Good afternoon and thank you for the kind introduction. As an NYU alumnus, it’s a pleasure to be here to share some of my thoughts on money markets and the role of the Fed.1 Markets, as usual, are quite good at keeping us all on our toes. Recent events in money markets are definitely worthy of discussion for central bankers, market participants, and academics alike. Today I plan to further that discussion, with the typical caveat that the views I express are my own and not necessarily those of either the Federal Reserve Bank of New York or the Federal Reserve System.
Implementing Monetary Policy
First, a short comment on my role. I am the Manager of the System Open Market Account, or SOMA, a position that was created in 1936 at the very first meeting of the modern Federal Open Market Committee (FOMC)2. The nature of the role has not changed much in the past 88 years. The SOMA Manager is responsible for overseeing the conduct of open market operations at the direction of the FOMC, supporting the implementation of monetary policy. I have the same goal as my predecessors: to do so effectively and efficiently.
The FOMC sets the stance of monetary policy to influence interest rates and overall financial conditions with the aim of promoting maximum employment and stable prices3. The primary tool for doing so is raising or lowering the target range for the effective federal funds rate (EFFR). The EFFR represents the volume-weighted median rate on overnight unsecured transactions in the federal funds market. When the Committee decides to move rates up or down, it is my job to make sure that happens. Ensuring that the EFFR remains within the target range set by the FOMC is what we refer to as ‘interest rate control.’
The FOMC has communicated its intention to implement monetary policy in an ample reserves regime. In this arrangement the Fed supplies enough reserves that control over the EFFR and other short-term interest rates is exercised primarily through setting administered rates, and active management of reserve supply is not required. Administered rates are rates that the Fed directly controls, and two of them are key for monetary policy implementation. One is the interest rate on reserve balances (IORB). The IORB represents the rate paid on reserve balances held at the Fed and is the Fed’s main way of influencing the EFFR4. The IORB is intended to provide a floor under EFFR, but participants in the federal funds market that do not earn IORB are willing to lend federal funds below that rate. To reinforce the floor the FOMC introduced the overnight reverse repo facility (ON RRP), which is available to a wide range of counterparties that are important in short-term funding markets, particularly Government-Sponsored Enterprises (GSEs), and money-market funds (MMFs)5. The ON RRP offering rate is the second key administered rate for monetary policy implementation.
Complementing all this, the Standing Repo Facility (SRF) —which I will return to later—and other standing liquidity facilities act as backstops helping to contain upward pressure in rates and supporting smooth market functioning6. The combination of these administered rates and facilities helps keep the EFFR well within the target range. In fact, since the formal adoption of this framework in January 2019, the EFFR has only spent one day outside that range - the implementation framework works extremely well.
Of course, the federal funds market is just one money market, and it is certainly not the largest. But changes in administered rates typically influence other money markets,7 and that allows the Federal Reserve to achieve broad rate control. Still, persistent pressure in one money market segment can often be transferred to other segments. In that sense, rate control depends on the smooth functioning of all major segments of money markets.
As I noted in a recent speech,8 the market for repurchase agreements (or repos) is of particular interest. The repo market is large, of course, but it is also the market in which the Fed conducts temporary open market operations, including those under the ON RRP and the SRF. Additionally, the activity of the Federal Home Loan Banks (FHLBs) in both the repo and federal funds markets creates an effective channel for dynamics in repo to influence fed funds rates and volumes.
That brings us to the topic that I would like to spend most of my time covering today. As seen on the first slide repo rates have moved gradually higher relative to administered rates and the EFFR over the past couple of years, while repo spreads have widened. We have also seen signs of greater pressures on certain days, particularly financial reporting dates like quarter-ends. In the rest of my remarks, I will review the recent repricing of repo markets, my views on what can be learned from reporting date dynamics, and the role and design of the SRF.
Quarter-End Dates
First, I want to make clear that there is considerable evidence that reserve supply remains abundant - quarter-end pressures do not appear to be induced by a scarcity of reserves. I will not rehash the usual set of indicators here9, but I will highlight that one—the elasticity of pricing in the fed funds market—is now publicly available with monthly updates10. That measure continues to show no statistically significant price reaction in the federal funds market to changes in reserve supply, as has been the case since the Fed began reducing its securities holding in June of 2022.
The EFFR has remained very stable relative to administered rates with no sign of upward pressure, even as Treasury repo rates have moved higher. As I noted recently, this increase in repo rates was spurred by the combination of growing demand for Treasury securities financing, reduced availability of Fed-supplied liquidity owing to the ongoing balance sheet runoff, and some frictions that seem to be impeding the smooth transfer of liquidity from those who have it in excess to those with a shortfall. This slow rise in repo rates has been modest and orderly overall. In that sense, it is an entirely expected and arguably healthy development—a sign that money markets are normalizing as liquidity declines and as Treasury net issuance continues to grow.
But I, like many market participants, have observed the greater levels of volatility in overnight repo around financial reporting dates like quarter-end. Since June 2022, when balance sheet runoff began, higher overnight repo rates on such dates have become more common, although the extent of the increases has varied. The most recent quarter-end saw a few days of elevated rates, with the Secured Overnight Financing Rate (SOFR), a broad measure of overnight Treasury repo activity, peaking roughly 21 basis points above its average in the prior week.
Although this may seem large, it is important to place this most recent quarter-end in the longer-run historical context. To start, reporting date pressures are nothing new; they have been a regular feature of money markets for a long time. Also, as seen on the second slide, temporary repo market pressures of 10-25 basis points, as measured by SOFR, were common in 2015-1711, a period during which reserves were clearly abundant. What we saw at the end of September 2024 was within that range. And, importantly, like in the past, pricing returned to normal quite quickly after the end of September.
However, as I mentioned, there is some evidence that intermediation frictions and tighter liquidity conditions are playing greater roles in repo market pricing of late, particularly around financial reporting and Treasury settlement dates.
These frictions likely reflect a range of factors, including regulatory costs, counterparty credit limits, and other operational limits. These factors, combined with declining liquidity and increasing collateral supply, have likely contributed to the observed increase in repo rates around September quarter-end. This conclusion is consistent with recent market outreach conducted by my colleagues on the Open Market Trading Desk here in New York12.
All this suggests that the recent widening of repo rates around quarter-ends should be carefully monitored, and my colleagues on the Desk and I are doing exactly that. But it does not on its own indicate that reserve supply is anything other than abundant—particularly given the message embedded in the other indicators I mentioned previously. It also does not suggest that repo intermediation is becoming prohibitively costly, or that money market conditions are at risk of becoming disorderly in the near term.
The Standing Repo Facility
When market lending rates start to drift higher relative to our administered rates, it is natural to ask what role the Fed’s facilities play in maintaining rate control and money market functioning. Recent events have focused market attention on our Standing Repo Facility (SRF), and I would like to take some time to discuss it here.
The SRF is a standing facility intended to support the effective implementation of monetary policy and smooth market functioning by putting a ceiling on the EFFR. It is designed to achieve this by dampening upward pressure on repo rates that can spill over into the federal funds market by offering liquidity to eligible counterparties, including primary dealers and certain banks, subject to a minimum bid rate. That minimum bid rate is determined by the FOMC and is currently set to the top of the target range for the EFFR. Therefore, if certain market rates rise above the SRF rate, or if greater levels of intermediation are needed, counterparties can turn to the SRF to fund their borrowing needs at a lower rate. The SRF can be effective in dampening upward pressure on rates even with little or no usage as long as it is a credible outside option for our counterparties.
That said, repo activity at rates above the SRF rate does happen. For example, on one day this past quarter-end we saw more than $500 billion of overnight repo against SRF-eligible collateral trading at rates above the SRF rate. Is it surprising? Is it a sign that the SRF is not effective?
The answer to both questions is a clear no. To see why, it is important to understand the structure of the repo market and how the SRF relates to the three major segments of repo activity.
Dealers are central players in the Treasury repo market, and, being dealers, they function mostly as intermediaries. Under most market conditions they borrow funds, collateralized by Treasuries and other securities, from money market funds, government-sponsored entities (GSEs), and other cash-rich entities as well as from each other, and they lend funds, also collateralized, at higher rates to levered market participants.
The first segment of the repo market that I want to discuss is the tri-party repo market, which is a client-to-dealer segment where dealers borrow from cash lenders against a generic basket of securities and settle transactions through a third-party. A portion of tri-party repo is centrally-cleared and netted through the Fixed Income Clearing Corporation (FICC)13, while the rest of the market trades bilaterally and is not centrally-cleared and netted. The SRF settles through this uncleared segment of the tri-party repo market. Much of the trading in this segment happens early in the morning but settles in the afternoon; maturities are largely overnight.
The second segment is the interdealer segment, where dealers trade with each other. These trades are used to redistribute liquidity from dealers with an excess of funding to those with a need. This redistribution supports smooth market functioning by facilitating the movement of aggregate liquidity throughout the financial system. Importantly, these interdealer trades are predominantly cleared and netted via services offered by FICC. Most activity here takes place early in the morning, but, unlike the tri-party segment, it also settles in the morning as opposed to the afternoon.
Finally, the third segment is the dealer-to-client repo market, where dealers provide financing to levered market participants, primarily hedge funds. This segment involves mainly bilateral transactions in which counterparties face each other directly. That said, the role of sponsored activity, in which a FICC member will sponsor or guarantee the performance of a client to have the trade cleared at FICC, has been growing significantly in recent years and may well continue to do so. Non-centrally cleared transactions are more likely to be longer maturities compared to the tri-party or interdealer markets. Like the interdealer segment, the bulk of dealer-to-customer activity occurs and settles in the morning.
With all this in background in mind, is the SRF designed to be an effective backstop for all three segments of the repo market? The answer is a clear no.
For one, borrowers in the dealer-to-client segment of the repo market (again, mostly hedge funds) do not have direct access to the SRF. Therefore, the SRF cannot be expected to directly dampen rate pressures in that segment.
Primary dealers, on the other hand, do have access to the SRF. In principle, they should be incentivized to use the SRF whenever market rates in the interdealer or dealer-to-client segments are higher than the SRF minimum bid rate. It is important to bear in mind, however, that the interest rate paid is not the only cost of engaging in repo transactions; there are other costs associated with operational overhead, regulatory constraints, and internal risk limits that are very important. For example, a dealer that borrows from other dealers in the interdealer market and lends in the dealer-to-client market generally faces little operational overhead because transactions in both of those markets settle in the morning, as I just noted. Instead, a dealer that borrows in the tri-party segment and lends elsewhere faces higher operational overhead because the borrowing portion of its transaction settles in the afternoon instead of the morning. Further, almost all interdealer activity as well as a sizeable fraction of client-facing lending and borrowing is centrally cleared. Cleared positions can be netted down when measuring the size of dealer balance sheets, which reduces the costs associated with those trades relative to uncleared positions and increases overall gross capacity.
The SRF, because it is conducted over the tri-party platform, settles in the afternoon and is thus not lined up with the early-morning nature of most interdealer and dealer-to-client activity. It is also not centrally cleared, and therefore cannot be netted against other positions. There is, of course, a price differential at which the SRF would be more attractive considering these costs, but this difference, which is understood to be substantial, explains why dealers may not turn to the SRF even when market rates are higher than the SRF rate.
Still, the SRF does serve as an effective backstop source of funding for dealers in the uncleared tri-party segment of the market, where the facility operates. Thus, we would expect the SRF to dampen upward pressure on that segment of the market, and that dampening should propagate indirectly to other segments of the market. Note that I said “dampen” upward pressure, not eliminate upward pressure, because the SRF was never intended to put a firm ceiling on repo rates. To the contrary, dealers using the SRF for funding should, in turn, offer that liquidity to clients and other dealers at a spread. Thus, we would expect a sizeable share of this ‘on-lending’ will take place above the SRF minimum bid rate even if the facility works as designed because, again, dealers need to earn a spread for their intermediation activity.
This brings us back to recent events. Although the volume of repo trading above the SRF minimum bid rate has been substantial, this past quarter-end it was isolated to the dealer-to-client and interdealer segments. Dealers were willing to borrow there at a rate higher than the SRF rate because of the overhead and balance sheet costs that I mentioned earlier were evidently larger than the benefit provided by the lower SRF interest rate.
Activity for which the SRF is an effective substitute—i.e., uncleared tri-party trading—has occurred almost exclusively below the SRF offering rate. Outside of the most recent quarter-end, we have seen little volume trading above the SRF rate.
As seen in the third slide even this past quarter-end the amount of tri-party repo transactions at rates above the SRF rate was very small, while a significant portion of interdealer and dealer-to-client activity was at or above the facility rate. It is not at all surprising, then, that the SRF was used only to a relatively small—but, importantly, non-zero—amount at that time. To the contrary, I view that utilization as an encouraging sign that, as intended and when the conditions are right, our counterparties are willing to avail themselves of the facility.
Of course, uncertainty remains. All of the above does not guarantee that the SRF will see material use if pressures in the tri-party or other repo segments intensify. Some have pointed to potential stigma for usage of the facility—though whether such stigma exists remains to be seen. I personally see no reason why it should, given the SRF is not set at a substantial penalty rate, is offered against only high-quality collateral, and is substantially similar to the heavily-used repo operations conducted in 2019 and 2020. As it stands thus far, the SRF has behaved as expected, and it has supported the Fed’s ability to maintain the EFFR in the FOMC’s target range, even through recent quarter-end pressure.
Conclusion
To sum up, maintaining control over short-term interest rates is essential to enabling the FOMC to use monetary policy in furtherance of its dual mandate of maximum employment and price stability. That makes the smooth functioning of money markets more than an esoteric, acronym-riddled island populated by highly specialized market participants and the occasional central banker. It is at the heart of what the Fed does. Consistent with that imperative, the Federal Reserve System, and the New York Fed’s Open Market Trading Desk, are constantly scrutinizing money markets for evidence of even minor stress.
Dates on which exogenous factors generate pressure on repo rates are one place to gather that kind of information. The most recent quarter-end turn was valuable in that respect. Volatility in repo rates was elevated relative to the past few months. But, at least in terms of magnitude, it was not particularly elevated when viewed with a longer-term perspective. While such events bear monitoring closely, on their own, they do not suggest banking sector liquidity is anything other than abundant.
If more acute pressures were to emerge, the SRF is available to provide funding to market participants at an administered rate. The SRF is intended to provide a ceiling for EFFR, not for repo rates. I would therefore expect there to be sizeable fraction of repo trades occurring above the SRF rate, even if the facility works as intended. That remains very much the case today, and indeed it is what happened this past quarter-end.
In the meantime, I expect policy rates to remain well-controlled. If anything, usage of the facility on September 30th, however small compared to the flows, evidences the likelihood that market participants will use it if and when it is economically attractive to do so. For now, although we are monitoring events closely, there are few obvious and foreseeable risks to our continuing to implement monetary policy efficiently and effectively at the direction of the FOMC.
Thank you, and I’m happy to take your questions.
