By Jeff Huther
ABA DataBank
The introduction of the Secured Overnight Financial Rate (or SOFR) as a widely adopted Libor alternative has changed short-term interest rate dynamics in ways that have not yet fully played out. Some of the differences in reference rates are due to how SOFR is constructed, and others are a result of how monetary policy is conducted. In this DataBank post, we dissect the differences and argue that, while the risk of persistent SOFR readings outside of the Fed’s 25 basis point target range for the fed funds rate is low, SOFR is likely to drift higher within that range and become more volatile, with implications for interest rates on bank assets and liabilities. While the Fed has not officially shifted its policy target from an interbank borrowing rate to SOFR, it has taken several steps in recent years to expand its role in the market underlying SOFR.
The effects of SOFR construction
Libor measured the cost of uncollateralized interbank lending. SOFR, as a repurchase agreement (repo) rate for Treasury securities, measures the cost of collateralized borrowing for large financial institutions, many of which are not banks. Lending in the repo market is dominated by large money managers, and borrowers are generally large securities dealers and hedge funds. Given this market structure, the shift from Libor to SOFR can be thought of as a shift in reference rates from Main Street to Wall Street. Much of the time, the forces affecting Main Street are the same as those that affect Wall Street. When differences arise, it is clear that SOFR is many things that Libor was not (see table below).
The Fed calculates SOFR using a large number of transactions. Daily repo volumes have averaged more than $1.5 trillion in 2024, a number that will rise following July 5, 2024, when regulators have required many more trades to be reported. The additional reporting includes trades by hedge funds and large asset managers; the net effect of expanded coverage will depend on how much volume of each entity type reports and market conditions. Hedge funds, as borrowers, typically pay higher rates than other borrowers while asset managers typically lend at, or near, published rates. During normal times, the additional volumes may increase stability in interest rates. During periods of stress, however, higher-risk borrowers typically see borrowing costs rise quickly and risk-averse lenders withdraw from the market.
The role of the Fed: ON RRP and SRF
Prior to the financial crisis, the Fed set a target for the fed funds rate, an interbank rate that was reliably tracked by Libor. In 2008, the Fed switched to targeting a 25 bp fed funds range instead of a single number (e.g., 5.25 percent to 5.50 percent) and, in 2013, the Fed (along with other international regulators) initiated a process to phase out Libor and transition to SOFR as a reference rate for private sector transactions. The Fed has developed two facilities (one as borrower, the other as lender) to support the repo market, which provide stability to SOFR readings. With its easily expandable balance sheet, the Fed narrows the range of SOFR outcomes by being a backstop borrower and backstop lender at fixed rates.
- Backstop borrowing. The Fed offers to borrow in the repo market through daily auctions at its Overnight Reverse Repo (ON RRP) facility. The minimum auction rate has been set at, or near, the bottom of the Fed’s target range since its introduction in 2013. Most of the Fed’s counterparties for this facility are money market mutual funds or Federal Home Loan Banks. These firms are also large lenders in the repo market, so borrowers generally have to offer to pay at least as much as the Fed is willing to pay in its ON RRP facility.
- Backstop lending. The Fed lends in the repo market through its Standing Repo Facility (SRF), which should help limit increases in SOFR during stress events. As with the ON RRP, SRF transactions are conducted through daily auctions. The credit risk inherent in lending, however, reduces the Fed’s set of acceptable counterparties to banks and its primary dealers (as of May 2024, there are only 36 participants). The minimum accepted auction rate has been the top of the FOMC’s target range, limiting its attractiveness as a day-to-day source of funding.
The excess liquidity provided by the Fed during the pandemic has not been fully absorbed yet, so while we have seen that the ON RRP has been effective in setting a floor for SOFR, we do not have evidence of the SRF’s effectiveness on setting a SOFR ceiling. Almost all repo rates have been below the SRF rate since its introduction in 2021.
The outlook for SOFR
As a bank funding rate, Libor typically rose during periods of financial stress. In theory, secured rates like SOFR may decline during stress events (as in flight-to-quality episodes). In practice, SOFR is vulnerable to the balance sheet constraints of the dealers that intermediate in the repo market, constraints that may be exacerbated by proposals for large banks to hold more capital. An example of the effects of constrained balance sheets occurred in mid-September 2019, when SOFR shot up following a relatively modest increase in funding demand in the absence of broader market disruptions (see Figure 1). Borrowers were unable to find lenders, even though banks had large cash holdings, leading the Fed to intervene to alleviate the imbalances. Given the ongoing balance sheet constraints on the large banks, future SOFR spikes should not be ruled out. And, of course, if capital requirements are increased for large banks, the higher intermediation costs are likely to increase the frequency of SOFR spikes.
The September 2019 episode highlights the effects of switching from a bank-based reference rate to a financing-based rate. The shock was unrelated to economic activity and bank financial health but had a significant effect on levered financial positions and led to a substantial change in the Fed’s balance sheet policy. The short duration of imbalances like the one in 2019 limits the costs of spikes in SOFR since most loan rates are based on longer-term averages rather than single day readings.
In 2024, SOFR has generally been flat at 5.31 percent, towards the lower end of Fed’s target range, except for days when financing pressures rise, such as the Treasury’s mid-month and month-end settlement dates. The broader pressures on SOFR, however, will eventually push the rate up within the Fed’s policy target range for overnight interest rates (which continues to be expressed in terms of fed funds, despite the Fed’s administrative focus on the repo market in recent years).
One source of pressure for higher SOFR is the Fed itself. Its ongoing balance sheet reductions reduce the amount of funds available for SOFR transactions, so the demand for repo funding is driving SOFR slowly higher within the Fed’s target range and increasing day-to-day volatility. The shaded area in Figure 2 represents the Fed’s 25 bps fed funds target range. The SOFR trend, shown in red, may slow in coming months now that the Fed has reduced the pace of quantitative tightening.
In the longer term, SOFR will rise and fall within the Fed’s target range depending on a set of factors not directly related to banks’ cost of funding. At the middle and end of months, when the U.S. Treasury issues large amounts of long-term debt, SOFR will rise. Demand for financing will rise and fall with hedge fund positioning. The supply of funding in the repo market will be driven by money market mutual fund balances, money market rates and the supply of U.S. Treasury bills. If the Fed’s SRF provides a ceiling on repo rates within the target range, quantitative tightening may eventually create a prolonged period of SOFR near the top of the Fed’s policy range, regardless of banks’ cost of funds.
