October 23, 2019 - At this point, most lenders know that LIBOR is likely to cease shortly after the end of 2021 and the Secured Overnight Financing Rate (“SOFR”) is very likely to be the replacement rate for syndicated loans and CLOs. While we know that, a major question remains: How do we make SOFR – a rate that is very different than LIBOR – work for the loan market? We explain some of the key issues for operationalization below.
First, as noted in Demystifying SOFR, there actually are four SOFR rates that need to be operationalized. Two of the SOFR rates – Forward Looking Term SOFR and SOFR Compounded in Advance – are very similar to LIBOR and should be relatively simple to operationalize. The other two – SOFR Compounded in Arrears and Simple Daily SOFR in Arrears – are very different and will be much more complicated to operationalize. But there is a decent chance that the syndicated loan market will adopt SOFR Compounded in Arrears, and so it is necessary to operationalize it – and quickly!
The “Known In Advance” SOFRs
Like LIBOR today, borrowers and lenders would know their SOFR in advance of the interest period in a Forward Looking Term SOFR and SOFR Compounded in Advance world. If Forward Looking Term SOFR existed, there would be a forward looking 1-month or 3-month rate published every day. Meanwhile, SOFR Compounded in Advance would be compounded before the interest period begins and therefore the rate would be locked and known at the beginning of the interest period. (Thus, we define these as “Known in Advance” SOFRs.) This is the how LIBOR works today: one rate is plugged into loan systems at the beginning of the interest period and counterparties know the exact amount they will pay or receive, and lenders can easily calculate daily accruals. Thus, for the “Known in Advance” SOFRs, accruals, notices and payment periods can all behave the same way as they do for LIBOR today. The major systems’ change is that for loans that transition from LIBOR to SOFR, a spread adjustment must be added to SOFR to make it more comparable to LIBOR.
So, this is very easy. Why isn’t it the sole solution? The first issue is that there is no guarantee that a Forward Looking Term SOFR reference rate will actually exist. It is contingent upon there being enough SOFR futures trading that a robust, stable and permanent forward looking term rate develops. That is by no means definite – and certainly not definite enough for all eggs to be put in this basket. Meanwhile, SOFR Compounded in Advance has the potential to be stale, particularly for longer tenors. For instance, in a rising rate environment, borrowers might be very interested in locking in long-tenored contracts to lock in a low rate for a long time. Meanwhile in falling rate environments, borrowers may trend toward shorter contracts to be able to replace the rate with a lower rate in the near term. This can make asset-liability management challenging. For these reasons, we cannot just assume that one of the “Known in Advance” rates will be the winner. And so, we also have to operationalize the SOFRs that are not “Known in Advance”. And we have to do it quickly.
The “Not Known in Advance” Rates
The “Not Known in Advance” Rates represent a significant change for the market. For these rates, the interest rate is accrued daily over the life of the loan contract. For instance, here is a hypothetical – and simplified – “Compounded Balance” approach to compounding daily SOFR[1]. In 30-day loan contract beginning April 1st, a lender/system would pull the April 1st interest rate on April 2nd and multiply it by the April 1st outstanding balance to determine the April 2nd interest. On April 3rd, the lender/system would pull the April 2nd SOFR and multiply it by the April 2nd principal plus accrued interest as of April 1st to determine the April 2nd interest accrual. And so on every day for 30 days. (For Simple Daily SOFR in Arrears, the process for pulling SOFR daily is similar, but the process is much simpler because the rate is not compounded.) This is clearly a very different from a world where the rate is known in advance. So what are some of the conventions and systems requirements that have to change to operationalize these “Not Known in Advance” rates?
Compounding: A critical convention is the recommended compounding methodology, which the ARRC Business Loans Working Group is developing. As of mid-October, there is ongoing debate on whether to compound the SOFR rate itself or compound the outstanding balance of the loan. However, most market participants agree that while SOFR should be compounded, any “spread adjustment” (which will be used to make SOFR more comparable to LIBOR) and the margin on the loan should not be compounded.
Business Day Conventions: Compounding only SOFR aligns loans with derivatives and hedges, which market participants have said is important. But there are other compounding conventions that are necessary to align with derivatives. For instance, derivatives only compound interest on business days, not weekends or holidays. On those days, simple interest is used. The loan market is likely to follow this convention.
Billing and Lookbacks: A “Not Known in Advance” rate introduces complexities in billing and paying interest. For instance, if a borrower has a 30-day loan that starts April 1st and accrues every day until April 30th, the final rate will only be known on May 1st. It is not fair for borrowers to be billed and pay interest the same day. The solution for loans may be a “look back”, which creates a multiple-day gap between when the final interest rate is known, when the borrower is billed, and when the interest is paid. In effect, instead of beginning to compound interest based on the rate published on April 1st, loans would “look back” five business days to March 25th to begin the interest compounding period. Thus, April 1st would use the March 25th SOFR, April 2nd would use the March 26th SOFR…and so on until April 25th. On April 25th, the agent would know the full 30-day interest amount and would be able to bill the borrower. The borrower would then have five days to pay the interest. (A particularly weedy issue – into which we will not delve here! – is whether to use an “observation” shift, which basically would align the business days in a lookback with the weighting of that business day. Again, this is how ISDA’s methodology works, and would be the methodology in any SOFR Index that is developed.)
Tools: If the discussion above seems complicated, that’s because it is. But tools are being developed to help market participants consume LIBOR without doing complicated calculations themselves. First, the Fed is planning to publish official “Compound Averages of SOFR”. The benefit here is that lenders could pull these rates from screens and put them directly into systems and not do any calculations. While these Compound Averages may be most useful for SOFR Compounded in Advance, the New York Fed also is looking to publish a “SOFR Index”, which would internalize all the compounding and holiday conventions and simply spit out a compounded rate when a practitioner selected a start date and an end date. This could both simplify the operationalization of SOFR in loan systems and provide a simple public “golden source” to check the accuracy of interest rates.
Next Steps: The ARRC’s Business Loans Working Group (BLWG) is developing a full list of recommended SOFR loan conventions. This list includes items mentioned above like compounding methodologies, spread adjustments, look backs and holiday schedules, as well as many other issues like rounding, break funding, SOFR floors and day count conventions. The BLWG is working with its counterparts in the UK, EU and Switzerland to align the recommended conventions as much as possible globally. These recommended conventions are being developed with business people, operations specialists and loan market vendors; this iterative process provides information for vendors to build systems that can internalize all variants of SOFR. And soon, the vendors should provide timelines on when systems should be SOFR-ready. For more information, please contact mcoffey@lsta.org or ehefferan@lsta.org.
[1] Compounding only the SOFR rate works if there is no intraperiod prepayment of loans. However, it may be necessary to compound interest on the outstanding balance of a loan if the outstandings fluctuate over the interest period.
