Perli: Reflections on the Early Days of Reserve Management Purchases and the Maintenance of Ample Reserves

Introduction

Thank you to the Money Marketeers of New York University for the invitation to speak this evening.¹

As the System Open Market Account (SOMA) Manager, it is my responsibility to brief the Federal Open Market Committee (FOMC) on financial market developments and oversee the Open Market Trading Desk's (the Desk) implementation of monetary policy as directed by the FOMC. A key focus for the Desk over recent years has been adjusting the size of the SOMA securities portfolio, consistent with the Committee's 2022 principles and plans to implement monetary policy efficiently and effectively in an ample reserves regime.²

When I spoke before this group last March, the Federal Reserve was nearing its third year of balance sheet reduction.³ Since then, the FOMC has taken two important steps in line with its 2022 plans. First, having determined that reserve balances were approaching the ample range, at its October 2025 meeting it decided to conclude the runoff of aggregate SOMA securities holdings effective December 1; and second, when it judged that reserves declined to the ample range in December, it instructed the Desk to begin reserve management purchases (RMPs) to maintain reserves within that range.⁴

In my remarks today, I will provide my perspectives on the shift in money market conditions that led to those decisions, how the Desk approaches RMPs, and some general factors influencing reserve management and the size of the balance sheet. I'll conclude with a few comments on the role of standing repo operations (SRPs) and the composition of our RMPs.

Before I go further, I will give the usual disclaimer that these views are my own, and not necessarily those of the Federal Reserve Bank of New York, the Federal Reserve System, or any other organization.⁵

The Transition from Abundant to Ample Reserves

I'll begin with some context on how I think about ample reserve conditions.

The FOMC has defined its ample reserves implementation framework as a regime in which rate control is exercised primarily through administered rates, and in which the active management of reserves is not required.⁶

The word ample does not refer to a specific quantity of reserves; rather, it refers to that range of reserves that makes the federal funds rate only modestly sensitive to short-term variations in reserve supply (Panel 1). A reserve supply that is less than ample results in scarce reserves, where the federal funds rate is highly sensitive to changes in reserve supply, and rate control requires active reserve management. Conversely, a reserve supply that is greater than ample results in abundant reserves, where the federal funds rate is unresponsive to changes in reserve supply and the quantity of reserves exceeds amounts needed for effective rate control-this is where our financial system operated between mid-2020 and late last year.

When reserves are ample, I would expect to see a few things in money markets, as outlined in Panel 2.⁷ First, the Effective Federal Funds Rate (EFFR) should be close to the rate of interest on reserve balances (IORB), with low volatility. Second, repo reference rates, such as the Tri-Party General Collateral Rate (TGCR) and Secured Overnight Financing Rate (SOFR), should not deviate too much from IORB, on average, but show a moderate amount of volatility, especially around "high-pressure" days in the repo market.⁸ Third, usage of the Fed's overnight reverse repo (ON RRP) operations should be minimal.⁹ Fourth, the Fed's SRP operations should see occasional usage because volatility in repo rates volatility is likely to push repo rates above the SRP rate at times. Fifth, reserve supply should meet bank demand at the aggregate level, but domestic banks may still borrow moderate amounts in the federal funds market and other markets to smooth out the distribution of those reserves. And sixth, for the same reason, some banks may manage intraday reserve positions by delaying payments until later in the day or using a moderate amount of intraday credit.

From the start of balance sheet runoff through last summer, there were incremental signs of money market tightening, particularly in repo markets, but overall reserve conditions remained abundant. This was evidenced by the stability of the EFFR, the low volatility in other money market rates, and a range of other reserve ampleness indicators I've discussed in the past.¹⁰

Complicating the picture last year, however, were dynamics stemming from the federal debt limit. Because of those dynamics, in the first half of 2025, the U.S. Treasury drew down the Treasury General Account (TGA), essentially Treasury's "checking account" at the Fed. This temporarily boosted what I'll call system liquidity-or, the sum of reserves and ON RRP-and eased money market conditions. After the debt limit increase in early July, these trends reversed as the Treasury quickly rebuilt the TGA balance to its targeted level, largely through a rapid increase in net bill issuance.

The TGA rebuild and the cumulative impacts of balance sheet runoff brought system liquidity to the lowest levels since runoff began (Panel 3) and induced a notable shift in money market conditions, as I illustrate on the next few slides.

As liquidity diminished, rate pressures in repo markets intensified. ON RRP usage declined, reaching de minimis levels by October; this reduced the pool of funds readily available to shift from ON RRP into private repo. Consequently, repo rates had to increase more meaningfully to entice additional cash lending.¹¹ Repo reference rates rose relative to IORB and sustainably printed above EFFR (Panel 4). Repo rate volatility and sensitivity to Treasury issuance also increased (Panel 5).¹² And amid firmer and more volatile repo market conditions, SRP usage became larger and more frequent (Panel 6).

Higher repo rates contributed to a relatively fast increase in the federal funds rate from mid-September through November, with the EFFR rising from 7 basis points to 1 basis point under IORB. This rise in the EFFR was much faster than observed in early 2018, when EFFR first increased during the previous balance sheet runoff period (Panel 7).

As usual, the cash lending activity of the Federal Home Loan Banks (FHLBs) played an important role connecting the repo and federal funds markets.¹³ Amid higher repo rates, FHLBs shifted some lending from federal funds to repo and used higher repo rates to negotiate higher federal funds rates.

The reserve ampleness indicators I've condensed into a spiderweb chart in past speeches summarized well the tightening of reserve conditions (Panel 8). The share of repo transactions taking place at rates above IORB and the share of bank payments occurring late in the day approached levels seen in Q1 2019. The share of domestic bank borrowing in the federal funds market also increased notably.

The other two indicators remained at benign levels, but one of the two-the estimated elasticity of the federal funds rate to changes in reserves, or in other words, the estimated slope of the reserve demand curve-may have been held down by the aftereffects of the debt limit episode and other factors.

I want to emphasize that it was not necessary that all indicators moved at the same time to conclude that reserves were entering the ample range; experience showed that some indicators move sooner and faster than others, and that's the very reason why Fed staff monitors more than one.

By late last year, one could debate exactly where within the ample range reserves might have been, but the totality of the evidence was clear that the supply of reserves had shifted from abundant to ample. The FOMC made exactly that determination at its December 2025 meeting and directed the Desk to start RMPs that same month.

The Desk's Approach to RMPs

As you know, so far, the Desk has been conducting reserve management purchases at a monthly pace of $40 billion. It's important to understand that the money market conditions I described earlier did not require an immediate injection of reserves at that pace to maintain rate control. However, it is our job at the Desk to look ahead, and in doing so, we anticipated a large and rapid drain of reserves amid tax payment flows into the TGA during April tax season.¹⁴ Absent any action, there was a high likelihood that those flows would have pushed reserves below the ample range in April, potentially creating challenges for effective rate control.

In principle, the FOMC could have waited until the expected flows materialized in April to fill the reserves gap, but this would have implied the need to purchase very large amounts of Treasuries in short order. That would have been impractical from an operational standpoint, and it could well have compromised the smooth functioning of the Treasury bill market. Instead, the FOMC opted for a more prudent and gradual approach and decided to begin purchases in December and smooth out the cumulative RMPs needed to accommodate the anticipated drain in reserves in April over several months.

By construction, this strategy may result in reserves climbing to the higher end of the ample range in the first part of April; higher reserve supply would put downward pressure on short-term interest rates and, potentially, even result in some larger take up in the ON RRP. Temporary increases in reserve supply to the higher end of ample could also occur in the future, due to fluctuations in the TGA or other factors. Variation in the monthly pace of RMPs over time allows us to address fluctuations in factors that at times materially drain and add reserves.

Outlook for the Balance Sheet

An adjustment to our monthly purchase pace is likely to happen soon. Beyond April, the TGA is likely to decline as the Treasury uses the funds it receives via tax inflows to pay its obligations. All else equal, the decline in the TGA will result in a corresponding increase in reserves. The precise size of future RMPs is hard to predict exactly at this stage. But, as the Desk said in its statement in December, the monthly pace can likely be significantly reduced after April.¹⁵ To account for uncertainty and other factors, that reduction may be somewhat gradual.

Looking further ahead, generally, the size of the SOMA portfolio and amount of RMPs the Desk will conduct to maintain ample reserves will be driven by trend growth and variation in Federal Reserve liabilities, with our assessment of money market conditions serving both as a critical input and important feedback mechanism.

The three major items that comprise over 90 percent of current Fed liabilities are currency (Federal Reserve notes), the TGA, and reserves, so I will focus on those (Panel 9). Other Fed liabilities, such as the Foreign and International Monetary Authority (FIMA) reverse repo pool, and deposits we offer to government-sponsored enterprises (GSEs) and designated financial market utilities (DFMUs) are also quite important for the financial system but have a relatively smaller impact on aggregate balance sheet size and reserve management.

Historically, currency has tended to increase with nominal economic growth and international demand for dollars. In the last few years, annual growth has been up to about 3.5 percent.

With respect to the TGA, Treasury's cash management policy since 2015 has been to hold an amount of cash sufficient to cover one week of outflows from the TGA.¹⁶ This amount has grown over time in nominal terms, mostly alongside the size of the economy and federal debt; accordingly, the U.S. Treasury has increased its assumed quarter-end TGA balance from $320 billion in early 2016 to $850 billion on average in 2025.¹⁷ Absent any changes in Federal Reserve assets, the trend growth in currency, the TGA, and other non-reserve liabilities drains reserves; RMPs are needed to offset this.

In addition to trend growth, some non-reserve liabilities, particularly the TGA, exhibit significant seasonal variation, which can lead to swings in the supply of reserves and influence RMP decisions, as is the case now. In recent years, aggregate non-reserve liabilities (excluding ON RRP) have increased by as much as about $250 billion on certain days and by as much as roughly $460 billion over a rolling 1-month time horizon (Panel 10). Over the past several years, the TGA has risen by somewhere between roughly $175 billion and $400 billion during April tax season and has gradually declined back to more normal levels thereafter (Panel 11). Both the level and the variation in the TGA may increase in the future as the economy grows, and the Desk will have to take all these changes into account when formulating future RMP plans.

Lastly, an important driver of RMPs and Fed balance sheet size is the underlying demand for reserves.

Reserves represent the largest liability on the Fed's balance sheet at around $3 trillion currently; they play a special role within the banking system given their safety and immediate liquidity. Banks' demand for reserves is driven by a number of factors, notably intraday payment needs, liquidity management objectives, and regulatory considerations.¹⁸

All else equal, nominal reserve demand may be expected to grow over time with bank assets and bank payment volumes. But, importantly, structural changes in the banking system, for example, as a result of different liquidity regulations, can lead to corresponding changes in reserve demand. In particular, future potential changes to bank regulatory liquidity requirements may eventually reduce demand for reserves.

If that were to happen, the demand curve for reserves would shift to the left (as illustrated in Panel 12), the quantity of reserves consistent with an ample reserves regime would be smaller than it otherwise would be, and the Desk would take that into account when formulating a plan for the size of the SOMA portfolio. Our process and implementation framework are well positioned to deal with changes in demand for reserves and other Fed liabilities.

The Role of Standing Repo Operations and the Composition of RMPs

Even as the Desk conducts RMPs and reserves fluctuate within the ample range, there can be days when private repo rates come under noticeable temporary upward pressure. That can certainly happen on high-pressure days in the repo market. But it can also happen because of reserve supply forecast errors or unexpected increases in the demand for reserves. Those are precisely the cases where SRP operations are useful. SRP operations are an integral part of our rate control toolkit, and I expect them to be used when it's economically sensible to do so-that is, when private repo rates rise above the SRP rate. By offering our SRP counterparties an alternative source of funding, SRP operations reduce incentives for them to borrow at rates above the SRP rate, help dampen upward pressure on money market rates, and therefore support strong rate control.

To conclude, let me say a few words about the composition of our purchases. The Committee directed the Desk to increase SOMA securities holdings through purchases of Treasury bills and, if needed, other Treasury securities with remaining maturities of 3 years or less to maintain an ample level of reserves. This is in addition to reinvesting all principal payments from agency securities holdings into Treasury bills. This strategy is consistent with the Committee's intention to move toward a portfolio constituted primarily of Treasury securities.¹⁹ Importantly, RMPs do not represent a change in the stance of monetary policy and should not be confused with Large-Scale Asset Purchase programs, such as those implemented in response to the COVID-19 pandemic or the Global Financial Crisis, where the Fed purchased longer-term securities to support market functioning and ease overall financial conditions. RMPs are instead aimed at supporting interest rate control by preventing reserve conditions from tightening too much.

So far, purchases have been conducted entirely in Treasury bills, and operations have been running smoothly, with orderly market functioning and little price impact. I expect that this will continue to be the case in the future. Should continued bill purchases lead to market strains, however, the Desk has the flexibility to purchase short-term coupon securities to preserve strong bill market functioning. Of course, the Desk could also revert to coupon purchases if the FOMC decided that the share of bills in the SOMA portfolio has reached its desired level.

Thank you, and I look forward to the discussion.

Presentation

¹ I would like to thank Eric LeSueur and Linsey Molloy for their assistance in preparing these remarks, and Navya Sharma, Rachel Wilson, and Zack Youngblood for their assistance with the presentation.

² Board of Governors of the Federal Reserve System, Principles for Reducing the Size of the Federal Reserve's Balance Sheet, January 26, 2022, and Plans for Reducing the Size of the Federal Reserve's Balance Sheet, May 4, 2022.

³ Roberto Perli, Current Issues in Monetary Policy Implementation, remarks before the Money Marketeers of New York University, New York City, March 5, 2025.

⁴ Board of Governors of the Federal Reserve System, Federal Reserve issues FOMC statement, October 29, 2025, and Federal Reserve issues FOMC statement, December 10, 2025.

⁵ A repo market indicator I discuss later makes use of data collected under the authority of the U.S. Department of the Treasury's Office of Financial Research (OFR). The views expressed in the speech do not represent the views of the OFR, the Financial Stability Oversight Council, or the U.S. Department of the Treasury.

⁶ Board of Governors of the Federal Reserve System, Statement Regarding Monetary Policy Implementation and Balance Sheet Normalization, January 30, 2019.

⁷ See also Julie Remache, Monetary Policy Implementation in an Ample Reserves Regime, remarks at the Fixed Income Analysts Society, Inc., Women in Fixed Income Conference, Federal Reserve Bank of New York, New York City, February 12, 2026.

⁸ These include days with high payment flows due to Treasury securities settlements or tax payments, and balance sheet reporting dates, when temporarily tighter balance sheet constraints reduce dealer repo intermediation capacities.

⁹ Usage of both ON RRP and SRP operations may increase around reporting dates due to temporary reductions in dealer repo intermediation capacity and other frictions.

¹⁰ See, for example, Roberto Perli, Money Market Conditions and the Federal Reserve's Balance Sheet, remarks at 2025 U.S. Treasury Market Conference, Federal Reserve Bank of New York, New York City, November 12, 2025. See also Gara Afonso, Domenico Giannone, Gabriele La Spada, and John C. Williams, When Are Central Bank Reserves Ample?, Federal Reserve Bank of New York Liberty Street Economics, August 13, 2024, and Gara Afonso, Kevin Clark, Brian Gowen, Gabriele La Spada, JC Martinez, Jason Miu, and Will Riordan, A New Set of Indicators of Reserve Ampleness, Federal Reserve Bank of New York Liberty Street Economics, August 14, 2024.

¹¹ See Perli (November 2025).

¹² See Lucy Cordes and Sebastian Infante, Repo Rate Sensitivity to Treasury Issuance and Quantitative Tightening, FEDS Notes, February 12, 2025.

¹³ See additional discussion of FHLBs' lending behavior in Roberto Perli, Balance Sheet Normalization: Monitoring Reserve Conditions and Understanding Repo Market Pressures, remarks at 2024 U.S. Treasury Market Conference, Federal Reserve Bank of New York, New York City, September 26, 2024, and Gara Afonso, Gonzalo Cisternas, Brian Gowen, Jason Miu, and Joshua Younger, Who's Borrowing and Lending in the Fed Funds Market Today?, Federal Reserve Bank of New York Liberty Street Economics, October 10, 2023.

¹⁴ This was confirmed by Treasury's latest quarterly refunding statement, which estimated that TGA balances could temporarily peak around $1.025 trillion, plus or minus $50 billion by late-April (compared to nearly $840 billion on March 25). See U.S. Department of the Treasury, Quarterly Refunding Statement of Deputy Assistant Secretary for Federal Finance Brian Smith, February 4, 2026.

¹⁵ See Federal Reserve Bank of New York, Statement Regarding Reserve Management Purchases Operations, December 10, 2025.

¹⁶ This one-week amount is subject to a minimum balance of roughly $150 billion. See U.S. Department of the Treasury, Quarterly Refunding Statement of Acting Assistant Secretary for Financial Markets Seth B. Carpenter, May 6, 2015, and Remarks by PDO Assistant Secretary McMaster Before the 2025 Annual Primary Dealer Meeting at the Federal Reserve Bank of New York, September 29, 2025.

¹⁷ See end-of-quarter cash balance assumptions in the U.S. Department of the Treasury's Quarterly Refunding Financing Estimates.

¹⁸ See, for example, responses on factors determining banks' lowest comfortable level of reserves in the September 2023 Senior Financial Officer Survey Results.

¹⁹ Board of Governors of the Federal Reserve System (January 2022).