From Libor to SOFR: What It Means for Markets

• The phase-out of the London Interbank Offered Rate (Libor), the most widely referenced short-term interest rate in the world, is now well underway, with most new syndicated loans shifting to the use of Secured Overnight Financing Rate (SOFR) term rates after 31 December 2021.

• Existing loans referencing one-week and two-month Libor tenors were required to switch reference rates as of year-end 2021, though the transition for some other Libor tenors has been extended to June 2023 to allow for a smoother transition.

• Many key uncertainties about the Libor-to-SOFR shift have now been addressed, including disparities between the two benchmarks’ credit risk economics, SOFR’s former lack of a term structure, and a lack of uniformity in loan documentation related to the benchmark update.

• That said, basis risk remains an issue and continues to bring uncertainty to the CLO market in particular, given that CLO assets and liabilities may reference different rates during the transition period. Potential disadvantages for older CLOs with liabilities tied to higher credit spread adjustments (CSAs) relative to new CLOs issued with negotiated, market-based spreads are another area to watch.

• While we believe any systemic failure as a result of the Libor shift is highly unlikely at this point, we do not expect the transition to be seamless.

The sunsetting of the London Interbank Offered Rate (Libor) and transition to Secured Overnight Financing Rate (SOFR) term rates is now well underway, ushering in a new era for syndicated loan and securities markets long tied to Libor, the most widely referenced short-term interest rate in the world. As of 31 December 2021, issuers have been strongly encouraged by regulators to cease entering into new Libor-based agreements for all tenors, and one-week and two-month Libor tenors were retired as of that date for existing transactions. An extension through June 2023 was issued for existing US dollar Libor transactions at other tenors to allow for a smoother runoff.

While enforcement of the Libor-to-SOFR shift could pose challenges, regulators have been firm in their communications, and we expect most issuers to adhere to the guidance. Indeed, the move to term SOFR, which the Alternative Reference Rates Committee (ARRC) officially recommended as Libor’s replacement in July of 2021, witnessed the first term loan issuances and collateralized loan obligation (CLO) tranches utilizing the rate in the fall of last year. Approximately $18 billion in loans tied to the new SOFR rates were issued between October and 7 December 2021.1 Still, despite encouraging movement to the new benchmark rate now underway, a handful of issuers have already taken advantage of regulatory gray areas and continue to issue Libor-based contracts.

Notwithstanding these developments, many of the uncertainties and unknowns that slowed the transition and worried investors have now been resolved, including disparities between the two benchmarks’ credit risk economics (and the need to finalize an appropriate credit spread adjustment), SOFR’s former lack of a term structure, and a lack of uniformity in loan documentation related to the benchmark replacement. That said, basis risk in CLOs remains a concern, and other unknowns remain as the transition unfolds.

Libor transition timeline: pivotal dates

Libor versus SOFR: resolving key differences

Despite its widespread use, concerns about the subjectivity of Libor – which lacks an active underlying market and relies on banks’ reporting to set the daily benchmark – drove the push to retire the rate, which has served as the reference rate for an estimated $223 trillion in transactions tied to the rate,2 including the roughly $1.3 trillion US syndicated loan market3 and the more than $850 billion US CLO market.4  

Unlike Libor, which has essentially become a hypothetical rate, SOFR is calculated using real market transactions and benefits from significant liquidity and ample market depth. Two key disparities between Libor and SOFR – relating to credit risk economics and the former lack of a term structure for SOFR – have now been addressed at least in part, providing greater confidence to market participants.

While Libor is an unsecured rate and reflects bank credit risk, SOFR is based on underlying repo transactions collateralized by US Treasuries, making it a near risk-free rate. Accordingly, SOFR term rates are likely to offer a lower spread than the Libor equivalent. This means a credit spread adjustment is needed to ensure comparability between the benchmark rates, as well as to reduce the potential for value transfer between borrowers and lenders.

The UK Financial Conduct Authority (FCA) announcement on 5 March 2021 that Libor would no longer be representative for one-week and two-month US dollar Libor after the close of the year triggered the fixation of ISDA fallback spreads for all existing securities from that date forward (see table below). It’s important to note that the fallback spreads were determined for use in ARRC-related fallback language for existing transactions, while for new transactions, the CSA can be negotiated. The codification of CSAs marked a pivotal moment in the Libor transition, providing clarity to market participants and accelerating the momentum behind the shift.

Fixed ISDA Fallback Credit Spread Adjustments Offered Clarity for Existing Deals

Source: Source: FCA, ISDA, Bloomberg as of 5 March 2021.

Another key discrepancy between the two rates had related to SOFR’s former lack of a term structure, which was necessary for it to work for many financial contracts. On 29 July 2021, the ARRC formally recommended CME Group’s forward-looking term rates for SOFR, which standardized the use of these rates and cleared up a critical uncertainty for markets.

Implications for the US syndicated loan market

We see several areas for loan investors to watch as the transition unfolds.

The potential for differences in loan economics. During downturns and market disruptions, SOFR has historically tightened, while credit-sensitive rates such as Libor have tended to widen. Looking back over roughly the past two decades (2000-2021), while the spread differential has remained within 5 to 30 basis points (bps) about 70% of the time, it has reached 100 bps or more about 6% of the time, including during the global financial crisis.5 While in practice the CSA may be higher or lower than historical differences would indicate, the different ways in which the two rates respond to macro conditions could clearly change the economics for loan investors, who should not expect to see the same level of widening on the base rate for SOFR than for Libor.

Spread Differential Between 3M Libor and Calculated 3M SOFR

Source: LSTA and Bloomberg data as of January 2022. Three-month SOFR represents Bloomberg-calculated SOFR compounded to a three-month tenor for ISDA fallback rates.

The presence of Libor floors and further potential for differences in economics. At the end of 2021, about 55% of loans in the market had Libor floors in place greater than zero, with an average floor of about 73 bps.6 This translated to a roughly 40-bp weighted average Libor floor for the average loan, whereas one-month and three-month Libor were at the time well below that floor, at about 10 and 21 bps, respectively. Similarly, SOFR plus the CSA would also fall below that floor.

What this boils down to is that Libor or SOFR floors could help mitigate the differences in economics between Libor and SOFR because the floors offer a higher payment rate than on either of the other options. When loans reprice or refinance, Libor or SOFR floors are negotiable as well, and we’ve seen the average floor come down over time from 95 bps at the end of 2020 to around 73 as of this writing.7 If floors continue to decline, which will depend, of course, on supply and demand technicals in the loan market, then the value of those floors would diminish as well.

Are contract changes and renewals ‘new’ transactions? We’ve also seen questions in the market as to whether refinancings and other contract changes would be classified as “existing transactions” that could be amended, or rather as “new transactions” to which the restrictions on issuance of Libor-based loans after 2021 would apply. The latest guidance from regulatory authorities8 indicates that refinancings will be viewed as new transactions, though a draw on an existing undrawn loan facility would not be viewed as a new transaction. Some uncertainty regarding other transactions still exists, however, such as the treatment of add-on term loans and transactions brought by non-bank loan arrangers, which may face less scrutiny from regulators than those brought by banks.

Fallback language addressing the Libor shift. The market has also now largely addressed a lack of adequate fallback language around the discontinuation of the benchmark rate for US syndicated loan contracts. Necessary updates included incorporating 1) guidelines for the transition from Libor to a new reference rate, 2) triggers to effectuate the transition, and 3) the methodology used to calculate the stated reference rate, including any necessary spread adjustments.

As discussed in our March 2020 Libor commentary, the ARRC had initially suggested two potential solutions: an amendment approach, which did not specifically identify SOFR, but instead provided more flexibility in the event that a different successor rate emerged; and a hardwired approach, which designated term SOFR as the primary preferred successor rate in credit agreements. While the market initially preferred the amendment approach, hardwired fallback language quickly became the market standard after the March 2021 FCA announcement of the cessation of Libor and the July ARRC recommendation to use term SOFR.

Secondary market trading of Libor-based securities. Lastly, a key development at the end of 2021 was regulatory guidance allowing for trading of Libor-linked assets in the secondary markets, which averted potentially severe disruptions to those markets.

Implications for the US CLO market

With CLOs representing a significant source of demand for US syndicated loans, the technicals of these markets are distinctly interconnected. We see several key areas for CLO investors to watch during the transition.

Asset-liability rate mismatches and basis risk create a more complex transition for CLOs. The transition to SOFR is inherently more complex for CLOs than for loans given that with CLOs, both the assets making up the collateral and the liabilities they issue are floating rate. For CLOs issued before 31 December 2021, basis risk already exists, with liabilities linked to three-month Libor and most assets linked to one-month Libor. The SOFR transition, however, adds additional basis risk, as the interest rate a CLO receives on loan payments may be based on SOFR, while its liabilities may pay out a Libor-based rate. This creates the potential for risk related to benchmark mismatches, and the impact on equity returns, while not yet clear, could be unfavorable. Such basis risk will likely persist for CLOs during the transition period through June 2023.

What is the impact of Libor or SOFR floors? At this writing, both three-month term SOFR plus the CSA and three-month Libor are less than the Libor or SOFR floor for some loans. For CLOs, the presence of non-zero floors on loans will mitigate some erosion of the base rate on the asset side in the near term.

CLO indentures add language to smooth the transition. Following the FCA’s July 2017 announcement on the future of Libor, most CLO indentures incorporated language to facilitate an orderly transition away from Libor. In May 2019, the ARRC formalized its recommended fallback language for securitizations, allowing for a change in the reference rate as long as one of several triggers is satisfied. Separately, legislation was passed and signed into law in New York State in April 2021 to address Libor-based contracts that lack fallback provisions; similar legislation later cleared the US House of Representatives in December 2021, but it has yet to be finalized as of this writing.

Among the recommended triggers are a public announcement by Libor’s administrator, the FCA, the Fed, or some other authority with jurisdiction over the administrator that Libor has or will cease to be published; and the use of the new reference rate by more than 50% of the underlying loans within the CLO (either of which would independently trigger the change). Another potential trigger found in some CLOs is if at least half of the CLOs in the market utilize the new rate; however, this typically would not be a mandatory trigger. The CLO manager could consider the economics in making this determination, and the AAA tranches could reserve the right to reject such a move.

Credit spread adjustments for new deals remain a wild card. Based on a handful of CLO partial refinancings priced so far, the implied CSAs for CLOs with SOFR-based tranches have ranged from 10 bps to 20 bps depending on the tenor, and these spread adjustments will become clearer in the coming weeks as more deals are issued. To be clear, however, no SOFR-based CLO tranche to date has priced with a separate CSA; instead, the economic differences are being reflected directly in the tranche’s negotiated spread. On the asset side of the equation, we’ve also started to see that while investors wish to preserve the CSA, some loan issuers are attempting to eliminate it and instead add it to the overall spread. The problem is that if demand significantly outstrips supply in the loan market, the resulting technicals would lead to tighter spreads, which means the CSA would essentially disappear from the market. This could mark a shift in economics that is unfavorable for CLO equity returns.

Potential economic advantages or disadvantages of SOFR- versus Libor-based CLOs. Another question is whether older CLOs that switch from Libor to SOFR could be at a disadvantage versus newer CLOs tied to SOFR since inception. This is given that hardwired fallback spreads could be set at higher ARRC-determined rates for the former than what a newer CLO might be based on, with spreads determined according to current market supply and demand. In short, there will be new-issue SOFR CLOs and those that must use ARRC recommended spreads, and the latter will likely face greater costs from liabilities. Generally speaking, we believe CLOs with liabilities linked to SOFR will be well positioned in the current macro environment.

Most existing Libor-based CLOs now contain certain triggers that require a switch to SOFR. Unlike newer CLOs with negotiated rates, these will switch to the hardwired spread adjustments, which could differ. However, these transitions will occur over time. And just like for loans, CLO liabilities could attempt to refinance based on market conditions, which gives them an opportunity to renegotiate the coupon or spread on the liabilities. 

Clarity is emerging as the transition unfolds

The syndicated loan market has successfully adapted to changes in the past, and we believe the Libor transition is just another, if significant, part of this evolution and will ultimately prove successful.

Although loan and CLO investors should not expect a seamless transition, we see minimal likelihood of any systemwide failures. Bottom line, we believe that while investors should be aware of potential bumps in the road, the markets will weather the transition and that attractive opportunities remain in these critical asset classes.