February 24, 2021 - Interest rate floors have been in flux in the past year – and floor economics currently are moving against lenders. Prior to the Covid-19 crisis, LIBOR floors had been waning. When the crisis hit – and interest rates were slashed to near zero – floors made a comeback. Today, the incidence of floors remains higher but, with borrowers having the upper hand, the level of LIBOR floor has contracted (See COW). Interestingly, the focus on floors extends into LIBOR fallback language as well, as Covenant Review flagged this week. We hit the Floors, so to speak, below.
Let’s start with some background on LIBOR fallbacks. It’s critical to understand that there are two types of ARRC LIBOR fallback language for business loans: First, there is Amendment Fallback Language, which was introduced in 2019 and was widely used through late 2020. The ARRC discourages use of amendment fallbacks today. Second, there is Hardwired Fallback Language, which was refreshed in 2020 and is recommended by the ARRC (and banking supervisors) today. Importantly, the fallback methods treat interest rate floors differently.
In the 2019 ARRC Amendment Fallbacks, the language was primarily intended to keep a replacement rate from going negative. Thus, the language states that “if the Benchmark Replacement as so determined would be less than zero, the Benchmark Replacement will be deemed to be zero for the purposes of this Agreement.” While parties certainly can introduce a non-zero SOFR floor into an amendment agreement, it is possible that a loan may have a non-zero LIBOR floor but, in effect, a SOFR floor of zero.
The 2020 Hardwired Fallback language and ARRC Conventions are more prescriptive. Philosophically, the ARRC Guiding Principles attempt to minimize the value transfer between the counterparties when a loan falls back from LIBOR to SOFR. For this reason, in its SOFR In Arrears Conventions for Syndicated Business Loans, the ARRC states that, “[i]f there are interest rate floors in an existing LIBOR credit agreement that converts to SOFR, the ARRC Guiding Principles should be interpreted as applying that floor to SOFR plus the associated spread adjustment recommended by the ARRC (emphasis added).” The ARRC spread adjustments will match ISDA’s adjustments based on the five-year historical median difference between LIBOR and SOFR, as published by Bloomberg.
Likewise, the 2020 ARRC Hardwired Fallback Language for Syndicated Loans point to the same floor being used for LIBOR and the successor rate. The fallback’s narrative explains that “[b]ecause the successor rate plus the spread adjustment together represent LIBOR’s successor it is appropriate for, and the ARRC Guiding Principles should be interpreted to mean that, the floor applicable to LIBOR would be applicable to the Benchmark Replacement (which includes both components).” The Hardwired Fallback’s contract language follows that philosophy, noting in the Benchmark Replacement definition that “[i]f the Benchmark Replacement would be less than the Floor, the Benchmark Replacement will be deemed to be the Floor for the purposes of this Agreement and the other Loan Documents.” The “Floor”, of course, is the interest rate floor provided in the original loan agreement.
This week, Covenant Review analyzed whether the LIBOR floor is being retained in fallbacks. Hardwired fallbacks typically have the same floor for LIBOR and SOFR (plus spread adjustment). Amendment fallbacks, however, can vary markedly. CR noted 21% of 2020 fallbacks and 30% of 2021 fallbacks had successor floors that were reduced to zero from a non-zero LIBOR floor. CR published a pie chart showing that the impact could be meaningful: 11% of deals could see floors drop from 100/125 bps to zero. Thus, though it may be a low risk, this theoretically could reduce lender income when a loan falls back from LIBOR.
Why is this a low(ish) risk and what are the mitigants? The LSTA and CR flag several reasons. First, borrowers we have spoken with generally intend to refinance their loan prior to the June 2023 transition date; the floor can be addressed actively in the refinancing. Second, if lenders are seeing an amendment fallback, they should check the floor language – nothing prohibits the inclusion of a deal’s negotiated non-zero LIBOR floor in the amendment fallback language. Third, the Covenant Review analysis should further encourage the use of hardwired fallbacks, which match the LIBOR floor to the SOFR (plus spread adjustment) floor. Fourth, the amendment that actually transitions a loan from LIBOR to a replacement rate at LIBOR cessation in 2023 could introduce the same floor.
At the end of the day, there are risks, but – with knowledge and action – the risks should be easily managed.