My basic theory of Libor, the London interbank offered rate, is that it is a function call. You want to have a contract that specifies a floating interest rate, one that changes (say) every quarter based on prevailing interest rates. One way to do that is specify in the contract that, each quarter, you will observe some market data and call some banks for quotes and do some calculations and produce a number, the number being the interest rate. The contract could spell out the entire methodology to take some facts about the world and convert them into an interest rate.
But the way Libor works in contracts is mostly not like that. The way Libor works in contracts is mostly by saying “the interest rate will be whatever Libor says it is.” (Plus a fixed spread.) Exactly how that is expressed varies, but it is generally expressed by reference to some source, either the official administrator of Libor (formerly the British Bankers’ Association, now Intercontinental Exchange Benchmark Administration) or a Bloomberg or Reuters page that displays the official Libor.[1]
And then ICE is in charge of figuring out what Libor is, and Bloomberg and Reuters are in charge of getting that information and displaying it, and your contract can just take it as a given. To write the contract, you don’t have to know the exact mechanics of how ICE calculates Libor by polling banks about the interest rate at which they can do unsecured short-term borrowing. If ICE adds banks to the panel that it polls, or deletes banks, or changes the wording of the question it asks them, or tells them to use more transaction data in answering the question, or changes its method of topping and tailing and averaging the answers—all of that just flows through to your contract automatically. You call the Libor function, it returns a value, you use the value, and you don’t really care how the function operates internally. The people who maintain the function can tinker with it, and you won’t even notice.
We are in the middle of a long and boring effort to get Libor out of contracts. Contracts—floating-rate loans, interest-rate derivatives, etc.—are no longer supposed to use the Libor function. The main reason for this change is that it turns out that the way that Libor was calculated, during and shortly after the 2008 financial crisis, was pretty bad: The BBA polled banks about their cost of short-term unsecured borrowing, and the banks lied about it, so Libor was, in an important sense, “wrong.”[2] A secondary reason for the change is that the eurodollar markets used to calculate Libor are not as active and important as they once were, so even a more honest calculation of Libor—the kind that ICE does now—may not reflect “true” interest rates the way Libor used to. And so regulators want banks and derivatives traders to stop using Libor and start using some other, more market-based interest-rate reference. In the U.S. this is mainly SOFR, the Secured Overnight Financing Rate, which is calculated by the New York Fed based on actual transaction data.
There are various problems with this transition, but the simplest and dumbest one is that there are a lot of contracts that say “the interest rate will be Libor” (plus a spread), and you have to go find all of them and get the contracting parties to agree to cross that out and write in “the interest rate will be SOFR” (plus a spread) or whatever. That is hard administratively—you have to find the contracts, you have to get the two parties to pay attention, etc.—but there is also an economic problem. Libor and SOFR are different; they measure different things; Libor is unsecured and SOFR is secured; SOFR is overnight and Libor comes in longer tenors. If your loan pays interest of six-month Libor plus 150 basis points, will it now pay the six-month SOFR futures rate plus 175 basis points, or six months of daily SOFR compounded in arrears plus 168 basis points, or what? The borrower will say “let’s change to SOFR but not increase the spread,” the lender will say “let’s change to SOFR and increase the spread a lot,” there will be some economics to be worked out, and there’s no guarantee that everyone will agree. And so banks are going out and trying to renegotiate trillions of dollars of contracts to replace Libor with something more sensible, but they kind of have to do that one client at a time.
The simple dumb solution would be to answer these questions, once and for all, by changing the internal mechanics of Libor. ICE could just wake up one day and say, “We will keep reporting Libor, but instead of being based on a panel of banks, it will be SOFR plus 20 basis points, that’s just what Libor means now.”[3] And then if your contract says “our interest rate will be Libor,” you will go to the Bloomberg or Reuters page that reports Libor, and it will keep reporting Libor, and the function will produce an answer just like before. But now the guts of the function will be based on SOFR—the good rate, the one regulators like, the one with a future—rather than the old and discredited method of calling up banks for their unsecured lending rates.
In practice it would be a bit tough for ICE to do this, and people who use Libor and don’t like how ICE answers the economic questions would get mad and sue it. On the other hand … New York could do it?
New York Governor Andrew Cuomo has proposed legislation that would help prevent hundreds of billions of dollars of financial contracts from descending into chaos when the London interbank offered rate expires.
Provisions to help troublesome Libor-linked contracts switch to replacement rates are contained in Cuomo’s state budget plan, which was published on Tuesday. Bankers, investors and regulators see such proposals as crucial to ensuring that a large swath of the global financial system isn’t disrupted. …
As home to the world’s biggest financial center, much of the debt falls under New York law. …
The U.K. hasn’t faced the same complications around sterling Libor, partly because of its different exit strategy. Proposals to keep publishing a “synthetic” Libor number that doesn’t require trading data from panel banks would help legacy contracts that can’t transition to avoid a cliff-edge scenario at the end of 2021, when the U.K. benchmark will likely retire.
In New York, the bill would allow contracts to instead use the replacement rate recommended by the Fed Board, New York Fed, or the ARRC.
Lots of financial contracts are governed by New York law, so the New York legislature can, within some limits, change what those contracts mean. Cuomo’s budget includes an article on “Libor Discontinuance,” which says (section 18-401, page 237) that “On the Libor replacement date, the recommended benchmark replacement shall, by operation of law, be the benchmark replacement for any contract, security or instrument that uses Libor as a benchmark,” unless there is a different fallback provision in the contract. If you trace through the defined terms, what that means is basically that when Libor stops publishing, any contracts that use Libor will automatically instead use a different function. The different function will be whatever is recommended by the Fed,[4] which is administering the Libor transition, and will presumably be (1) SOFR, (2) termed out in some way (using futures curves or compounding to compute a longer-term rate from overnight SOFR), (3) plus a spread (to reflect the difference between secured and unsecured rates). The law doesn’t choose what the function will be; it leaves it up to the Fed.
This is actually pretty similar to the U.K.’s “synthetic Libor,” though the U.K. approach is a bit more direct:
The term began circulating in the leveraged loan and floating rate bond markets after the UK government announced in late June 2020 that it intends to amend the UK regulations of benchmark interest rates to give the Financial Conduct Authority (“FCA”) enhanced powers, including the power to select a new calculation methodology to any benchmark. This includes the power to direct the administrator of LIBOR to change the methodology of LIBOR if the FCA determines that the current LIBOR methodology (i.e., polling of panel banks) is no longer representative of the market and if it would be both more appropriate and feasible to change to an alternative methodology. This new methodology would result in a new interest rate being published as the “screen rate” instead of LIBOR. In other words, the Intercontinental Exchange intends to publish a new rate on the same screen and in the same location on the screen where it had previously published LIBOR.
However, it is important to understand that, as envisioned by the FCA, this new rate would not replicate LIBOR through some synthetic calculation. Rather, in the accompanying FAQs the FCA explains that the new methodology would follow the market consensus that emerges on how to calculate fair alternatives to LIBOR. For most currencies, this will be a risk-free rate chosen by the applicable LIBOR currency area, adjusted for the relevant term of the contract, and with a fixed credit spread adjustment added. In other words, for the USD LIBOR market, “synthetic LIBOR” would likely be SOFR plus a modifier.
The U.K. approach would directly change what shows up on the Libor screen. Contracts wouldn’t have to change at all; they’d continue to call the Libor function, but by U.K. law that function would now work differently. The New York approach would just automatically change all the contracts written under New York law. The contract—the piece of paper documenting the trade—would still say “the interest rate is Libor,” but lawyers would know that, as a matter of law, you have to read those words to mean “the interest rate is SOFR plus a spread.”
Nobody thinks any of this is a good solution. Here is a speech from last March by Edwin Schooling Latter of the U.K. FCA, warning people that synthetic Libor is a terrible idea, because “parties who rely on regulatory action enabled by this legislation, will be giving up their control over the economics of their contracts.” Much of the proposed New York legislation is about exempting contracts from the law if the contract parties pick some other fallback, some other way to get around Libor. Obviously it is better for each contract to be carefully renegotiated to reflect the economic intent of the parties, rather than some dumb one-size-fits-all approach imposed by Albany.
But that is easy to say, and hard to do. What is the economic intent of the parties? Often they have different intents: Borrowers want to pay a lower rate, lenders want to get a higher rate, etc. Often they have no particularly clear view on what sort of interest rate they want and how it should be calculated. That was why they used Libor, why they called some externally administered function and let someone else give them a number for their interest rate. That number was widely accepted, it was the “normal” number, and that’s what they wanted; they just wanted whatever the interest rate was. For those people, sure, whatever, let New York law tell them to use what the New York Fed thinks the interest rate is. They want someone else to tell them a number. Why not do that?
Bloomberg Opinion
by Matt Levine
Jan 21, 2021