In the US, the Federal Reserve convened the ARRC to introduce a “more robust” reference rate and to facilitate the transition. This led to the Secured Overnight Financing Rate (SOFR) which launched and began publishing rates in April 2018. One-month and 3-month futures contracts began trading on the CME in May 2018. 2 Like SONIA and €STR, SOFR is based on actual transaction data, and is calculated using a weighted-average trimmed mean. Moreover, the volume of transactions in its underlying market — the US Treasury repurchase agreement (repo) market — is high, at about $1 trillion daily.13
ARRC has endorsed SOFR as its preferred alternative to LIBOR for certain derivatives and other financial contracts, and has set an ambitious timetable to encourage its use. Yet it has made clear that SOFR is not an exclusive solution.14 There are drawbacks and challenges associated with SOFR, and despite official-sector enthusiasm for SOFR, reception has been mixed.
Volatility has been an issue for SOFR. For instance, in mid-March, with markets roiled by COVID-19, SOFR dropped to 0.26% before doubling the next day, and then falling again. In September 2019, idiosyncratic strains in the repo market sent SOFR above 5%. This volatility has raised concerns that SOFR may not be an appropriate replacement for LIBOR,15 although the use of compounded averages for SOFR help to smooth the impact,16 and volatility should decrease as the volume and variety of products referencing the benchmark increase.
In cash bonds, government-sponsored entities have dominated SOFR-based issuance. The Federal Housing Finance Administration has prohibited the Federal Home Loan Banks from entering into LIBOR transactions or investments,21 and Fannie Mae and Freddie Mac will stop accepting LIBOR-based ARMs by the end of 2020.22 They have created new ARM products indexed to the 30-day SOFR average.23 Financial institutions and corporations, however, continue to prefer LIBOR. Given the low private-sector market volume, it may take a cessation of LIBOR to drive SOFR issuance.
Developing a term reference rate based on SOFR will also be key, although derivatives volume must increase before this can happen. In H2 2020, CME17 and LCH18 transitioned from using the Fed funds rate to SOFR to discount cleared swaps, which has resulted in increased volumes. ARRC has set a goal of creating the term reference rate in H1 of 2021.19
Importantly, there is private-sector interest in a US dollar alternative to SOFR that, like LIBOR, is sensitive to credit conditions. This is largely because floating-rate financial contracts based on SOFR can present risky imbalances, particularly for certain banks. During credit-market volatility, banks’ borrowing costs may rise. Products, such as floating rate business loans tied to LIBOR, serve as a hedge against such rising costs, because LIBOR is correlated with banks funding costs: as funding costs rise, so do the proceeds from floating-rate instruments tied to LIBOR.
In contrast, SOFR may move in the opposite direction from banks’ borrowing costs. Because SOFR represents the cost to fund treasury collateral, it does not have a credit component. Therefore, it does not reflect the cost of funding bank balance sheets. This can be an issue during market volatility, when fund flows put downward pressure on SOFR in anticipation of central bank policy easing. A drop in SOFR would decrease proceeds from floating-rate products tied to SOFR. During the 2009 financial crisis the spread between LIBOR (which reflects banks’ funding costs) and SOFR would have exceeded 50 basis points (bps) for about one year. During the H1 2020 downturn, the SOFR/LIBOR spread exceeded 50 bps for 41 days, peaking at 140 bps. This widening is a significant concern for smaller and medium-sized banks that rely on more credit sensitive funding, because they don’t hold Treasury portfolios needed to obtain funding via the repo market.
To address the need for a credit-sensitive reference rate, regulators have hosted a series of meetings with a Credit Sensitivity Group, composed mainly of regional banks, to identify potential solutions. Various possibilities have been explored, but no preferred option has been identified. In November, following a series of workshops convened by the Federal Reserve Bank of New York, senior bankers conceded that it is unlikely that a credit-sensitive rate would be ready prior to LIBOR’s demise, pushing banks to peg lending products to the risk-free rate.
One possibility gaining momentum among banks is the use of AMERIBOR, the alternative to LIBOR developed by the American Financial Exchange, reflecting the unsecured borrowing costs of more than 1,100 US lenders. In May, Fed Chairman Jerome Powell wrote “While [AMERIBOR] is a fully appropriate rate for the banks that fund themselves through the American Financial Exchange (AFX) or for other similar institutions for whom AMERIBOR may reflect their cost of funding, it may not be a natural fit for many market participants.”23