corporation “both be liable for the underpayment of tax, and entitled to the overpayment of tax.” This kept the restrictive adjective “same” while creating exceptions for mergers and consolidated returns. The principles it applied were:

• In a statutory merger, as a matter of law the surviving corporation is liable for any underpayment and entitled to any overpayment of the non-surviving corporation, so netting is permissible even for balances from tax years prior to the merger. (Situations 5 and 7)
• But in an acquisition in which both corporations survive, netting is impermissible because neither is liable for underpayments or entitled to overpayments of the other. Actual payment of the underpayment is not sufficient, absent legal liability for it. (Situation 6)
• All members of an affiliated group are severally liable for underpayments of the consolidated return, Treas. Reg. section 1502-6(a), so they can net their own overpayments against the group’s underpayments. (Implied from the reasoning for Situations 8 and 9)
• But a subsidiary is not liable for the group’s underpayment for tax years for which it was not a member of the group. Therefore, the subsidiary can’t net the group’s underpayment against its own overpayments. This applies when the subsidiary was acquired or came into existence after the year of the group’s underpayment (Situations 3 and 4), or if the subsidiary was owned by the parent during that year but was not part of the affiliated group (Situations 8 and 9).
• It is unclear whether subsidiaries would be entitled to some or all of the group’s overpayments for tax years when they were members of the groups and therefore able to net the group’s overpayments against their own underpayments. It is theoretically possible but that will depend on the facts and circumstances of the particular case. (Situations 1 and 2) The IRS reiterated this position in Chief Counsel Advice 2004-11003 but later walked back this concession in Chief Counsel Advice 2007-07002.

This was not a perfect, or complete, analysis but it was a good start. When I first read it years ago, I thought it was an interesting way to frame the analysis. Although we sometimes use a shortcut and say that this issue is whether Company A and Company B are the “same taxpayer,” this FSA focuses on whether a single taxpayer – Company A – has equivalent overlapping overpayment and underpayment. But in doing so, it allows under appropriate circumstances the attribution of overpayments or underpayments on Company B’s tax returns to Company A. There is solid support for many of the attributions in FSA 200212028.

Litigation ensued and the courts began fleshing out the nuances. In that litigation, the government pushed for a much more restrictive interpretation than in FSA 200212028. At various times, DOJ advanced three alternative arguments that would substantially limit or eliminate the ability of affiliated groups to obtain the benefits of netting. First, the TIN associated with the overpayment must be identical to the TIN associated with the underpayment. Thus, an overpayment or underpayment of the affiliated group could never be netted against an underpayment/overpayment of individual members; further, balances of the surviving corporation in a statutory merger could never be netting against pre-merger balances of the non-surviving corporation. Second, even if the TIN were identical, the taxpayer must not have undergone any substantial change between the two years. Thus, any addition or removal of members of an affiliated group or merger with another corporation – among other changes – would eliminate the ability to net interest. This would effectively eliminate netting for the largest corporations, which are those most likely to benefit from netting. Third, Congress did not intend netting to be available for overpayments or underpayments by consolidated returns at all. The Federal Circuit has accepted the first argument, with exceptions carved out, but has not (yet) adopted the more extreme interpretations.

Balances that arose prior to consolidation: Energy East Corp. v. United States, 645 F.3d 1358 (Fed. Cir. 2011), aff’g 92 Fed. Cl. 29 (2010). Energy East Corporation acquired Central Maine Power Company (“CMP”) in 2000 and Rochester Gas & Electric Corporation (“RG&E”) in 2002. Both subsidiaries became part of Energy East’s affiliated group and were included in consolidated returns from that point forward. The refund claim requested that CMP’s and RG&E’s overpayments for 1995, 1996, and 1997 be netted against Energy East’s underpayment for 1999. Thus, all of the balances began before the subsidiaries were acquired by Energy East.

The Court of Federal Claims used a dictionary definition of “same” – “being one without addition, change, or discontinuance: identical.” By that definition, CMP and RG&E were no longer the same taxpayers after the acquisition by Energy East, as they became part of an affiliated group. I’m often uneasy about a court’s use of dictionary definitions and this is no exception. Not only did it open the door for DOJ to argue for an overly narrow definition, but it also fails to recognize multiple different shades of meaning. For example, since I got married 31 years ago, I have moved twice, got a law degree, changed jobs four or five times and professions twice, lost weight, qualified for Medicare, etc. In one sense, yes, people would say I’m “not the same person” I was 31 years ago. But “same person” is also commonly used more broadly so that “not the same person” would refer to, say, major psychological changes – or The Invasion of the Body Snatchers. “Same” just raises the questions “to what extent” or “in what essentials” and becomes a vaguely ontological inquiry. So, dictionary definitions are a (minor) pet peeve just because of the flexibility and nuances of human language makes them too susceptible to manipulation and an ineffective guide to Congressional intent. I’m not a fan of the CFC’s definition of “same” for this issue.

The CFC also noted that “although they later became members of the consolidated group, [Energy East], CMP, and RG&E were different taxpayers with different employer identification numbers at the time of their overpayments and underpayments.” There was no support for treating them as the same for tax years prior to joining the group. This temporal requirement became an important part of the analytical framework, although I argue in Part 2 that it’s not necessarily the right way to look at this issue.

The Federal Circuit agreed that the consolidated return regulations provided no basis for concluding that individual members of the group should be treated as the “same” for years prior to their joining the group. It also rejected Energy East’s alternative argument that the focus of the “same taxpayer” determination should be when interest (to be netted) was accruing. Instead, the court concluded that they had to be the “same taxpayer” on the date of the underpayment and overpayments, based on the “last antecedent rule.” It said that “by the same taxpayer” referred to “equivalent overpayments and understatements” in Section 6621(d); thus, “the statute provides an identified point in time at which the taxpayer must be the same [by virtue of being members of an affiliated group], i.e., when the overpayments and underpayments are made.” As I’ll discuss in Part 2, this may not be precisely correct, at least in other contexts.

The court did not state that these three companies were the same taxpayer for years in which they were included in consolidated returns. That determination was not necessary for the decision, as the parties agreed the companies were not the same taxpayer for the years when the overpayments and underpayments were made.

Under the “attribution to a single entity” framing of FSA 200212028, this decision was clearly right. CMP and RG&E had no connection to Energy East’s underpayment for 1999 that would allow attribution under existing rules. Under a “two entities are treated as the same taxpayer” framing, it’s not as clear. I’ll return to that in Part 2, in the discussion of the Wells Fargo case.

That’s it for Part 1. Stay tuned for Part 2, where the action heats up.