What are the implications of the Fed slowing down its balance sheet reduction process?
During the years of expansionary monetary policy, the Federal Reserve embarked on an asset purchase programme aimed at injecting liquidity into the economy and stimulating it, with its assets peaking at 35% of US GDP in mid-2022. The inflationary crisis required a restrictive monetary policy which included reducing the size of the central bank’s balance sheet in order to withdraw liquidity from the financial system. In 2025, the Fed announced a slowdown in the pace of its balance sheet reduction process beginning in April.
In March 2025, the Federal Reserve (Fed) announced a slowdown in the pace of its balance sheet reduction process beginning in April. This decision reflects both the fact that its reserves are close to the level considered appropriate and the need to avoid liquidity tensions in the financial markets. In this article, we assess the impact of this decision and update our projections for the Fed’s balance sheet with respect to the estimates made at the end of last year.1
- 1. See the Focus «Balance sheets: the not-so-visible normalisation of monetary policy» in the MR11/2024.
During the years of expansionary monetary policy (2019-2022), the Fed embarked on an asset purchase programme aimed at injecting liquidity and stimulating the economy, which resulted in an unprecedented accumulation of assets that peaked at 35% of US GDP in mid-2022. Thereafter, the inflationary crisis required a restrictive monetary policy, which included reducing the size of the central bank’s balance sheet in order to withdraw liquidity from the financial system. This programme, known as quantitative tightening (QT), has been implemented through a passive strategy that consists of not renewing the portfolio of Treasuries and mortgage-backed securities (MBSs) as they reach maturity.2 Since its inception in June 2022, the QT programme has been implemented in different phases, with the balance sheet being reduced at varying rates. Initially, the process was accelerated, but from mid-2024 its pace was gradually moderated. Since April 2025, the Fed has further slowed the pace of its balance sheet reduction, limiting monthly redemptions of Treasuries to 5 billion dollars, while maintaining the cap for MBSs at 35 billion, which has allowed the Fed to reduce its balance sheet to 28% of GDP. Had the pace of reduction announced in June last year been maintained, our estimates suggested that, by the end of 2025, the Fed could have reduced it to around 25% of GDP. Now, at the reduction rate in force since April, the balance sheet is set to reach a similar but slightly higher level of around 26% of GDP, and in order to reach the levels estimated previously, it would need at least an additional six months.
- 2. When the Fed purchases financial assets, it pays for them by creating new reserves that are deposited in the sellers’ accounts in the Fed, thus increasing the liquidity of the system. On its balance sheet, those assets are reflected as assets for the Fed, while the reserves are created as liabilities. When the Fed sells those assets or allows them to mature, it does not receive any cash income, but rather decreases the balance of the reserves, thus reducing liquidity.
So far, the markets have absorbed the QT and the change in its intensity without any major issues or turbulence arising. However, the main risk is the Fed withdrawing more liquidity than desired, which could lead to episodes of financial stress. The Fed’s goal is to reduce the balance sheet to a level with «ample» reserves, that is, not so abundant as in the past, but sufficient to ensure that the financial system can operate without liquidity restrictions and that the effective federal funds rate (EFFR) is not overly sensitive to daily fluctuations in the level of reserves. This equilibrium level, which is difficult to estimate accurately, is identified more by its effects than by any specific measure, so the Fed uses various indicators to assess whether the volume of liquidity is adequate.
One of these indicators is the elasticity of the EFFR with respect to reserves, that is, how much the interest rate varies in response to changes in the amount of reserves. When the elasticity is high (which, for a demand curve, it means that it becomes more negative), small changes in reserves have a greater impact on the EFFR, making it difficult for the Fed to control rates and, therefore, to implement its monetary policy. Conversely, if the elasticity is close to zero, then the EFFR hardly changes, even with large variations in reserves. Currently, we are in this second case (see second chart), with abundant reserves and liquidity needs covered. Ideally, the QT process would end when the elasticity goes from almost zero to slightly negative. However, it is difficult to identify that point, given that over time changes occur in both the demand and supply of reserves that are beyond the Fed’s control.
A second measure is the amount of reserves that non-banking financial institutions hold through overnight financing operations known as ONRRPs.3 With the start of the new cycle of rate hikes, the Fed used ONRRPs to incentivise these entities to deposit their excess liquidity and prevent them offering it in the market, which would apply downward pressure on the EFFR, making it difficult to implement monetary policy. By offering an attractive interest rate on ONRRPs, the Fed convinced them to park their liquidity, turning that rate into a lower bound for the EFFR. As the first chart shows, this has been the component of liabilities that has been reduced the most since the beginning of the QT process, falling from 2.5 trillion dollars in June 2022 to 632 billion, bringing it close to pre-pandemic levels.4 The moment non-banking institutions are left with no excess liquidity, the ONRRPs will be reduced to zero, and the QT process will drain reserves directly from the banks, which could begin to shift the elasticity of the reserves from zero into negative territory.
Based on these indicators, the Fed has suggested that total reserves of around 10%-11% of GDP5 would be a good target level, and they are currently at around 13%. If we assume that the Fed maintains the current rate of reduction, then its reserves would reach 10%-11% beginning in Q2 2026. Therefore, in the absence of any frictions in liquidity, the Fed has the green light to continue to reduce its balance sheet over the next year.
- 3. ONRRPs (Overnight Reverse Repurchase Agreements) are one-day agreements in which the Fed borrows money by offering securities as collateral in order to reduce liquidity in the financial system.
- 4. The other liability accounts have remained relatively stable: bank reserves held by banks in deposits in the Fed went from 3.1 trillion dollars to 3 trillion, cash in circulation (notes and coins) actually increased slightly from 2.3 trillion to 2.4 trillion; and the US Treasury General Account (TGA) in the Fed went from 0.8 trillion dollars to 0.7 trillion.
- 5. Christopher Waller. «A Conversation with Federal Reserve Governor Christopher Waller». The Brookings Institute, 2024.
As the Fed approaches the desirable level of reserves but, as we have seen, in the absence of any precise measure to know when to halt QT, it prefers to slow down so as not to subject the financial system to an episode of liquidity stress. Fed Chair Powell has described it as being like a plane slowing down as it comes in to land, in order to reach the runway on target. Moreover, such prudence is even more important in the current context.
The sovereign debt markets in the US have been the protagonists of high volatility due to the uncertainty surrounding the Trump administration’s economic policy. The erratic implementation of tariffs has led to see-saw movements in sovereign yields. Furthermore, the doubts surrounding fiscal sustainability, with a deficit close to 7% of GDP and with no sign of it being materially corrected any time soon, have put additional pressure on sovereign yields. The impact of this situation is already apparent, with a weak auction of 20-year securities in mid-May. A sudden withdrawal of the Fed as a major buyer of Treasuries poses an additional risk to financial stability, and while the Fed does not explicitly acknowledge this, it will certainly want to avoid it.
- 1. See the Focus «Balance sheets: the not-so-visible normalisation of monetary policy» in the MR11/2024.
- 2. When the Fed purchases financial assets, it pays for them by creating new reserves that are deposited in the sellers’ accounts in the Fed, thus increasing the liquidity of the system. On its balance sheet, those assets are reflected as assets for the Fed, while the reserves are created as liabilities. When the Fed sells those assets or allows them to mature, it does not receive any cash income, but rather decreases the balance of the reserves, thus reducing liquidity.
- 3. ONRRPs (Overnight Reverse Repurchase Agreements) are one-day agreements in which the Fed borrows money by offering securities as collateral in order to reduce liquidity in the financial system.
- 4. The other liability accounts have remained relatively stable: bank reserves held by banks in deposits in the Fed went from 3.1 trillion dollars to 3 trillion, cash in circulation (notes and coins) actually increased slightly from 2.3 trillion to 2.4 trillion; and the US Treasury General Account (TGA) in the Fed went from 0.8 trillion dollars to 0.7 trillion.
- 5. Christopher Waller. «A Conversation with Federal Reserve Governor Christopher Waller». The Brookings Institute, 2024.
