When Does the Statute of Limitations Start on the Erroneous Refund Suit? (4/29/21; 6/30/24)

I recently read and posted a comment to the following blog discussion:  Keith Fogg, Late Filed Erroneous Refund Suit (Procedurally Taxing Blog 4/22/21), here.  I wanted to explore further the issue of the starting date for the statute of limitations for erroneous refund suits.  Readers will recall that § 7405 permits an erroneous suit for refund which is "erroneously made.”  Section 6532(b) says that the statute of limitations to initiate the erroneous refund suit is “2 years after the making of such refund, except that such suit may be brought at any time within 5 years from the making of the refund if it appears that any part of the refund was induced by fraud or misrepresentation of a material fact.”  (The Regulation, § 301.6532-2, merely repeats the statute.)  Both the 2- and 5-year statutes of limitations require a starting date.

The Procedurally Taxing Blog linked above discusses a case, United States v. Page, No. 3:20-cv-08072 (D. Ariz. April 16, 2021),  here, involving the following facts:

• 5/5/17 – IRS mails taxpayer an erroneous refund check for $491,104. 
• ??/??/?? – Taxpayer receives the erroneous refund check
• 4/5/18 – Taxpayer cashes the erroneous refund check.
• ??/??/?? – IRS writes to request return of the erroneous refund
• 12/19/19 – Taxpayer responds by returning $210,000.
• 331/20 – IRS sues for erroneous refund.

The Government moved for default judgment.  Taking seriously its obligation to test the validity of the Government’s claims even on default judgment, the Court held (Slip Op. 3):

Under Ninth Circuit law, “[t]he refund is considered to have been made on the date the taxpayer received the refund check.” United States v. Carter, 906 F.2d 1375, 1377 (9th Cir. 1990); see also O’Gilvie v. United States, 519 U.S. 79, 91 (1996) (“[T]he law ordinarily provides that an action to recover mistaken payments of money accrues upon the receipt of payment.” (internal quotation marks and citation omitted)).

Under the facts outlined above, that meant that the statute of limitations foreclosed the erroneous refund suit absent fraud (which the Government did not allege).

There are two problems with the holding.  First, I think it makes no sense (as I noted in my then off-the-cuff comment to the Procedurally Taxing Blog entry because I don’t think the Government could have brought an erroneous refund suit until (i) at the earliest the refund check was cashed and perhaps (ii) until the refund check cleared.  I address that below.  Second, on a more procedural issue, I think the language of the Ninth Circuit precedent was dicta and not persuasive dicta, so the Page Court was not bound by it.  I address that issue below.

Accrual of the Erroneous Refund suit

Suits in Anglo American jurisprudence are subject to an accrual requirement.  The right to the action must have accrued before the action can be litigated and, as a consequence, the statute of limitations does not commence until the right to bring the suit accrues.  It thus would make no sense to say that, although there is a potential right of action, the right of action can commence before it accrues sufficiently that the claimant could actually bring a lawsuit on the claim.

So, applying logic (I discuss law below), I ask the question whether the IRS can bring an erroneous refund suit upon the taxpayer’s receipt of the erroneous refund check before at least cashing the check?  What if the taxpayer never cashes the check, could the IRS still sue for refund?  I think that is nonsense.  Thus, it would seem to me that the IRS will have to await the taxpayer actually receiving something of value (cash for a refund check) before suit could be brought and even await the check clearing.  Of course, the Government has an easy remedy before the check is cashed – i.e., it can cancel the check.  I suspect that the Government would be laughed out of court if it filed an erroneous refund suit before the check is cashed.  Another remedy that might be available is a replevy action to recover the check.  But a replevy action is not an erroneous refund suit. So just common-sense logic tells me that an erroneous refund suit would not accrue before, at the earliest, the taxpayer cashes the check.

In United States v. Wurts, 303 U.S. 414 (1938), the following facts were presented:

• 3/15/32 – IRS approves a refund (subsequently determined to be erroneous).
• 4/30/32 – IRS mails the refund check
• ??/??/?? – Taxpayer receives the check
• ??/??/?? – Taxpayer cashes the check
• ?/??/?? – Check clears.
• 4/26/34 – Government sues for erroneous refund

If the accrual date for the erroneous refund suit were when the IRS approves the refund on 3/15/32, the suit was untimely.  Interpreting the plain meaning of the statute, the Court reasoned that the internal IRS approval of the refund cannot be what the statute required – the making of the refund.  The Court said (p. 418, footnote omitted):

It would require language so clear as to leave room for no other reasonable construction in order to induce the belief that Congress intended a statute of limitations to begin to run before the right barred by it has accrued. Obviously, the Government had no right to sue this taxpayer to recover money before money had been paid to him. The construction urged by respondent would allow the statute of limitations to begin to run against recovery on an erroneous payment before any such payment is made. As said by a House Committee in reporting on a statute of limitations contained in a revenue act, n9 "Logically the period of limitation should run from the date of payment, since it is at that time that the right accrues."

We are of opinion that Congress did not intend the limitations of § 610 to run against the Government until the Government's right "has accrued in a shape to be effectually enforced." 

This statute does not begin to run against the Government when a claim is erroneously allowed. It begins to run from the date of payment. The judgment below is not in accord with this construction of the statute and is

Reversed.

Note how the Court uses accrual concepts, familiar in the law, to support its result.  The Court did not address the issue of when the payment is made thus causing the action to “accrue.”  The possibilities on the date of payment are:

• Date of receipt of check
• Date of cashing (depositing) the check
• Date of check clearance

Of course, only if and when the check clears would the Government have the right to sue for an erroneous refund for only then is it out any money that it can claim to have been erroneously refunded.  Also, perhaps importantly to starting a statute of limitations, the Government will not know when the taxpayer received the check.  As in the Page case, the facts do not include the date of receipt of the check, but it is rare that the Government will know that date (at least that is true before the USPS started tracking mail).

In O’Gilvie v. United States, 519 U.S. 79, 91 (1996), the facts were (pp. 90-91):

• 7/6/90 - IRS writes refund checks (and presumably mailed them).
• 7/9/90 - taxpayers receive refund checks
• ??/??/90 - Refund checks cashed
• ??/??/90 - Refund checks clear
• 7/6/92 - Government files erroneous refund suit

The Court said (p. 91-92):

The children concede that they received the refund checks on July 9, 1990, and they agree that if the limitation period runs from the date of receipt—if, as the Government argues, that is the date of the "making of" the refund—the Government's suit was timely. But the children say that the refund was made on, and the limitations period runs from, the date the Government mailed the checks (presumably July 6, 7, or 8), in which case the Government brought this suit one or two or three days too late.

In our view, the Government is correct in its claim that its lawsuit was timely. The language of the statute admits of both interpretations. But the law ordinarily provides that an action to recover mistaken payments of money "accrues upon the receipt of payment," New Bedford v. Lloyd Investment Associates, Inc., 363 Mass. 112, 119, 292 N. E. 2d 688, 692 (1973); accord, Sizemore v. E. T. Barwick Industries, Inc., 225 Tenn. 226, 233, 465 S. W. 2d 873, 876 (1971) ("`[T]he time of making the . . . payment . . . was the date of actual receipt' "), unless, as in some States and in some cases, it accrues upon the still later date of the mistake's discovery, see Allen & Lamkin, When Statute of Limitations Begins to Run Against Action to Recover Money Paid By Mistake, 79 A. L. R. 3d 754, 766-769 (1977). We are not aware of any good reason why Congress would have intended a different result where the nature of the claim is so similar to a traditional action for money paid by mistake—an action the roots of which can be found in the old common-law claim of "assumpsit" or "money had and received." New Bedford, supra, at 118, 292 N. E. 2d, at 691-692. The lower courts and commentators have reached a similar conclusion. United States v. Carter, 906 F. 2d 1375 (CA9 1990); Akers v. United States, 541 F. Supp. 65, 67 (MD Tenn. 1981); United States v. Woodmansee, 388 F. Supp. 36, 46 (ND Cal. 1975), rev'd on other grounds, 578 F. 2d 1302 (CA9 1978); 14 J. Mertens Law of Federal Income Taxation § 54A.69 (1995); Kafka & Cavanagh, Litigation of Federal Civil Tax Controversies § 20.03, p. 20-15 (2d ed. 1995). That conclusion is consistent with dicta in an earlier case from this Court, United States v. Wurts, 303 U. S. 414, 417-418 (1938), as well as with this Court's normal practice of construing ambiguous statutes of limitations in Government action in the Government's favor. E. g., Badaracco v. Commissioner, 464 U. S. 386, 391 (1984).

We concede the children's argument that a "date of mailing" interpretation produces marginally greater certainty, for such a rule normally would refer the court to the postmark to establish the date. But there is no indication that a "date of receipt" rule has proved difficult to administer in ordinary state or common-law actions for money paid erroneously. The date the check clears, after all, sets an outer bound.

The Court said that the Government argued that the date of receipt of the check was the date of "making of" the refund.  But, actually, the O’Gilvie Court's holding was only that the date of mailing rule argued by the taxpayers was not the proper rule, noting at the end that "the date the check clears, after all, sets an outer bound."  And, importantly, Commonwealth Energy Sys. cases were after Ogilvie (1998 for trial level and 2000 for Court of Appeals) were after O'Gilvie with persuasive reasoning for the date of clearing, which establishes a date certain rather than date of receipt which does not.

In United States v. Commonwealth Energy Sys. and Subsidiary Cos., 235 F.3d 11, 14 (1st Cir. 2000), the Court held that the statute starts when the check cleared.  

Finally, in United States v. Greene-Thapedi, 398 F.3d 635, 639 (7th Cir. 2005), the Court said explaining Commonwealth Energy and distinguished the Ninth Circuit opinion in Carter and explained its own reasoning:

Only one court of appeals has addressed the precise question before us. It concluded that the "making of such refund" occurs on the date when "the check cleared the Federal Reserve and payment to the taxpayer was authorized by the Treasury." United States v. Commonwealth Energy Sys. and Subsidiary Cos., 235 F.3d 11, 14 (1st Cir.2000). In Commonwealth Energy, the IRS issued a sizable refund check, which was received by the taxpayer on July 27, 1995. The government filed suit on July 30, 1997, two years and three days later, claiming that the refund was issued in error. Id. at 13. In deciding how to measure the limitations period, the First Circuit concluded that a check-clearing rule was superior to a date-of-receipt rule both because it construed the limitations period in favor of the government and because it provided a clear reference point for all concerned. Id. at 14-15. ("Although the Treasury cannot know for certain when a check is received by a taxpayer, it can know when that check clears, and determine whether or when to file suit accordingly."). Noting that Treasury records reflected authorization of payment to the taxpayer on August 2, 1995, the court held that the government's suit was not barred by the two-year limitations period when it was filed on July 20, 1997. Id.

Greene-Thapedi counters that there are two other appellate decisions that apparently adopt a date-of-receipt rule. In our view, however, those cases do not support her position. Paulson v. United States, 78 F.2d 97 (10th Cir.1935), is of little or no help to her. In the same vein as the Supreme Court's decision in Wurts, Paulson held that the two-year limitations period did not commence with the date on which the IRS Commissioner issued a refund but instead "when the money is paid." 78 F.2d at 99. The court did not clarify what it meant by "when the money is paid," but that phrase seems to us to be consistent with the First Circuit's position. Indeed, the Tenth Circuit commented on the fact that the IRS could cancel payment on a refund check at any point prior to payment. Id. ("Ordinarily a statute of limitation does not begin to run until a suit could be brought. Certainly it cannot be contended that a suit of this nature may be maintained and judgment secured after the schedule of refunds and credits has been approved, but before the money is paid to the taxpayer."). United States v. Carter, 906 F.2d 1375 (9th Cir.1990), is just another case in which the court was presented with a choice between the date of mailing and the date of receipt. As the Supreme Court later did in O'Gilvie, the Ninth Circuit chose the date of receipt. Id. at 1377-78. It said nothing about the date of negotiation, as there was no need to do so.

We conclude that the First Circuit's approach in Commonwealth is sound, for the reasons given by that court. Factual disputes are more likely to arise when a court is asked to determine the date that a taxpayer received a refund check in the mail. By contrast, a court can determine with near certainty the date on which the Treasury authorized payment on the check. This rule permits both the government and the taxpayer to know exactly when the limitations period commences. As applied to these facts, the two-year statute of limitations began to run on December 19, 1999, when the Treasury department records show that Greene-Thapedi received payment on the refund check. The government's suit, which was filed on November 28, 1999, was thus not time-barred.

So, my conclusion is that the date the check clears is the "right" answer.  Hopefully, the Ninth Circuit will be convinced of that as well and clarify its prior statements. [Added 6/30/24 8:00 am: the Ninth Circuit reached the same conclusion on appeal. United States v. Page, ___ F.4th ___, ___ WL ___, 2024 U.S. App. LEXIS 15505 (9th Cir. 2026).]

Mea culpa:  I stated that my comments on the Procedurally Taxing Blog post were off-the-cuff.  Indeed, they were.  I had not bothered to check my own Federal Tax Procedure book which in the 2020 Practitioner Edition on page 232 n. 1044, here, says:

n1044 The courts have held that the statute starts running when the taxpayer, having deposited the erroneous refund check, the erroneous refund check then clears the Federal Reserve and payment is actually made by the IRS. O'Gilvie v. United States, 519 U.S. 79, 91 (1996); United States v. Commonwealth Energy Sys. and Subsidiary Cos., 235 F.3d 11, 14 (1st Cir. 2000); and United States v. Greene-Thapedi, 398 F.3d 635 (7th Cir. 2005).

Was the Ninth Circuit Precedent Really Controlling on the Page Court?

The precedent cited by the Page court was, in my reading, dicta and thus would not be controlling on the Page court unless it is persuasive.  I don’t think it is persuasive for the following reason (posted as a comment to the Procedurally Taxing Blog entry on Page):

The Ninth Circuit held that the two-year statute of limitations was an affirmative defense that the taxpayer must prove. Under the facts, the taxpayer negotiated the check within the two-year window. The thing that was uncertain under the facts was when the taxpayer received the check. Technically, all the Ninth Circuit resolved was that the taxpayer did not properly lay the factual predicate to assert the claimed legal defense. The Court did not have to even make a statement about the timing of accrual of the erroneous refund suit.

Thus, the Ninth Circuit could have said something like this: “The taxpayer claims that the two-year window started on the date of receipt of the erroneous refund check. The taxpayer failed to show the factual predicate for that legal claim. Hence, the taxpayer loses without this Court having to resolve that issue.” If that is right, the resolution of the legal issue was not necessary to the holding and thus dicta. For the reason noted above, nonsensical dicta at that.

JAT Comment:  Readers may have noted that I did a mea culpa above about not checking the Federal Tax Procedure Book first.  I thought I would share an anecdote about not checking the book first.

The first time I learned the lesson was when a former student in a class on Real Estate Taxation at University of Houston Law School taught it to me in the 1980s.  As I was pontificating (close but not the same as blustering or bluffing), a student asked a question.  I said that I can’t answer that question, but she and I could walk through the steps to get to the right answer.  (I was going to use it as a teaching experience for the student and myself.)  Another student jumped in to say something to the effect that “Mr. Townsend, you don’t have to guess at the answer, it is right here in the book.”  Oops.  After that I repeated that anecdote to every class urging the students to jump in at appropriate times to move the class along.  (That student who called me out was one of my best ever, but claimed much later that he did not recall the incident; you can be sure that I did because it caused instant but temporary embarrassment which I got over.)

That same phenomenon played out in writing this blog, but I caught it myself after writing the heart of the blog.  I made off-the-cuff comments on the Procedurally Taxing Blog that a case wrong and nonsensical.  I offered no authority to support my claim.  In exploring the issue today, I went through the research and only toward the end decided to check my Federal Tax Procedure Book.  Had I checked first, I would have spared some time doing research that I had already done years ago and summarized.  And, I would have made this blog entry shorter.  Still, since I had done the research today, I decided that I should leave it on the blog in case anyone can find it useful.  

I am sure that is not the last time I will make the mistake of not checking my own writings either good or bad (in this case they were good, but if I don't have the discipline to check, I am sure in some cases it may be bad).