How the Litigious Bird Caught the (Banque) Worm

The transactional plumbing of corporate debt payment systems is hardly where one expects to find watershed legal moments; and it usually lives up to that mundane reputation. But every so often, real disputes emerge, and they are often doozies. Such was the case with a $1.8 billion loan facility that Revlon Inc. entered into with a syndicate of lenders half a decade ago. This particular loan, in fact, had the honor of being embroiled in controversy not once, but twice within a single year. And the second of these imbroglios seems destined to cast a long shadow on the law, not just in the humdrum payment systems universe, but for contract law as a whole.

But first some backstory. In 2016, Revlon took advantage of historically low interest rates to secure the loan in question, with a syndicate of several sophisticated counterparties. The terms of the loan stipulated that Revlon had to make periodic interest payments up until 2023, when the loan and the balance became payable. To help fulfill its end, Revlon contracted with Citibank, N.A. (“Citi”), which was designated administrative agent for processing periodic payments and making appropriate bank transfers.

All seemed to go according to plan until the spring of 2020, when Revlon — which had fallen into financial difficulties necessitating additional capital infusions — executed a series of transactions that effectively undercut the collateral backing the loan. Several lenders immediately cried foul, alleging that Revlon’s actions not only substantially undermined the loan’s value, but they also breached the debt contract, entitling lenders to immediate payoff. For its part, Revlon insisted that the terms of the loan agreement permitted such collateral prestidigitation, that its actions therefore did not constitute breach, and (most of all) that the balance of the loan was not due and payable until 2023. The dispute on this issue has yet to be resolved – more on that below.

It was against this backdrop that the second snafu unfolded. Later in 2020, several lenders opted to exercise their rights under an arrangement with the company to transition into a new debt position. Such transactions – called “roll ups” – are common in this area and were permitted under the loan agreements here. But executing them is often tricky, since they typically occur at interim points between scheduled interest payments. In such situations, the lender must generally make good on whatever partial interest has accrued as of the date of the roll-up and transfer the principal balance on that lender’s account into a new species of debt. The process can be cumbersome in practice, and it has therefore become common to make the partial interest payments to all lenders, even those that are not rolling up their claims. Such categorical payments usually concede a small benefit to these other lenders, but do so in the name of administrative ease.

Making these unscheduled payments, however, remains cumbersome, and at Citi it still required a manual override of its automated platform to pay out the partial interest and to retire the principal of the rolling debtholders (transferring their balances to other debt ledgers). The process here evidently involved several manual entries made in less-than-intuitive locations. Through a series of mishaps and crossed wires (documented at length in court proceedings), these manual entries were mis-entered, and a nightmare scenario ensued: Citi not only paid out the partial interest (around $7.8 million) to all lenders as planned, but it also inadvertently repaid the principal of the loan – returning to lenders all of their principal balances (around $900 million). A few hours later, Citi employees discovered the error and sent several panicked notices to lenders informing them of a “fat finger” mistake, asking them to return the erroneous principal payments. Although several lenders cooperated, the others (representing around $500 million in principal) dug in, refusing to return the cash and daring Citi to sue if it wanted its money back.

Citi did just that. The dispute eventually landed in U.S. District Judge Jesse Furman’s courtroom in the Southern District of New York for an animated bench trial during the fall of 2020. The principal question was whether Citibank could obtain restitution for unjust enrichment under New York state law, or whether the lucky lenders were entitled to walk away with their unexpected bounty. In typical restitutionary actions involving mistaken benefit transfers, Citi would have a strong claim, and many commentators (including one of us) opined that the ultimate outcome would favor Citi. That said, restitution is a strange and unpredictable bird, and the lenders marshaled a full-throated defense, spotlighting a 30-year-old precedent in New York, Banque Worms v. BankAmerica Int’l, 570 N.E.2d 189 (N.Y. 1991), which they claimed allowed them “finders-keepers” rights. That argument is commonly known as the discharge-for-value defense, and it states that the recipient of a mistaken payment, lacking “knowledge that the money was erroneously wired,” could keep the funds and “should be able to consider the transfer of funds as a final and complete transaction, not subject to revocation.” This principle is animated by a longstanding policy goal of maintaining the finality of bank transactions, especially wire transfers that occur frequently throughout every single day.

In a noteworthy opinion issued on February 16, 2021, Judge Furman ultimately sided with the lenders in their discharge-for-value defense, holding that the recipients were not “on constructive notice of Citibank’s mistake at the moment they received the August 11th wire transfers.” Even though the lenders were promptly notified of the mistake after the fact, he held, the “magic moment” of fund transfer had already occurred, and the aforementioned judicial policy favoring finality of payments controlled. The transferred funds could not be clawed back, and the lucky lenders could keep it in satisfaction of their debt claims.

Citi will no doubt appeal the holding, and we suspect their appeal will focus (among other things) on two parts of Judge Furman’s opinion. The first is his holding that the discharge-of-value defense is available even if the balance of the loan is not yet due and payable at the time of the mistaken transfer. This is an issue of law that is a relatively close call (at least to us), and it was addressed somewhat elliptically in the Banque Worms case.

Second, and relatedly, Citi will likely challenge the court’s conclusion that the lenders were not on constructive notice that the payment was made mistakenly. Here, Judge Furman opined that sophisticated banks like Citi can be reasonably expected to have procedures in place to prevent the possibility that clerical mistakes like this will ever happen. Consequently, he concluded, no reasonable person in the lenders’ shoes would ever consider an unscheduled full payment to be the product of a mistake. In the parlance of Princess Bride character Vazzini, inferring a clerical error in this context would be nothing short of “inconceivable.”

This reasoning strikes us as curious on several fronts. First, the very fact that this type of error did actually happen undermines at least the strongest form of inconceivability – that a reasonable person would place zero weight on the likelihood of an error. Indeed, several of the lenders appear to have discussed that very possibility when the funds appeared without notice. (Op. 20-22). It warrants observing, moreover, that a large fraction of lenders were evidently convinced that there had been a mistake, and they accordingly returned the principal payments to Citi. Thus, while a reasonable lender might assess the probability of mistake to be low, it seems suspect to eliminate that prospect entirely.

Bearing that point in mind, and a bit more pedantically, Judge Furman’s reasoning here seems to give short shrift to another law – the law of probability – in the form of the infamous Bayes rule. To see why, it helps to think of Judge Furman’s holding on constructive notice as a formal Bayesian proposition about the probability that Citi could ever make a clerical error (or “Mistake”) of the sort described above when executing a transfer of funds:

Here, the opinion concludes that the value of “A” is extremely small (and well below 50 percent) for a sophisticated actor such as Citi. While this is likely true, the conclusion does not fully account for another important consideration: That an Unannounced Full Paydown (or “UFP”) had in fact occurred, and thus the unadorned probability above is only part of the picture. A more probative measure in the circumstances would focus on the probability of a mistake conditional on an Unannounced Full Paydown arriving, or Pr{Mistake | UFP}. Accordingly, Bayes rule implies the following relationship:

Note the two additional relevant terms here: “B”, the probability of a full payment given this type of clerical error; and “C”, the probability that Revlon would suddenly spring an unannounced full payment on unsuspecting lenders. And all the while, in order to carry their burden, the lenders would minimally have to demonstrate that this entire compound expression falls short of 50 percent.

Let’s consider the additional ingredients in turn:

  • Start with C – the unconditional probability of an unannounced full paydown. As noted multiple times in Judge Furman’s opinion, full payments such as this generally are preceded with an agent notice to the recipients; yet none was issued or received in this case.  Moreover, given the litigious backstory described above, where Revlon insistently maintained that it had no intention whatsoever of caving to the lenders’ acceleration demands on their steeply-discounted debt, it seems even more implausible that Revlon would suddenly have a 180-degree change of heart, caving in with neither an announcement nor settlement conditions. While it is certainly true that “early paydowns do happen” (Op. At 67), the history and context of these parties’ relationship make that prospect exceedingly unlikely.
  • Now consider B – the probability that an unannounced payment would occur conditional on making the type of clerical error that occurred here. Given the nature of the error as described in the opinion, it would seem that this probability is close (if not equal) to 100 percent. While different types of clerical errors might result in actions other than a full paydown, it was this specific type of error that the court seemed to concentrate on. (Op. at 66).

Combining these points, the lenders would have to prove that an unannounced full paydown – for parties actively litigating that same issue – would be far more likely (indeed by a factor of two) than a clerical error. Given the parties’ history, this would seem to be a difficult hill to climb, and Judge Furman’s opinion is not overly convincing in demonstrating how (if at all) they surmounted it.

To be sure, one might easily label our analysis above as little more than academic armchair speculation. And that label is probably accurate. But that’s also the point: Because the discharge-for-value doctrine is an affirmative defense, the lenders were required to carry the burden to prove it. Unclear cases, idle speculation, or evidentiary “ties” should have been resolved in Citi’s favor.

But that is clearly not how Judge Furman saw it. And significantly for the purposes of appeal, Judge Furman was the fact finder. Citi’s road to a successful appeal will therefore have to overcome the deference that is traditionally accorded to the judge’s interpretation of the facts discussed above. While an appeal still seems warranted in our view, there is a good chance that Judge Furman’s opinion will be upheld because of this deference, and the shadow of Banque Worms will be extended further. Should that occur, Citi would ironically take over the very position that the lucky lenders just vacated. Revlon’s erstwhile administrative agent will now be a principal creditor, even inheriting the plaintiff’s chair in a pre-existing collateral lawsuit that has been ongoing for the better part of a year. There may thus be more exciting chapters of this saga left to write.

If the holding is affirmed, what are the likely consequences for future lenders, borrowers, and payment agents? In our view, the outcome does not bode well for the sensible evolution of the law. Citi and its brethren will no doubt find ways to adapt to the new landscape, hiring additional compliance, legal, and IT personnel to reduce mistake rates even further. The incremental costs of doing so would likely be significant, but they will ultimately be passed on to future lenders and borrowers in the form of higher fees and delays. Future lenders will also adapt to this outcome, taking notice that it was the non-cooperative lenders who made bank (quite literally) in this dispute. The lenders who cooperated and worked with Citi collaboratively to correct the error, in contrast, ended up playing the sucker. Such incentives seem hardly consistent with the goal of minimizing the costs of executing and performing contracts – a key pillar of contract law and policy – even if they will add to the ranks of gainfully employed lawyers down the road.

There is, however, another potential response to this outcome that we think may be worth anticipating: contract evolution. There is significant legal authority that the restitutionary principles underlying this opinion are themselves default rules, and subject to alteration by contract.  Indeed, the key provision in the Uniform Commercial Code that underlies this area of law, UCC-4A § 501, states (in relevant part):

(a) Except as otherwise provided in this Article, the rights and obligations of a party to a funds transfer may be varied by agreement of the affected party.

And here, courts have recognized and reaffirmed the broad power that parties have to tailor their agreements to alter or eliminate these restitutionary principles. Should Judge Furman’s opinion be upheld on appeal, then, we predict that future parties to loan agreements may begin to change, limit, or waive entirely the discharge-for-value doctrine, replacing it with an express allocation of rights that is less likely to amplify waste, strategic behavior, and transaction costs.  And when that happens, the worm will most definitely have turned.

This post comes to us from Sneha Pandya, a third-year student at Columbia Law School, a managing editor of the Blue Sky Blog, and recipient of the 2021 Millstein Public Service Fellowship, and from Professor Eric Talley, faculty co-director of the Millstein Center and Isidor and Seville Sulzbacher Professor of Law at Columbia Law School.


  1. Justin

    They have already addressed this in a contractual right to reclaim the funds. Many deals have put in “Erroneous Distribution” provisions that allow agents to reclaim payments that are made in error, as determined by them (plus interest). This seems like a reasonable middle approach.

    • Eric Talley

      Thanks Justin — I saw some news reports about this development too; they evidently didn’t even want to wait for appeal. If you have specimen language for this new provision, I’d be interested in seeing it.

Leave a Reply

Your email address will not be published. Required fields are marked *