Ample liquidity for a safe and efficient banking system

Good morning. Thank you all for joining us again today. Yesterday’s discussions were so enriching, and I’m excited to build on them this morning.

As you know, the Federal Open Market Committee (FOMC) reduced interest rates earlier this week and announced it would end the runoff of the Fed’s asset holdings as of Dec. 1. In my remarks this morning, I’ll discuss the stance of monetary policy. I’ll then turn to the topic of this conference and describe how the Fed’s balance sheet fosters a safe and efficient liquidity environment for the U.S. banking system. These are my views and not necessarily those of my FOMC colleagues.

Monetary policy

I would have preferred to hold interest rates steady at this week’s FOMC meeting. Congress gave the FOMC a dual mandate: to pursue maximum employment and stable prices. The labor market remains balanced and cooling slowly. Inflation remains too high, taxing the budgets of businesses and families, and appears likely to exceed the FOMC’s 2 percent target for too much longer. This economic outlook didn’t call for cutting rates.

While the government shutdown has reduced the availability of national statistics, a wide range of alternative data sources continue to provide visibility into the state of the economy. Those sources include private-sector indicators, continuing administrative data such as unemployment claims, regional surveys run by many of the Federal Reserve banks, and the many conversations that my colleagues and I have with business and community contacts every week.

The labor market remains roughly balanced. At 4.3 percent, the latest reading on the unemployment rate was up only slightly over the past year on net.

Payroll job gains fell markedly in 2025. But slow job gains don’t necessarily mean there’s more slack in the labor market. Labor supply has fallen at the same time as demand, particularly due to changes in immigration policy and labor force participation. In consequence, despite the drop in job growth, we’re not seeing a rapidly widening gap between the number of jobs available and the number of people who want work.

My staff estimates that break-even payroll growth, the number of net new jobs needed to hold unemployment steady, has fallen to around 30,000 jobs per month. The Bureau of Labor Statistics has delayed its September employment report. However, private-sector data suggest September payroll growth remained near the level that would keep the market in rough balance. Likewise, layoffs and unemployment claims have stayed low, although I’m mindful of recent layoff announcements by some large employers.

Resilient consumer and business spending continues to support employment. Consumer spending growth has slowed from last year, but only to a pace that equals or slightly exceeds the longer-run trend. Stock market gains are fueling wealthier households’ demand. Large companies are investing enthusiastically in artificial intelligence and data centers. The picture for lower-income families and smaller businesses is softer than the aggregate, but it remains stable.

The Dallas Fed Weekly Economic Index combines 10 daily and weekly data series to deliver a rapid signal of the state of the economy. It is averaging 2.4 percent. That’s consistent with strong gross domestic product growth in the third quarter.

The risks to the labor market do lie mainly to the downside. In this low-hiring environment, the job market could have difficulty absorbing any significant pickup in layoffs from the current low level. Buoyant asset valuations can sometimes snap back without much warning, which might take the wind out of consumer spending. And the federal government shutdown will pose greater risks to economic activity the longer it continues. But these are all risks the FOMC will be able to closely monitor and address promptly if there is evidence they are becoming more likely to materialize.

Turning to the price stability side of the mandate, inflation is still too high and too slow to return to target. The FOMC targets a 2 percent inflation rate as measured by the annual change in the price index for personal consumption expenditures, or PCE. Inflation by that measure exceeded the target in each of the past four years. It’s on track to do so again this year. The index rose 1.8 percent just in the first eight months of the year, and forecasters expect it to end the year up about 2.9 percent.

While inflation has come down significantly from the postpandemic peak, it’s still not convincingly headed all the way back to 2 percent. The Blue Chip Economic Indicators surveys dozens of private-sector forecasters about their economic outlooks. The Blue Chip consensus outlook is for 2.6 percent PCE inflation in 2026 and around 2.4 percent in 2027, followed by fluctuations between 2.1 and 2.2 percent all the way out to at least 2031—never all the way back to target.

The FOMC’s 2 percent inflation target is a serious commitment. We’ve reaffirmed it repeatedly, most recently in the “Statement on Longer-run Goals and Monetary Policy Strategy” we issued this August. Our obligation to the public is to deliver on this commitment.

To be sure, inflation persistently running several tenths above the target is not the emergency that inflation persistently several percentage points above the target would be. And we have an equally serious obligation to pursue maximum employment. The FOMC’s long-run strategy calls for a balanced approach to our two objectives. I carefully weigh the potential labor market costs of measures to reduce inflation. But labor demand and supply remain in balance. The FOMC already mitigated downside risks by cutting rates at its previous meeting, in September. The remaining risks to employment are ones we can monitor closely and respond to if they are becoming more likely to materialize, not ones that currently warrant further preemptive action. For those reasons, I did not see a need to cut rates this week. And I’d find it difficult to cut rates again in December unless there is clear evidence that inflation will fall faster than expected or that the labor market will cool more rapidly.

Balance sheet normalization

The FOMC also decided to stop reducing the Fed’s asset holdings as of Dec. 1, ending a phase of balance sheet normalization that began in mid-2022. In contrast to the action on interest rates, the decision to end asset runoff was one I supported. Money market conditions indicate the Fed’s balance sheet is now much closer to a normal size, following the expansion in response to economic and financial stresses during the pandemic and the subsequent runoff.

More than just a return to normal, the reduction in the balance sheet forms part of how the Fed promotes safe and efficient liquidity for our nation’s banks. Of course, the primary responsibility for banks’ funding and liquidity risk management rests with bankers. Still, the Fed’s balance sheet underpins some aspects of bankers’ decisions, and that’s what I’d like to talk with you about today.

Efficient reserve supply

The Fed’s single largest liability, nearly half the balance sheet in total, is bank reserves. U.S. dollar reserves are the safest and most liquid asset in the world. Reserves are assets to the banks that hold them and, correspondingly, liabilities for the Fed. A dollar of reserves is always worth exactly one dollar: no more, no less. Reserves are also the only asset a bank can use immediately to meet a withdrawal or send a payment.

If you’re holding a loan to one of your customers, a corporate bond, or even a Treasury bill, you have to sell that asset or borrow against it to get money before you can make a payment. Reserves, by contrast, already are money. You can make a payment with reserves right away. You never have to monetize them first. In normal times, it’s straightforward for healthy banks to borrow against assets to get cash. But when markets come under stress, or when a bank needs to quickly obtain unusual amounts of liquidity, monetizing assets can be more difficult. Reserves help a bank manage its liquidity risk and serve customers efficiently through thick and thin.

As authorized by Congress, the Fed pays interest on banks’ reserve holdings. These interest payments have no material cost to taxpayers because, over time, the Fed covers them out of interest earnings on its assets. However, interest on reserves plays a crucial role in the efficiency and safety of the banking system.

When the interest rate on reserve balances (IORB) is close to interest rates on other money market instruments, it doesn’t cost banks anything to hold reserves to meet their liquidity needs. If reserves paid less than market rates, on the other hand, or no interest at all, each dollar of reserve holdings would cost a bank the difference between market rates and IORB. Banks would have a strong incentive to avoid holding reserves.

The Fed didn’t pay interest on reserves until 2008, and the old system was inefficient and risky. Because reserves were costly to hold, banks tried to hold as few reserves as possible. With few reserves on hand, banks often throttled outgoing payments until they received incoming money each day. Banks also arranged their business models to reduce regulatory reserve requirements, and then they tried to minimize their actual reserve holdings by lending anything above the regulatory minimums to other banks.

Considering that it doesn’t cost the Fed to provide reserves, all those gymnastics represented wasted effort for the economy. And that’s the best case. If costly liquidity held banks back from providing services their customers would have valued, the old system of scarce reserves put a brake on economic growth—for no good reason. Moreover, if banks were taking unnecessary liquidity risk because they avoided holding the safest, most liquid asset in existence, the financial system was more vulnerable to shocks—again, for no good reason.

Since 2008, the Fed has met banks’ demand for reserves with interest on reserves close to market rates. In 2019, the FOMC decided to continue implementing monetary policy in the long run in this framework of ample reserves. This is, quite simply, a better approach: better for banks, better for the households and businesses that are their customers, and better for the U.S. economy. I’m proud the Fed adopted it, and I’m committed to continuing it.

If reserves are so great, why did the FOMC run off asset holdings from 2022 to 2025 and, consequently, also reduce reserve supply? Like I used to tell my kids when we went to the ice cream shop, it’s possible to have too much of a good thing. The FOMC acquired large quantities of assets during the pandemic to support smooth market functioning, sustain the flow of credit and provide economic stimulus. Those purchases helped keep the public health crisis from turning into a financial crisis or recession. But among their side effects was that the purchases made reserves not just ample but abundant or super-abundant—more plentiful, that is, than the quantity banks wanted to hold with market rates near IORB.

It’s clear reserve supply exceeded banks’ demand because market interest rates fell significantly below interest on reserves. In 2022, rates on overnight repurchase agreements, or repos, collateralized by Treasury securities ran 10 or more basis points below IORB.

Supplying too many reserves is just as inefficient as supplying too few reserves. When aggregate reserve supply is too high, the banking system holds reserves instead of more productive assets. And liquidity is too expensive for financial institutions other than banks. To best serve the economy in normal times, the Fed needs to strike a balance and supply an efficient quantity of reserves—neither too many nor too few.

The transition to ample reserves

There are many indicators of reserve ampleness, but the most significant one to me is the position of money market rates relative to IORB. In an efficient system, market rates should be close to, but perhaps slightly below, interest on reserves on average over time.

“On average” is key there. Market rates can fluctuate from day to day. Bringing the average level close to IORB also requires some tolerance for modest, temporary moves above IORB.

Those fluctuations should help, in fact, in returning the Fed’s balance sheet to an efficient size. Many banks have adapted their business models to an environment with more-than-ample reserves and market rates consistently below interest on reserves. These adaptations likely increased the aggregate demand for reserves compared with the prepandemic level. There’s a ratchet effect: To meet banks’ reserve demand in the short run, the Fed now must supply more reserves than it would have previously.

Since late summer, market rates have risen significantly relative to IORB. The rise in market rates relative to IORB will likely make many banks want to cut back on reserve holdings. There is only so much a bank can do, though, to reduce its reserve needs from one day or week to the next. It takes time to shed unwanted deposits or to adjust the composition of assets and liabilities so the bank needs less immediate liquidity. By sustaining a period of money market rates averaging close to IORB, including some modest, temporary swings higher, the FOMC can give banks the incentive to make longer-term adjustments. In turn, those adjustments will undo the ratchet effect, allowing the FOMC to meet banks’ long-run demand for reserves, rather than the temporarily higher short-run demand. This will let the Fed operate with a smaller, more efficient balance sheet over time while continuing to foster safe and efficient liquidity for the banking system.

However, it is important not to take those incentives too far. Rates averaging above IORB in a sustained way would be inefficient.

The FOMC has followed a two-part strategy to help banks adapt and bring the financial system to reserve levels that will be ample in the long run.

First, we reduced reserve supply gradually. After beginning asset runoff in 2022, we twice slowed the pace of runoff to produce a gentler glide path for aggregate reserves. The gradual decline in reserves gave banks more time to change their business models and their own balance sheets during the transition from abundant to ample reserves.

Second, the Fed has established strong tools to put a ceiling on money market rates and to provide additional liquidity if needed.

The Fed’s ceiling tools

The discount window is available every business day to healthy banks, with primary credit offered against a wide range of collateral at an interest rate equal to the top of the fed funds target range. The Federal Reserve banks continue to work to make the window more usable. We’re taking on board the public feedback we received from a formal request for information last year. The industry wants transparency about the amount a bank can borrow, prompt approval of loans and seamless movement of collateral. We hear that feedback, and we’re working to modernize our operations. For example, our Discount Window Direct tool now lets bankers request loans online, and we’re working toward allowing firms to manage collateral there, as well. Yesterday’s panel addressed the interactions between the discount window and the Federal Home Loan Banks (FHLBs). We’re collaborating with the FHLBs to streamline those connections.

The Fed’s Standing Repo Facility (SRF) also offers funding to eligible counterparties at the top of the target range against Treasury and agency mortgage-backed securities. I was disappointed to see rates on a large share of tri-party repo transactions exceed the SRF rate at times this week. I anticipate that primary dealers will use the facility to obtain repo funding when it is economical to do so. Drawing on the SRF when the rate is economical is a sound way for a primary dealer to serve the market. With rates averaging higher than they were just a few months ago, the likelihood of the SRF rate becoming economical on some days is higher. Dealers may now need to step up their readiness to access the SRF in response to rate moves.

But the way the Fed implements SRF operations can also affect counterparties’ willingness to use it. This year, the Open Market Trading Desk at the Federal Reserve Bank of New York added morning-settlement SRF operations to provide funding earlier in the day. Counterparties have responded positively to that enhancement. In my view, central clearing would further strengthen the facility’s effectiveness. The Fed’s counterparties could net down centrally cleared SRF borrowing against onward lending to other firms. Netting would reduce counterparties’ costs and risks of intermediating SRF funds to the broader market, helping liquidity to flow efficiently through the financial system. Streamlined SRF borrowing could be particularly helpful in the event of financial stress. Times of stress are the times when it’s most crucial for the Fed to be able to deliver the liquidity the system needs. Yet times of stress are also exactly when market participants try to de-risk their balance sheets by pulling back from intermediation.

Moreover, central clearing of Fed operations would align with the private-sector repo market’s ongoing shift toward central clearing. Broader central clearing of Treasury repos will support financial stability by netting down many risks and providing strong, standardized risk management and transparency for the risks that remain. The Securities and Exchange Commission (SEC) has mandated central clearing of many Treasury repos between private sector participants by June 2027. I am pleased by the industry’s progress on the legal, operational and technological work to implement this important change, and I encourage industry leaders to maintain focus on the remaining steps to complete it.

Effective ceiling tools will allow the FOMC to permit modest interest rate fluctuations without risking large upward spikes in rates.

The outlook for money market conditions and reserves

In recent months, money market rates have moved up toward and sometimes above IORB. After averaging 8 to 9 basis points below IORB in the first eight months of 2025, the tri-party general collateral rate (TGCR) averaged slightly above interest on reserves in September and October. TGCR is a rate on overnight repos collateralized by Treasury securities. It’s a safe rate in a liquid and competitive market, and I view it as the cleanest single measure of money market conditions. TGCR is now running at or above where I’d like to see relative to IORB for a long-run environment of ample reserves. That makes it appropriate to end asset runoff, as the FOMC decided to do.

It is not certain where money market spreads will go from here and how the supply of reserves will need to evolve. Temporary factors including tax payments, Treasury issuance and banks’ quarter-end and month-end balance sheet adjustments have contributed to the upward pressure on repo rates in recent months. If those pressures recede, rates could come down relative to IORB. Banks may also continue to adapt to the new configuration of rates. There might then be room for modest further decreases in reserve supply. Once runoff ends, the Fed can further reduce reserve supply by holding assets constant for a time and allowing decreases in reserves to offset trend growth in other liabilities such as currency. But if the recent rise in repo rates turns out not to be temporary, the Fed in my view would need to begin buying assets to keep reserves from falling further and maintain an ample supply of reserves.

The size and timing of these reserve management purchases should not be mechanical, in my view. While purchases will need to offset relatively predictable trend growth in currency, reserve demand will likely also change over time in response to economic growth, changes in the banking and payments businesses, and adjustments in regulations. Reserve supply will need to roughly track those developments to remain efficient. I currently view the average spread of TGCR to IORB as the central, though certainly not only, indicator of how reserve supply needs to evolve.

The FOMC’s operating target

You may have noticed that I focused on a repo rate, TGCR, and not the fed funds rate, even though the FOMC targets the fed funds rate in setting monetary policy. That was intentional. In my view, the fed funds operating target is outdated. The fed funds rate measures rates on unsecured, overnight loans between banks and other entities, such as FHLBs, that hold reserves. This rate made sense as an operating target given the way money markets worked in the 1990s, when the FOMC began publicly targeting fed funds. However, the world has changed since then in two crucial ways.

One change is the move to supplying ample, interest-bearing reserves. Because reserves now earn a competitive interest rate, banks are less motivated to lend out extra reserves at the end of the day to peers whose reserves have fallen short.

Some say the Fed should reduce reserves to the point of scarcity to revive interbank trading. But ample reserves don’t cost anything, and they make the banking system safer and more efficient, as I discussed earlier. Creating artificial scarcity just to force banks to manage it would be wasteful. Although it would be competition of sorts, it wouldn’t be the kind of competition that matters for a vibrant economy. Banks should compete to best serve their customers, not to scrounge for liquidity that the Fed can easily provide.

In addition, while banks should be prepared to source contingent liquidity when needed, they don’t need to trade every day in the fed funds market to maintain that readiness. I’ve compared this before to building codes and fire drills. Modern building codes prevent many fires. People these days get much less practice escaping burning buildings. But no one suggests returning to old-fashioned, dangerous construction methods so buildings would burn down more often, and firefighters could get more experience. Fire drills work fine! In the same way, ample reserves reduce liquidity risk, and banks can employ drills and other readiness exercises to ensure they’re prepared to address the risk that remains.

The Global Financial Crisis (GFC) demonstrated the risks to firms on both sides of an unsecured, short-term interbank loan. The lender is at risk because if the borrower defaults, there’s no collateral to fall back on. The borrower is at risk because if lenders become worried, funding can dry up suddenly. The other key change since the 1990s is that post-GFC capital and liquidity regulations recognize these risks and create disincentives for unsecured, short-term interbank lending.

Put it together, and fed funds is no longer a vibrant interbank market that measures the marginal cost of money for a wide range of institutions. Today, the fed funds market is dominated by an arbitrage trade. FHLBs, which don’t receive interest on reserves, lend to foreign banks, which receive interest on reserves but don’t pay for deposit insurance. This concentrated base of lenders and borrowers makes the fed funds market idiosyncratic and fragile. From July 2023 through mid-September of this year, the spread of the fed funds rate to IORB didn’t change by a single basis point even as money market conditions fluctuated meaningfully and other money market rates moved in response. More recently, the fed funds rate has come up only gradually despite large daily swings in repo rates. Fed funds appears to act more like a lagging average of past money market pressures than a timely indicator. In addition, any change in the FHLBs’ incentives or funding demand can cause fed funds volume to drop sharply, as occurred when FHLBs needed to meet higher advance demand during the March 2023 banking stresses.

If the connection between fed funds and broader money markets ever broke, the FOMC would need to adopt a different target. I don’t think making important changes under time pressure is the best way to promote a strong economy and financial system. As I described in a recent in-depth essay with a colleague, I believe the time has come for the FOMC to modernize the target rate. Proactively changing to a more robust target, such as TGCR, would allow time for a smooth transition so the private sector could adjust.

Let me emphasize that modernizing the target rate would not change the FOMC’s monetary policy strategy. The operating target specifies how the FOMC measures and influences monetary conditions. It’s entirely separate from the question of how restrictive or accommodative those conditions should be to achieve the FOMC’s goals.

Targeting TGCR would not require a larger Fed balance sheet or larger, more frequent or more complex operations than targeting the fed funds rate. For many years, the FOMC has set a target range for the fed funds rate that is 25 basis points wide. The Fed’s existing tools are generally effective at keeping TGCR within that target range. However, as I noted, dealers may need to increase their readiness to draw on the SRF in an environment where it is in the money more often, and the Fed should also continue to enhance the facility’s usability. In my view, the FOMC does not need pinpoint control of the operating target. Still, targeting a repo rate would require more tolerance for volatility within the target range than we have historically seen in the fed funds rate.

There are important trade-offs between potential replacements for the fed funds rate. Besides TGCR, market participants often mention the Secured Overnight Financing Rate (SOFR) as a candidate. SOFR is a widely adopted reference rate with well-developed derivatives markets, and it includes centrally cleared trades that seem likely to grow in volume in light of the SEC mandate. However, TGCR is a cleaner measure of monetary conditions because it’s not affected by the spread dealers require to intermediate between repo market segments. Although the TGCR market is smaller than the SOFR market, TGCR still incorporates more than $1 trillion in daily transactions among a wide range of counterparties.

To reiterate, these are only my views. But I believe the FOMC should modernize the target rate.

Thank you.