Reversing the print machine

What is Quantitative Tightening, how does it work and what are its potential effects

Bottom line

After bloating its balance sheet to ~$9tn following the pandemic-related measures, the Fed intends to reduce its size, also known as quantitative tightening (QT). However, the increasing risks of stagflation combined with geopolitical tensions are leaving the Fed with little flexibility to operate without disrupting the financial markets and the borrowing ability of the U.S. government.

The Treasury Quarterly Funding Announcement on May 2nd, along with the potential release of the FOMC QT schedule on May 4th, should give more clarity to investors regarding the supply of assets, influencing risk premium and asset prices.

Executive Summary

All Dollars are not created equals

  • The Fed does not print actual money. It only controls the level of Reserves, which are the primary levers for its quantitative policies.

  • Reserves are different from Deposits and do not necessarily move in tandem.

  • Quantitative policies modify the balance sheet of the various financial actors and have the potential to change the risk premium for all asset classes.

An unsuccessful precedent and different macro conditions

  • The first and only QT experiment started in late 2017, subsequently to the various QE programs used to counter and smooth the 2008 financial crisis.

  • This first QT quickly threatened to jeopardize financial liquidity and forced the Fed to restart a “technical” QE as early as 2019. The Fed managed to reduce only 15% of its balance sheet during this first QT episode.

  • The current macroeconomic conditions are very different, notably for the rising risks of stagflation. The shape of the rates' curves, high inflation, higher debt, and geopolitical uncertainties lessen the chances for a smooth QT.

Multiple options, multiple outcomes

  • There are various methods for the Fed to shrink its balance sheet, but the end result is the same: an increase in net debt supply with a reduction of Reserves, forcing the private sector to step in and change its asset composition.

  • The Treasury Department and the Fed need to work together to implement a smooth QT.

  • The march FOMC minutes provided some hints regarding the intended size of QT. The schedule of the Treasury issuance will be particularly scrutinized to gauge the effect of QT on liquidity, money tightness, and, eventually, risk premiums.

All Dollars are not created equals

There are two types of money: one when transacting with the Fed and another when transacting with everyone else. More precisely, the Fed only deals with financial institutions that have an account with it, such as the Treasury Department, a primary dealer, or selected banks, hereafter referred to as Banks. The other financial players, such as smaller banks, pension funds, or hedge funds, collectively known as Non-Bank Entities (NBE), do not have a direct account at the Fed.

To better understand the various forms of money, let us first look at the balance sheet composition of the Fed and the Banks.

Fed’s Balance Sheet

The asset side of the Fed’s balance sheet is mainly composed of gold, foreign currencies, various sorts of U.S. debt (e.g., Treasuries notes and bills, mortgage-backed securities), and loans to its member banks.

The liability side of the Fed’s balance sheet is made of the total (paper) currency in circulation. Additionally, like a regular bank, the money sitting in the account of its direct clients (Banks), known as Bank Reserves, is also part of the Fed’s liabilities. Reserves today account for about 70% ($6.2tn) and currency for about 25% ($2.2tn) of the Fed’s total liabilities. These reserves can only be used to transact with other entities having their account at the Fed, i.e., other Banks.

These reserves sit mainly in three types of accounts:

  1. Member Banks' Accounts ($3.8tn);
  2. Treasury General Account ($0.6tn);
  3. Reverse Repo Program ($1.8tn).

The Fed controls reserves. They can only be exchanged between entities with an account at the Fed and cannot leave the Fed’s balance sheet. Thus, bank reserves are living in a closed system. The Fed determines the total amount of reserves in one of the three main accounts by creating (Quantitative Easing) or destroying (Quantitative Tightening) them.

Like Banks, the Treasury Department has its checking account equivalent at the Fed, known as the Treasury General Account (TGA). It is the account that receives tax payments or the proceeds from the sale of U.S. debt. It is also the account used to fund government spending. This chart shows the effect of the massive sale of U.S. debt and the ensuing distribution via the government’s checks during the pandemic.

Finally, the third main account type is the Reverse Repurchase Program (RRP) account used by market participants for their repo transactions with the Fed. A Repurchase Agreement (repo) is a secured money market transaction. A market participant in need of cash can sell an asset, typically a U.S. debt, to the Fed, agreeing to repurchase it the next day for a fee (repo rate). Given the massive supply of debt during the Covid-19 crisis and the lack of alternatives to use it as collateral, the RRP increased significantly.

Thus, transactions between the Fed and its partnered Banks are only made with a type of money known as Reserves, which, added to the currency in circulation, represent the monetary base. However, while denominated in U.S. dollars, reserves cannot be used to buy goods and services and do not directly affect CPI. Thus, in anticipation of what we will discuss later, the creation of Reserves to purchase assets, also known as Quantitative Easing, did not directly impact CPI. 

Banks’ Balance Sheet

The asset side of the balance sheet of a commercial bank with a Fed account comprises regular bank assets (e.g., cash, securities, loans, real estate) and Bank reserves that sit in its Fed account. On the liabilities side, there are the bank deposits, which are the money of its customers, i.e., NBEs. These U.S. dollar-denominated bank deposits reflect, roughly speaking, the actual money chasing goods and services. Transactions involving NBEs are done via bank deposits, which are very close to M2 broad money. They are created by commercial banks when they make loans or purchase assets. For example, when somebody takes a loan to purchase a house, the bank credits their account with a new deposit, which will be used to make the transaction with the seller. Once the transaction settles, the seller’s account will eventually be credited with that deposit. Thus, the commercial bank created money out of thin air. Therefore, deposits are heavily linked to bank credit as well.

Reserves vs. Deposits

Hence, while both the Bank Reserve and the Bank Deposit are measured in USD, these are different types of dollars. The former is created by the Fed, cannot buy any goods and services, and stays within the Fed-Banks eco-system, while the latter is created by the commercial banks and can flow into the real economy.

Therefore, the Fed does not print money but only creates and destroys Bank Reserves, a form of money used within the Fed-Bank system. Thus, the central bank does not directly control or print broad money, the currency used for chasing goods and services (represented by M2). Deposits are created by banks when they buy assets or grant loans. Therefore, commercial banks are the actual money printers, while the Fed only prints reserves.

Many wrongly assume that Bank Reserves and Deposits are moving in tandem. For example, the central bank can decide to engage in Quantitative Easing and buy trillions of securities via creating bank reserves. But suppose the Banks decide that the economy is not strong enough, and people will likely default on their loans as private companies looking to make profits. In that case, they may grant no loans at all and thus not "transmit" the Fed's impulse to the rest of the economy.

There are periods when Bank Reserves go down while Deposits continue their uptrend.


Transactions between participants do not have the same impact on changes in reserves or deposits. For example, let us analyze how M2 behaves when the Treasury issues new debts. If the buyers of the newly issued Treasuries are the Banks, M2 will likely increase, but if the final buyers are NBEs, M2 does not change. Indeed, when the Banks purchase a U.S. debt from the Treasury, only an accounting entry within the Fed happens. Funds are going from the Banks’ reserve account to the Treasury’s TGA account, not affecting M2. Then, the Treasury might spend the funds received from the sale into the real economy (e.g., pay for salaries), which eventually increases M2 via new bank deposits. However, when the NBEs purchase a U.S. debt, they use bank deposits, effectively reducing M2. As the Treasury eventually spends back the proceeds in the real economy, the money goes back into deposits leading to an increase in M2. Thus the net effect on M2 is null.

Besides, as NBEs encompass a very large category of financial actors, we can further distinguish these entities between the high-velocity NBEs, such as retail investors, and the low-velocity NBEs, such as pension funds. When the former purchase a Treasury, the disposable amount of money left for consumption gets smaller, reducing consumer demand and consequently demand-side inflation. As we will explain later, this is one of the mechanisms through which QT could reduce inflation. 

Quantitative policies

In this paragraph, we first explain how quantitative policies modify bank reserves and deposits and the balance sheet of the various financial actors and then how they affect risk premium.

To support financial markets during and after the 2008 crisis, the Fed engaged in Quantitative Easing (QE), an unconventional policy measure at the time, based on buying debt securities. Once Fed Fund interest rates hit zero, the Fed resorted to QE to further ease financial conditions, mimicking what Japan did in 2001. The Fed did three rounds of QEs from 2008 to 2014 to continue supporting the recovery, expanding its balance sheet by $3.5tn to $4.5tn. 

To buy these securities, the Fed creates reserves for the seller of the securities, thus expanding its balance sheet. The Fed takes these securities as assets and credits the selling Bank with reserves, which are now a new Fed’s liability. This process is similar to a commercial bank that increases the bank deposit of a customer buying a new house with a bank loan. The bank has the house as an asset and the newly created bank deposit (from the loan) as a liability.

In the below graph, we single out the currency component of the Base Money. The decoupling between the currency and the base money clearly shows the digitally printed dollars used for QE, i.e., the created bank reserves. Between 2008 and 2014, the expansion happened in three distinct waves.

Quantitative Easing also modifies the balance sheet composition of other financial players. Indeed, let us assume that the central bank is buying debt securities from an NBE. First, since the central bank only deals with Banks and not with NBEs, the transaction starts with a Bank purchasing the debt security from the NBE and crediting the latter with deposits. Then, the Bank sells the debt security to the central bank, which credits the Bank with newly created reserves. Thus, instead of having debt security as an asset, the NBE now has deposits. Hence, QE changes the private sector assets composition and helps increase the availability of money and credit. 

In the chart below, the blue line (Banks’ cash) above the orange line (Banks’ Assets minus Liabilities) shows a measure of the banks’ “excess cash”, driven by the QE rounds, and somewhat explains the easing of monetary conditions. By and large, we observe that when the excess cash is positive (negative), monetary conditions are easy (tight).

Quantitative Tightening (QT) is a contractionary monetary policy that reverses the impact of QE by shrinking the Fed’s balance sheet, eventually removing liquidity from the system. The Fed aims to normalize its monetary policy by reducing the level of reserves created during QE by either receiving principal repayments and not rolling them or actively selling its assets. By reducing reserves from the system, the Fed implicitly forces a reshuffle of the private sector balance sheet towards holding more Treasuries.

Through regulatory mechanisms, the private financial sector is forced to absorb what the Fed will not buy (or even sell), i.e., more Treasuries and Mortgage-Backed Securities (MBS) as the counterpart to owning fewer reserves. Banks will own more bonds and fewer reserves, while NBEs will own more bonds and fewer bank deposits. Excluding the current excess cash in the system, the private sector would need to sell other assets to make room for this adjustment, a headwind for assets along the risk premium curve.

In practice, QT shrinks the balance sheets of the Fed and the Banks, but a priori leaves unchanged (in terms of size) those of the Treasury Department and NBEs. Indeed, any form of QT will reduce the Fed’s reserve liabilities (its goal) and, consequently, the Banks’ reserve assets. However, from the Treasury Department’s side, nothing changes in the size of the balance sheet. The change is in the ownership composition, i.e., the Fed is replaced by the private sector, which buys Treasuries with bank deposits. Consequently, the balance sheet of NBEs is not affected in terms of size but only in terms of asset composition.

Quantitative policies’ effects on risk premiums

The risk premium is the return over a non-duration safe asset (e.g., cash) one is expected to be compensated for by investing in a financial asset, i.e., taking a market risk. In other words, by investing in risky financial assets and refraining from consuming today, investors expect to earn more than the cash return. Practically, the risk premium is a function of the availability of money and credit against the supply of assets. When the former is ample and the supply of assets is low, risk premiums are low, but when money and credit are tight, and the supply of assets is high, risk premiums are high.

First, during the QE, the Fed became a significant buyer at any price and bought a lot of debt by creating reserves, driving down risk premium, and resulting in a tailwind for risk assets. During QT, the Fed can be considered a net seller of assets, potentially driving up risk premiums and creating a headwind for risk assets. Indeed, without the Fed as a buyer, someone else will have to step in to buy the debt security and will certainly require a higher risk premium than the Fed would have, as the latter was insensitive to the risk premium. Thus, quantitative policies ultimately affect risk premium by modifying the demand/supply of assets.

Besides, the risk premium could also be affected by the choices of the Treasury in terms of its issuance program. For example, we assume that through RRP, there is enough demand for bills; hence no higher yield will be demanded by investors should the Treasury issue very short-term debt. However, investors will require a higher return from the Treasury to buy long-term debt, i.e., higher yields, as they will have to change their asset composition. As this yield increases to satisfy buyers’ demand, risk premia across other assets also expand, as NBEs will need to sell other assets to make room to buy the newly issued Treasuries to comply with regulatory requirements.

QT will reduce deposits, with the NBEs forced to buy Treasuries, and so, with less availability of cash, it may ultimately reduce consumer demand. It is then a tool to fight demand-side inflation, but in our opinion, it is a weak tool, similar to QE, which was a tool that could have massively increased inflation but did not.

An unsuccessful precedent and different macro conditions

The first Quantitative Tightening

After several rounds of QE, low inflation, a healthier job market, and five hikes, the Fed started to normalize its balance sheet in 2018 by letting its debt assets mature and eliminating the reserves created to buy them. The initial market reaction was negative, with a rapid drop of almost 15%. During the December 2018 FOMC press conference, Fed Chairman Jerome Powell confirmed QT was on auto-pilot, triggering another leg down in the market. He pivoted a few days later, hinting at a data-dependent Fed, relieving markets.

As QT continued, the repo rate, usually close to the Fed fund rate, spiked to 10% in September 2019. The real reason behind this spike is still debated. Some explain that the cash withdrawal to pay corporate taxes and the strong demand for cash to settle newly issued Treasuries created a cash shortage. Thus, participants who needed cash to meet regulatory requirements were paying a considerable premium. The Fed immediately injected $75bn into the repo market to support it. This event practically halted QT and forced the Fed to start a “technical QE” that continued until Covid-19 hit the financial markets, obliging the Fed to resume “official” QE. In other words, QT might have caused this cash shortage, and its consequences were not fully appreciated.

Market differences with 2018

There are many differences between the state of the markets and macroeconomy in 2018 and now. However, let us focus only on the crucial ones.

First, at the onset of the first QT, the curve was much steeper than now. Currently, the flat-to-inverted shape of the curve is reminiscent of late-cycle tightening, which suggests an economic contraction is within sight. Hence, a further tightening of financial conditions may lead to a recession. The Fed Funds or Eurodollar futures are even pricing a rate cut during the second semester of 2023.

Second, in 2018, the Fed was concerned with disinflation. Today, inflation measured via the CPI is at a 40-years high. This directly affects the Fed’s policy because of its mandate on price stability. They are thus more focused on fighting inflation than anything else. Three successive waves have caused post-pandemic inflation. The first was caused by the lack of supply related to the supply chain disruptions. Then, strong consumer demand from the high disposable income from Covid-19 aids fueled higher goods prices. During these first two waves, higher inflation was witnessed primarily for consumer goods. Note that the latter were in deflation since the mid-‘90s and acted thus as a natural brake for global inflation. Finally, as the Covid-19-related disruptions were vanishing, the eruption of war in Eastern Europe brought the third wave, driving up the price of food and energy, whose impacts on inflation and growth are a much more significant concern.

The current level of debt ($29tn) makes the sustainability of higher rates a real issue for the U.S. government. It directly affects its ability to serve its debt in the public market. Current interest payments amount to more than $550bn. With higher rates, the cost of the new debt issued by the Treasury will massively increase, and more issuance will be needed to finance it. However, as long as real interest rates remain negative, debt servicing should not be a concern.

Post-pandemic dynamics

The Covid-19 crisis increased the federal debt by more than $5tn, and the Fed’s balance sheet expanded by a similar amount. Thus, the Fed bought all the pandemic-related U.S. debt by and large. Since M2 also increased post-Covid, we can assume that some funds stayed at the TGA or went to NBEs, which either purchased assets or kept them as deposits. The Treasury then decided to use the TGA “excess cash” to fund federal spending, reducing subsequent debt issuance and increasing M2.

The lack of newly issued debt combined with the massive increase in savings from the private sector drove the overnight repo rate below zero, hindering Fed’s monetary policies. In other words, there was so much cash and nowhere to invest it that the return on the cash went below the Federal Funds rate. Therefore, the Fed created the RRP to help steer these “excess savings”. The RRP started around March 2021 and has now close to $1.7tn, which can help smooth any QT as it will chase any new short-term debt offering a yield, as we shall analyze later.

Multiple options, multiple outcomes

First, the natural supplier of debt is the Treasury. According to the government needs, tax receipts, maturing debts, interest rates, and many other considerations, the Treasury decides on the new debt's size and maturity. This will have a specific effect on the QT implementation. For example, the treasury can use the cash in the TGA Account to reimburse the Fed on its debt holdings. In that case, no new debt is even issued. On the contrary, it can estimate that with current market conditions and the potential to have much higher rates in the future, it is better to issue 30-year bonds and lock historically-low interest rates. With the Fed reducing its balance sheet, only the private sector would buy such bonds, but only after selling other assets.

Second, the Fed, which ceased to be a net buyer in March 2022, is now reducing the size of its balance sheet. It can do so by receiving the principal repaid at expiry on a bond it holds and simply "burning" that money by not reinvesting it in another bond and reducing the corresponding amount of Reserves in its liabilities.

Passively maturing or actively selling

While the Fed has various degrees of control over the engineering of QT, it all sums up to stop being a buyer and potentially become a seller of assets in the market. But as we observed with the previous QT episode, the Fed doesn't know how low the reserves can fall before a negative impact is felt in the real economy. A more aggressive QT would consist in actively selling its assets, hence forcing the price down. The treasury can also contribute to an aggressive QT by issuing long-duration securities, where RRP cannot help. While the mechanism is complex, we could reasonably deduct that the “extra” money sitting on the RRP is chasing short-term securities. Thus, it would not play a role if the treasury opted to issue long-duration bonds. A softer QT approach would see the Fed and the Treasury coordinating their actions to limit the net new supply of U.S. debt. The Fed would not actively sell assets, while the Treasury would reduce new debt issuance. 

The last FOMC minutes indicate the plan is to reduce the balance sheet by more than $1tn a year, or about $95bn a month. According to the available details, the balance sheet would shrink at a maximum monthly pace of $60bn in Treasuries and $35bn in MBS, almost double the rate of the previous QT episode. However, policymakers would attain this pace in steps over three months or longer, depending on market conditions. In other words, they can start by reducing the balance sheet by $35bn, then $45bn, and then $60bn per month.

Looking at the current debt holding of the Fed, the SOMA (System Open Market Account) Holdings, we can assess if the Fed intends to passively mature or actively sell. We first observe that more than half of the debt is due in three years.

Next, we also observe that the Fed should not actively sell its holdings for the following months since more than $60bn is maturing, and they could re-invest less.

On the other hand, regarding MBS, which are long-duration debt, the Fed should rather sell them. The Fed had bought about $1.4tn of MBS during the pandemic. Thus, it would take about three years for the Fed to return to its pre-pandemic level. Actively selling MBS may also reduce housing market inflation (currently at 5%), accounting for about one-third of the CPI.

Mortgage rates recently increased significantly, preventing the housing market from overheating. When the Fed starts selling MBS, mortgage rates are likely to continue their ascension. Again there is a risk that, if too high, mortgage rates would mute the demand and potentially cause a housing market crash. This would not be politically acceptable considering how housing is one of the two main pillars of American wealth, along with 401(k).

Treasury’s collaboration and Yield Curve control

With the 1y forward yield curve already inverted and short-term interest rates rising, the margin for maneuver of both the Treasury and the Fed is limited to avoid a market meltdown. The choice of the size and the maturity distribution for the Treasury to issue new debt and the Fed's intention to re-invest its bond proceeds could heavily influence the yield curve and risk premium.

As QE aims to ease financial conditions by lowering yields and providing ample liquidity to the market, QT, especially if coupled with a net new supply of debt, can tighten financial conditions, make yields rise, and remove liquidity. Higher rates on the long end of the curve are somewhat reasonable to get a steeper curve. Still, they are alarming given the high level of debt and potentially positive real rates if long-term inflation expectations stay anchored.

Thus, should the Treasury decide to issue short-term debt, there is enough liquidity within the RRP to absorb ~$1tn of new supply, and the potential negative effect on other assets would be contained. Therefore, the actions of the Treasury will impact asset prices as much as the actions of the Fed. Suppose the Treasury decides to soften the effect of QT on asset prices by carefully choosing the size and maturity of new debt issuance. In that case, the impact of such a QT on inflation will also be lower. On the other hand, should the Treasury decide to fight inflation too, it would expand the negative effect of QT on asset prices.


  • Strong collaboration between the Fed and the Treasury.  If the Treasury coordinates its supply with the Fed’s QT intentions, it will be easier for the private sector to step in and allow for a smooth balance sheet unwinding by the Fed. 

  • Adverse Event. An adverse event can disrupt the Fed’s ability to reduce its balance sheet. Wars or issues in financial stability could even force the Fed to reverse its course and expand its balance sheet. 

  • Strong Demand for U.S. debt. If the demand for U.S. debt increases thanks to its relatively higher returns, the Fed could shrink its assets more efficiently by selling to willing buyers without affecting the markets 


  • Inflation not going down. Should inflation remain high, the Fed would need to be extremely restrictive to fight it, actively selling its long-term assets, pushing down prices, and ultimately destroying wealth. 

  • QT produces unexpected effects. Like the previous attempt, QT can engender a black swan with unforeseen consequences on the market or the economy that could force the Fed to reverse its course.

  • Yields rise.  If yields are rising, a QT may exacerbate this rise. A too fast and substantial rise in yields could push down both the stock and housing markets. Moreover, the debt cost would potentially significantly increase. 



This report has been produced by the organizational unit responsible for investment research (Research unit) of atonra Partners and sent to you by the company sales representatives.

As an internationally active company, atonra Partners SA may be subject to a number of provisions in drawing up and distributing its investment research documents. These regulations include the Directives on the Independence of Financial Research issued by the Swiss Bankers Association. Although atonra Partners SA believes that the information provided in this document is based on reliable sources, it cannot assume responsibility for the quality, correctness, timeliness or completeness of the information contained in this report.

The information contained in these publications is exclusively intended for a client base consisting of professionals or qualified investors. It is sent to you by way of information and cannot be divulged to a third party without the prior consent of atonra Partners. While all reasonable effort has been made to ensure that the information contained is not untrue or misleading at the time of publication, no representation is made as to its accuracy or completeness and it should not be relied upon as such.

Past performance is not indicative or a guarantee of future results. Investment losses may occur, and investors could lose some or all of their investment. Any indices cited herein are provided only as examples of general market performance and no index is directly comparable to the past or future performance of the Certificate.

It should not be assumed that the Certificate will invest in any specific securities that comprise any index, nor should it be understood to mean that there is a correlation between the Certificate’s returns and any index returns.

Any material provided to you is intended only for discussion purposes and is not intended as an offer or solicitation with respect to the purchase or sale of any security and should not be relied upon by you in evaluating the merits of investing inany securities.