## 1. ECB Balance Sheet and International QT Experiences ## Source details **Canonical URL:** [1. ECB Balance Sheet and International QT Experiences](https://www.imf.org/-/media/files/publications/cr/2023/english/1eurea2023002.pdf) ## Other formats - [Markdown version](/-/media/files/publications/cr/2023/english/1eurea2023002.pdf.md) - [Structured JSON version](/-/media/files/publications/cr/2023/english/1eurea2023002.pdf.json) --- ### Executive Summary - The ECB Governing Council (GC) initiated quantitative tightening (QT) to reduce the Eurosystem stock of securities holdings accumulated since January 2015 through quantitative easing (QE). - QE supported aggregate demand and transmission by compressing bond yields; QT aims to remove monetary accommodation, alleviate QE side effects, reduce the Eurosystem’s footprint in financial markets, and ensure issuer limits do not hinder the ability to restart large-scale asset purchases. - QT should play a limited role in calibrating the monetary stance; the policy rate should remain the primary instrument because: - Considerable uncertainty exists about QT’s effects on yields and financial conditions. - QT could increase fragmentation in euro area debt markets and interfere with monetary transmission. - A relatively fast pace of QT would reduce the necessary conventional policy space from a monetary risk management perspective. - Recommended QT design features: - Pre-announcing a QT pace and dates to reassess speed to allow market participants to adjust expectations and smooth price adjustment of bonds. - Allowing deviation from the capital key over a reasonable time horizon as a result of passive QT. - Embedding an “escape clause” to be activated in case of severe market dysfunction creating a severe impediment to smooth transmission of monetary policy. ### A. Context - The ECB initiated asset purchase programs in January 2015 (APP) and March 2020 (PEPP), amounting to €4.9 trillion by end 2022, corresponding to 60 percent of euro area GDP. - With inflation rising since 2021, the ECB halted QE in June 2022 and, in July 2022, raised rates for the first time since July 2011. - After cumulative rate increases of 300bps, the GC began normalizing the Eurosystem’s monetary securities holdings through a measured and predictable pace starting March 1, 2023. - Between March and June 2023, the Eurosystem’s securities holdings were being reduced by a monthly average of €15 billion. - Starting July 2023, the GC planned to stop all APP reinvestments, increasing the pace of APP portfolio reduction. By end 2023, the GC planned to review its operational framework to provide information regarding the size of the ECB’s balance sheet in a steady state. - Unwinding the QE portfolio requires understanding its past impact and associated risks: large holdings depressed sovereign yields, contributed to disappearance of the unsecured interbank market, and reduced availability of marketable safe assets. ### B. Rationale for QT - QE influences long-term yields via signaling and portfolio rebalancing channels: - Signaling: announcements act as forward guidance on the expected path of short-term interest rates. - Portfolio rebalancing: asset purchases compress the term premium by reducing net supply of long-term assets held by the private sector, working through limited substitutability (“preferred habitat theory”) or duration risk extraction. - Estimated magnitude of QE’s effect on yields in the euro area: - Studies suggest yields fall on average 3.6 to 5.4bp per €100bn of asset purchases. - Applying these elasticities to the €4.9 trillion stock of QE purchases implies the level of accommodation to 10-year GDP-weighted zero-coupon yields of the four largest euro area jurisdictions is on average between 176 and 265bp. - These estimates carry large uncertainty and are state contingent; effects can be much higher in periods of market stress. - QE’s side-effects on safe securities and collateral: - €4.3 trillion in public sector bond purchases between June 2015 and June 2022 reduced the share of sovereign free float to 40 percent out of €10.3 trillion in outstanding euro area sovereign bonds by end 2022. - German Bunds’ free float dropped to 20 percent of issued stock by end 2022 (threefold decline), while Italy, France, and Spain fell to 56, 51, and 47 percent, respectively. - Locked up government securities peaked at €542 billion in Q3 2021 due to Eurosystem refinancing operations. - A scarcity measure based on the difference between the yield of an asset and the similar asset swap rate confirms increased scarcity, particularly with PEPP during 2020-22; the average asset swap spread in recent months remains almost double the pre-QE average. - QE and interbank market effects: - Asset purchases created abundant bank reserves, shifting the unsecured overnight lending rate close to the Deposit Facility Rate (DFR) and moving from a corridor to a floor system. - The resulting high level of excess reserves reduced the relevance of unsecured interbank transactions and may make exit more difficult. - Financial and political implications of large asset holdings: - Rising policy rates created an interest rate mismatch producing profit losses in the Eurosystem because reserves are remunerated at the DFR, which increased by 400bps from July 2022 to June 2023, while assets earn fixed coupons. - Losses in 2022 were offset by releases from risk provisions and buffers accumulated from profits over 2012-2021; over time, losses will decline as assets mature and reinvested income rises. - Temporary losses have halted transfers to fiscal authorities and can raise political scrutiny, though losses do not impair monetary policy conduct if the central bank’s commitment to the inflation target remains credible. ### C. Policy Considerations Associated with QT - Uncertain effects of QT on financial conditions: - Evidence on QT’s impact on yields is sparse; the Fed’s 2017–2019 balance sheet reduction of $640 billion (about 2.8 percent of GDP) at a monthly average pace of $28 billion produced mixed estimates. - Some studies found no statistically significant impact on the 10-year Treasury yield, while others associate a reduction of the Fed’s balance sheet by 1 percent of GDP with an increase in 10-year yields equivalent to a 6-10bp increase of the Fed Funds rate. - State contingency of QT effects: - Like QE, QT effects are likely state contingent and could be large during periods of market dysfunction. - If implemented when markets are calm and monetary policy is not constrained, QT’s impacts may be mild. - The signaling channel of QT differs from QE; QT does not imply forward guidance on keeping interest rates near the effective lower bound and could instead convey information about the pace of rate hikes and the terminal interest rate needed to achieve the inflation target. - QT “surprises” have larger effects when markets are turbulent; evidence suggests the interest-rate equivalent of QT can be almost three times higher during crisis periods. ### Projected immediate impact on yields and policy-rate equivalence - Using the QE elasticity symmetrically for QT and assuming absence of non-linearity, the ongoing monthly QT pace of €15 billion until the end of June and doubling thereafter would translate into an increase in the term premia of 12-18bp over the next 12 months. - Extrapolating impacts on policy rates identified for the Fed translates into a substitution with the policy rate in the range of 15 to 22bp. - This substitution is small relative to the cumulative 400bps increase in the ECB’s deposit facility rate between July 2022 and June 2023. - Much of the faster-than-expected balance-sheet rundown compared to market expectations in December 2021 has been gradually priced into the yield curve, implying a likely small marginal impact going forward. ### Risks that could amplify impacts and fragmentation concerns - QT may have larger impacts than currently predicted if marginal buyers are not forthcoming at a time of large sovereign financing needs; increased bond yields may reflect expectations about future bond net supply. - For the first time in many years, euro area governments face lower ECB support in rolling over maturing debt, raising the question which actors will absorb the net supply and at what spread levels. - Euro area banks are a natural candidate to step in given historical appetite, with current holdings below historical highs in most countries. - QE historically helped reduce fragmentation; unwinding that support could risk reintroducing fragmentation if QT is not predictable and cautious. - The PEPP’s flexible reinvestments and, if needed, the Transmission Protection Instrument (TPI) are highlighted as defenses against fragmentation; sovereign solvency concerns should be handled via Outright Monetary Transactions and ESM macro-adjustment programs. ### Market-functioning and crisis tools - Tools designed to support market liquidity should be used to address severe financial market dysfunction (central bank lending facilities, reverse auctions, outright asset purchases). - The separation principle is recommended: monetary policy tools (policy rates and QE/QT) should be distinct from emergency support tools aimed at market functioning or financial stability. - The pace of balance-sheet reduction could be reassessed in stress episodes; targeted interventions may accompany passive QT as seen in prior episodes (example: Bank of England during the LDI crisis). ### QT and interaction with the policy-rate path - If QT effects on the monetary stance remain small, there is no tradeoff between rebuilding conventional and unconventional policy space. - The QT objective of supporting the monetary stance is state dependent: desirable during tightening cycles; weaker case during loosening cycles ceteris paribus. - Short-term policy rates should remain the primary tool to fine tune the monetary stance; QT should proceed gradually and predictably. - Since the start of rate hikes, long-term yields have responded almost one-to-one with expected short rate hikes, indicating transmission from short-term to long-term rates has not been impaired by Eurosystem securities holdings. ### Proposed QT implementation and decision-tree - The current ECB approach—gradual and predictable unwinding with passive runoff of the APP portfolio—is appropriate to learn about yield impacts and monitor transmission risks. - Active outright sales are not warranted given the relatively shorter average maturity of the ECB’s asset portfolio and the modest initial roll-off. - Decision-tree branches (past learning phase): - Tightening cycle—QT can serve all objectives. If impact on yields remains small and balance sheet far from desired endpoint, the ECB could increase the pace to reduce its footprint. With APP reinvestments stopped, reducing PEPP reinvestment and later ceasing it fully could lead to a roll-off of €624 billion per year, with an expected impact on yields in the range of 22-33bp. - Loosening cycle—If the balance sheet remains far above the desirable endpoint and QT impact is minimal, QT could continue; if QT significantly tightens financial conditions, QT should pause in a loosening cycle to avoid increasing ELB risk. ### Estimates of terminal size, pace required, and reserve considerations - Beyond minimum reserve requirements and currency in circulation (about 13 percent of 2022 GDP or €1.7 trillion), demand for bank reserves has increased since 2009 for precautionary and prudential reasons. - Average excess reserves measures (back-of-the-envelope): - Average excess reserves after the GFC and prior to QT (2010-2022): €1.1 trillion. - Average throughout the QE period (2015-2022): €1.7 trillion. - More recent average (2017-2022, covering the floor system period): €2.1 trillion. - These imply a total Eurosystem balance sheet of between €3 and 4 trillion or 20-30 percent of 2022 GDP as a potential steady-state range. - Reaching a balance sheet of that size over the next 7 years would mean an annual reduction of €554-694 billion, corresponding to at least three times the current monthly QT pace. - If continued beyond June 2023, the current QT pace would keep the Eurosystem balance sheet twice as large as the Fed’s or the Bank of England by end 2024. - The ECB 2021 estimated that the floor required excess liquidity (FREL) could be somewhere between €400 billion and €1 trillion, implying a lower bound of the terminal balance sheet compatible with a floor system between €2.2 and €2.8 trillion. ### Operational guidance and principles for QT communication - Continue a cautious, observable approach to ascertain the appropriate balance sheet size and decide to slow or stop QT by observing the reserve satiation point. - Principles for the QT roadmap: - Early communication of QT principles to make the process predictable and eventually “invisible” even as market conditions change. - Pre-announcing regular but infrequent dates at which the speed of QT is re-assessed and pre-announcing any changes to the QT pace so market participants can adjust pricing. - Embedding an “escape clause” in the QT setup to permit pausing the program in the event of a severe and systemic shock that impedes smooth transmission of monetary policy. ### International experience (selected lessons) - Major central banks have typically initiated QT by partially or fully ceasing reinvestments (passive QT) or by selling securities ahead of maturity (active QT); approaches differ by pace, type, and maturity structure of holdings. - Fed: passive QT via caps on reinvestment (US$30 billion per month for treasuries and US$17.5 billion per month for MBS initially in May 2022), with caps doubled after three months and caps expected to remain in place over the medium term. - BoE: ceased reinvesting maturing gilts in February 2022 and initiated sale of non-financial investment-grade corporate bonds; communicated gradual and predictable QT with discretion to alter plans and temporarily reversed QT during severe dysfunction in October 2022. - RBNZ and RBA: examples of managed sales to Treasury or stopping reinvestments; both emphasized gradual, predictable approaches and reserved the right to change pace or halt sales if conditions change. - Commonalities: pre-announced QT start months in advance, emphasis on predictability, gradualism, and discretion to adjust QT in response to financial and economic conditions; QT decisions not used as forward guidance for policy-rate paths. *Source: IMF — “Quantitative Tightening by the ECB: Why and How?” (June 29, 2023).* ### 1. ECB Balance Sheet and International QT Experiences ________________________________ 4 ### 1. ECB Balance Sheet and International QT Experiences ### Executive Summary - The ECB Governing Council (GC) initiated quantitative tightening (QT) to reduce the Eurosystem stock of securities holdings accumulated since January 2015 through quantitative easing (QE). - QE supported aggregate demand and transmission by compressing bond yields; QT aims to remove monetary accommodation, alleviate QE side effects, reduce the Eurosystem’s footprint in financial markets, and ensure issuer limits do not hinder the ability to restart large-scale asset purchases. - QT should play a limited role in calibrating the monetary stance; the policy rate should remain the primary instrument because: - Considerable uncertainty exists about QT’s effects on yields and financial conditions. - QT could increase fragmentation in euro area debt markets and interfere with monetary transmission. - A relatively fast pace of QT would reduce the necessary conventional policy space from a monetary risk management perspective. - Recommended QT design features: - Pre-announcing a QT pace and dates to reassess speed to allow market participants to adjust expectations and smooth price adjustment of bonds. - Allowing deviation from the capital key over a reasonable time horizon as a result of passive QT. - Embedding an “escape clause” to be activated in case of severe market dysfunction creating a severe impediment to smooth transmission of monetary policy. ### A. Context - The ECB initiated asset purchase programs in January 2015 (APP) and March 2020 (PEPP), amounting to €4.9 trillion by end 2022, corresponding to 60 percent of euro area GDP. - With inflation rising since 2021, the ECB halted QE in June 2022 and, in July 2022, raised rates for the first time since July 2011. - After cumulative rate increases of 300bps, the GC began normalizing the Eurosystem’s monetary securities holdings through a measured and predictable pace starting March 1, 2023. - Between March and June 2023, the Eurosystem’s securities holdings were being reduced by a monthly average of €15 billion. - Starting July 2023, the GC planned to stop all APP reinvestments, increasing the pace of APP portfolio reduction. By end 2023, the GC planned to review its operational framework to provide information regarding the size of the ECB’s balance sheet in a steady state. - Unwinding the QE portfolio requires understanding its past impact and associated risks: large holdings depressed sovereign yields, contributed to disappearance of the unsecured interbank market, and reduced availability of marketable safe assets. ### B. Rationale for QT - QE influences long-term yields via signaling and portfolio rebalancing channels: - Signaling: announcements act as forward guidance on the expected path of short-term interest rates. - Portfolio rebalancing: asset purchases compress the term premium by reducing net supply of long-term assets held by the private sector, working through limited substitutability (“preferred habitat theory”) or duration risk extraction. - Estimated magnitude of QE’s effect on yields in the euro area: - Studies suggest yields fall on average 3.6 to 5.4bp per €100bn of asset purchases. - Applying these elasticities to the €4.9 trillion stock of QE purchases implies the level of accommodation to 10-year GDP-weighted zero-coupon yields of the four largest euro area jurisdictions is on average between 176 and 265bp. - These estimates carry large uncertainty and are state contingent; effects can be much higher in periods of market stress. - QE’s side-effects on safe securities and collateral: - €4.3 trillion in public sector bond purchases between June 2015 and June 2022 reduced the share of sovereign free float to 40 percent out of €10.3 trillion in outstanding euro area sovereign bonds by end 2022. - German Bunds’ free float dropped to 20 percent of issued stock by end 2022 (threefold decline), while Italy, France, and Spain fell to 56, 51, and 47 percent, respectively. - Locked up government securities peaked at €542 billion in Q3 2021 due to Eurosystem refinancing operations. - A scarcity measure based on the difference between the yield of an asset and the similar asset swap rate confirms increased scarcity, particularly with PEPP during 2020-22; the average asset swap spread in recent months remains almost double the pre-QE average. - QE and interbank market effects: - Asset purchases created abundant bank reserves, shifting the unsecured overnight lending rate close to the Deposit Facility Rate (DFR) and moving from a corridor to a floor system. - The resulting high level of excess reserves reduced the relevance of unsecured interbank transactions and may make exit more difficult. - Financial and political implications of large asset holdings: - Rising policy rates created an interest rate mismatch producing profit losses in the Eurosystem because reserves are remunerated at the DFR, which increased by 400bps from July 2022 to June 2023, while assets earn fixed coupons. - Losses in 2022 were offset by releases from risk provisions and buffers accumulated from profits over 2012-2021; over time, losses will decline as assets mature and reinvested income rises. - Temporary losses have halted transfers to fiscal authorities and can raise political scrutiny, though losses do not impair monetary policy conduct if the central bank’s commitment to the inflation target remains credible. ### C. Policy Considerations Associated with QT - Uncertain effects of QT on financial conditions: - Evidence on QT’s impact on yields is sparse; the Fed’s 2017–2019 balance sheet reduction of $640 billion (about 2.8 percent of GDP) at a monthly average pace of $28 billion produced mixed estimates. - Some studies found no statistically significant impact on the 10-year Treasury yield, while others associate a reduction of the Fed’s balance sheet by 1 percent of GDP with an increase in 10-year yields equivalent to a 6-10bp increase of the Fed Funds rate. - State contingency of QT effects: - Like QE, QT effects are likely state contingent and could be large during periods of market dysfunction. - If implemented when markets are calm and monetary policy is not constrained, QT’s impacts may be mild. - The signaling channel of QT differs from QE; QT does not imply forward guidance on keeping interest rates near the effective lower bound and could instead convey information about the pace of rate hikes and the terminal interest rate needed to achieve the inflation target. - QT “surprises” have larger effects when markets are turbulent; evidence suggests the interest-rate equivalent of QT can be almost three times higher during crisis periods. *Source: IMF — “Quantitative Tightening by the ECB: Why and How?” (June 29, 2023).* ### 15. The impact on yields of the QT program announced by the ECB over the 12 months ahead ### 15. The impact on yields of the QT program announced by the ECB over the 12 months ahead ### Projected immediate impact on yields and policy-rate equivalence - Using the QE elasticity symmetrically for QT and assuming absence of non-linearity, the ongoing monthly QT pace of €15 billion until the end of June and doubling thereafter would translate into an increase in the term premia of 12-18bp over the next 12 months. - Extrapolating impacts on policy rates identified for the Fed translates into a substitution with the policy rate in the range of 15 to 22bp. - This substitution is small relative to the cumulative 400bps increase in the ECB’s deposit facility rate between July 2022 and June 2023. - Much of the faster-than-expected balance-sheet rundown compared to market expectations in December 2021 has been gradually priced into the yield curve, implying a likely small marginal impact going forward. ### Risks that could amplify impacts and fragmentation concerns - QT may have larger impacts than currently predicted if marginal buyers are not forthcoming at a time of large sovereign financing needs; increased bond yields may reflect expectations about future bond net supply. - For the first time in many years, euro area governments face lower ECB support in rolling over maturing debt, raising the question which actors will absorb the net supply and at what spread levels. - Euro area banks are a natural candidate to step in given historical appetite, with current holdings below historical highs in most countries. - QE historically helped reduce fragmentation; unwinding that support could risk reintroducing fragmentation if QT is not predictable and cautious. - The PEPP’s flexible reinvestments and, if needed, the Transmission Protection Instrument (TPI) are highlighted as defenses against fragmentation; sovereign solvency concerns should be handled via Outright Monetary Transactions and ESM macro-adjustment programs. ### Market-functioning and crisis tools - Tools designed to support market liquidity should be used to address severe financial market dysfunction (central bank lending facilities, reverse auctions, outright asset purchases). - The separation principle is recommended: monetary policy tools (policy rates and QE/QT) should be distinct from emergency support tools aimed at market functioning or financial stability. - The pace of balance-sheet reduction could be reassessed in stress episodes; targeted interventions may accompany passive QT as seen in prior episodes (example: Bank of England during the LDI crisis). ### QT and interaction with the policy-rate path - If QT effects on the monetary stance remain small, there is no tradeoff between rebuilding conventional and unconventional policy space. - The QT objective of supporting the monetary stance is state dependent: desirable during tightening cycles; weaker case during loosening cycles ceteris paribus. - Short-term policy rates should remain the primary tool to fine tune the monetary stance; QT should proceed gradually and predictably. - Since the start of rate hikes, long-term yields have responded almost one-to-one with expected short rate hikes, indicating transmission from short-term to long-term rates has not been impaired by Eurosystem securities holdings. ### Proposed QT implementation and decision-tree - The current ECB approach—gradual and predictable unwinding with passive runoff of the APP portfolio—is appropriate to learn about yield impacts and monitor transmission risks. - Active outright sales are not warranted given the relatively shorter average maturity of the ECB’s asset portfolio and the modest initial roll-off. - Decision-tree branches (past learning phase): - Tightening cycle—QT can serve all objectives. If impact on yields remains small and balance sheet far from desired endpoint, the ECB could increase the pace to reduce its footprint. With APP reinvestments stopped, reducing PEPP reinvestment and later ceasing it fully could lead to a roll-off of €624 billion per year, with an expected impact on yields in the range of 22-33bp. - Loosening cycle—If the balance sheet remains far above the desirable endpoint and QT impact is minimal, QT could continue; if QT significantly tightens financial conditions, QT should pause in a loosening cycle to avoid increasing ELB risk. ### Estimates of terminal size, pace required, and reserve considerations - Beyond minimum reserve requirements and currency in circulation (about 13 percent of 2022 GDP or €1.7 trillion), demand for bank reserves has increased since 2009 for precautionary and prudential reasons. - Average excess reserves measures (back-of-the-envelope): - Average excess reserves after the GFC and prior to QT (2010-2022): €1.1 trillion. - Average throughout the QE period (2015-2022): €1.7 trillion. - More recent average (2017-2022, covering the floor system period): €2.1 trillion. - These imply a total Eurosystem balance sheet of between €3 and 4 trillion or 20-30 percent of 2022 GDP as a potential steady-state range. - Reaching a balance sheet of that size over the next 7 years would mean an annual reduction of €554-694 billion, corresponding to at least three times the current monthly QT pace. - If continued beyond June 2023, the current QT pace would keep the Eurosystem balance sheet twice as large as the Fed’s or the Bank of England by end 2024. - The ECB 2021 estimated that the floor required excess liquidity (FREL) could be somewhere between €400 billion and €1 trillion, implying a lower bound of the terminal balance sheet compatible with a floor system between €2.2 and €2.8 trillion. ### Operational guidance and principles for QT communication - Continue a cautious, observable approach to ascertain the appropriate balance sheet size and decide to slow or stop QT by observing the reserve satiation point. - Principles for the QT roadmap: - Early communication of QT principles to make the process predictable and eventually “invisible” even as market conditions change. - Pre-announcing regular but infrequent dates at which the speed of QT is re-assessed and pre-announcing any changes to the QT pace so market participants can adjust pricing. - Embedding an “escape clause” in the QT setup to permit pausing the program in the event of a severe and systemic shock that impedes smooth transmission of monetary policy. ### International experience (selected lessons) - Major central banks have typically initiated QT by partially or fully ceasing reinvestments (passive QT) or by selling securities ahead of maturity (active QT); approaches differ by pace, type, and maturity structure of holdings. - Fed: passive QT via caps on reinvestment (US$30 billion per month for treasuries and US$17.5 billion per month for MBS initially in May 2022), with caps doubled after three months and caps expected to remain in place over the medium term. - BoE: ceased reinvesting maturing gilts in February 2022 and initiated sale of non-financial investment-grade corporate bonds; communicated gradual and predictable QT with discretion to alter plans and temporarily reversed QT during severe dysfunction in October 2022. - RBNZ and RBA: examples of managed sales to Treasury or stopping reinvestments; both emphasized gradual, predictable approaches and reserved the right to change pace or halt sales if conditions change. - Commonalities: pre-announced QT start months in advance, emphasis on predictability, gradualism, and discretion to adjust QT in response to financial and economic conditions; QT decisions not used as forward guidance for policy-rate paths. *Source: IMF staff analysis (chapter excerpt).* --- _Source: https://www.imf.org/-/media/files/publications/cr/2023/english/1eurea2023002.pdf_