## wpiea2023145-print-pdf

## Source details

**Canonical URL:** [wpiea2023145-print-pdf](https://www.imf.org/-/media/files/publications/wp/2023/english/wpiea2023145-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2023/english/wpiea2023145-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2023/english/wpiea2023145-print-pdf.pdf.json)

---

### Summary
- The Eurosystem, having grown its footprint vastly, faces a period of loss-making as rising policy rates lift the remuneration of bank reserves while assets churn more slowly.
- QE acted as a fixed-for-floating swap that removed duration risk from the private sector and created bank reserves, exposing the Eurosystem to interest rate risk during tightening.
- Ten-year (2022−31) projections find losses that are large but temporary and recoupable under the baseline; asset churn into higher-yielding securities materially improves net income.
- Faster QT can reduce cumulative net income because savings from faster reserve reduction are exceeded by lost interest income from lower reinvestments.
- Parallel upward shifts in yield curves are profit enhancing; lower banknote demand reduces profits.
- Policy conclusions (high level):
  - Temporary and recoupable loss-making obviates need for capital contributions or state indemnities; losses can be offset against future net income following the US Federal Reserve’s approach.
  - Fiscal impacts will be material even without capital support: annual taxes and transfers of 0.1−0.2 percent of GDP may face long interruptions (one case stretching to 11 years).
  - More-conservative profit distributions in the future steady state could smooth dividends.
  - Loss-making must remain orthogonal to monetary policy decision-making; credibility rests on delivering the price stability mandate.

### Projections and Scenarios (2022−31)
- Baseline scenario:
  - Uses market rate paths and announced policies.
  - Indicates two years of loss-making for the Eurosystem.
  - Indicates 0−5 years of loss-making for the five NCBs.
- Sensitivities:
  - Faster QT reduces cumulative net income due to lost interest income from lower reinvestments despite faster reserve reduction.
  - Parallel upward shifts in yield curves increase profits.
  - Lower banknote demand reduces net income.
- Valuation effects:
  - Unrealized valuation losses treated as of limited relevance for reporting given amortized cost accounting; income-sharing rules used to derive NCB-level P&L projections.

### Balance Sheet Evolution and Key Statistics
- Historical and peak magnitudes:
  - End-2000 total assets: €836 billion.
  - End-2000 currency in circulation: €371 billion.
  - End-2000 bank reserves: €125 billion (of which €116 billion satisfied 2 percent minimum reserve).
  - By mid-2012 excess liquidity: almost €1 trillion.
  - Pandemic period (March 2020 − mid-2022): euro area national public sector net debt issuance ≈ €1.7 trillion; Eurosystem net purchases on the secondary market: €1.9 trillion; excess liquidity grew by €2.5 trillion.
  - By end-2021: Eurosystem balance sheet = €8.6 trillion; monetary policy assets = 56 percent of euro area GDP; QE book at amortized cost = €4.7 trillion; bank reserves = €4.3 trillion.
- Eurosystem consolidated balance sheet (selected € billion, 2021 vs 2022):
  - Securities of euro area residents denominated in euro: 4,886; 5,102
  - Banknotes in circulation: 1,544; 1,572
  - Liabilities to euro area credit institutions: 4,371; 4,077
  - TLTRO and other claims on euro area credit institutions: 2,229; 1,355
  - Gold and gold receivables: 559; 593
  - Revaluation accounts: 555; 586
  - Capital and reserves: 113; 115
  - Total assets / Total liabilities: 8,564; 7,956
- Reserve remuneration and implementation:
  - Reserve tiering introduced; multiple initially set at six.
  - Minimum required reserves remunerated at the MRO rate until December 20, 2022, and thereafter at the DFR.
  - Return of the DFR to positive in September 2022 led to a large sweep of bank reserves from the current account to the deposit facility.

### Eurosystem Structure, Income Sharing, and NCB Heterogeneity
- Institutional features:
  - Eurosystem comprises multiple NCBs (now 20); ECB Governing Council sets base money volume; issuance delegated to NCBs.
  - Decentralized structure implies fiscal spillovers of Eurosystem profits and losses differ by country; disaggregated NCB-level net income projections are necessary.
- Key rules shaping outcomes:
  - Parametric limits: per-issue and per-issuer limits set at 33 percent in most cases; national public debt purchases generally mirror the ECB capital key; PEPP holdings not consolidated into PSPP holdings.
  - Bond allocation and risk-sharing: of every €100 of public sector debt acquired, €80 domestic national debt by NCBs, €10 supranational by NCBs, €10 national debt by ECB; credit losses on national debt held by NCBs remain each NCB’s responsibility; about two-thirds of QE portfolio (end-2022) comprised NCB holdings of home sovereign debt.
  - Pooling rules (Decision (EU) 2016/2248): ECB collects net monetary income, aggregates it, and allocates back to NCBs based on capital key; monetary income measured at actual rate of return or reference rate currently set at the MRO rate; expenses on bank reserves shared fully.
- Distributional implications:
  - Home bias in NCB bondholdings and capital-key-based pooling means NCBs in lower-debt countries with lower funding costs generally face larger and more persistent losses relative to peers.

### Methodology for Projections and Key Assumptions
- Projection mechanics:
  - Monthly iteration to end-2031; earnings and expenses applied monthly to prior balances.
  - Pre-provision net losses subtracted from “other assets” and “other liabilities”; when pre-provision net income reverts to positive, payouts assumed zero until capital, reserves, and general provisions replenished to end-2021 levels; thereafter 2008−21 average payout ratios applied.
  - Starting point: end-2021.
- Seven-block balance-sheet view:
  - Assets: QE book (amortized cost), TLTROs (roll off at maturity), other assets (revalued monthly).
  - Liabilities: banknotes, bank reserves, other liabilities.
  - Seventh block: net TARGET2 position.
- Key parameter assumptions (selected):
  - Policy rate path: baseline DFR trajectory from €STR forwards as of end-February 2023; peak DFR = 3.5 percent around mid-2023, falling to 2.6 percent three years later; linear convergence to nominal neutral rate = 2.3 percent by end-2028; 50 basis point MRO margin over DFR assumed constant.
  - Yields on QE portfolios: initial rates = 2021 period-average returns; Jan 2022–Feb 2023 rates move with observed 10-year sovereign yields; thereafter 10-year bund yield moves in lockstep with baseline DFR. Synthetic 10-year spreads over bund (2015−22 averages): Netherlands = 16 basis points; France = 36 basis points; Spain = 102 basis points; Italy = 165 basis points.
  - Reinvestment: new bonds acquired through reinvestment assumed to yield projected interest rate at purchase.
  - TLTRO / QT: baseline QT path = APP shrinking by €15 billion per month during March−June 2023, then zero reinvestment; PEPP assumed full reinvestment throughout.
  - Other balances: foreign assets and gold constant; foreign assets earn 1-year US treasury bill rate; other assets grow at uniform rate = 6.6 percent per year (banknote growth 2012−21 average).
  - Reserves response: each €1 of TLTRO repayment matched by €0.70 reduction of bank reserves and €0.30 reduction of government and non-EU deposits (pattern since Nov 2022). Each €1 of QT drains reserves one-for-one until reserves hit a floor set at 1 percent of total bank assets.

### Baseline Results and NCB-Level Outcomes
- Buffers and starting balances:
  - General provisions at end-2021: €128 billion.
  - Capital and reserves at end-2021: €55 billion.
  - Revaluation accounts at end-2021: €555 billion.
- Aggregate Eurosystem projection:
  - After roughly breaking even in 2022, projected large losses in 2023−24 before returning to profit in 2025 and recouping losses by 2027.
  - Cumulative losses over 2023−24: about €55 billion, equivalent to 0.5 percent of euro area GDP.
  - These losses exceed half of the Eurosystem’s general provisions; for Bundesbank and Banque de France they begin to erode capital and reserves.
- Top-five NCB outcomes (selected):
  - Bundesbank: cumulative losses projected to peak at almost 1.2 percent of GDP in 2025; would exceed €20 billion general provisions + €6 billion capital and reserves (end-2021); capital and reserves projected negative for seven years.
  - Banque de France: cumulative losses peak at 0.7 percent of GDP by 2024; provisions consumed; capital and reserves projected negative for four years (remain positive but below statutory target before provisioning).
  - Banco de España: cumulative losses peak at 0.2 percent of GDP in 2024; provisions reduced but capital and reserves intact.
  - De Nederlandsche Bank: cumulative losses peak at 0.3 percent of GDP in 2024; provisions reduced but capital and reserves intact.
  - Banca d'Italia: just avoids loss-making.
- Historical context and payout patterns:
  - 2008−21 pre-provision, pre-tax net income averaged 0.3 percent of euro area GDP annually.
  - 2008−21 average: 40 percent of pre-provision, pre-tax net income of the top-five NCBs flowed to provisions and retained earnings.
  - Corporate tax payments 2008−21: Banque de France paid out 46 percent of post-provision net income; Banca d’Italia 22 percent.
  - QE years (2015−21) profit transfers: Banca d'Italia averaged 0.3 percent of GDP annually; Banco de España averaged 0.2 percent of GDP annually.

### Valuation and Buffer Sensitivities
- Notional valuation effects on QE book (amortized cost accounting caveat):
  - Static valuation losses (if marked to market, no sales assumed): end-2022 estimated valuation losses = €758 billion (5.7 percent of euro area GDP); peak 2023 = €1,015 billion (7.2 percent of GDP).
  - Marking QE book to market would wipe out all buffers including revaluation accounts for Eurosystem and top-five NCBs; paper treats this as moot for reporting under amortized cost.
- Buffers per NCB (selected end-2021 figures):
  - ECB general risk provisions: €8.2 billion; capital = €8.2 billion (capital + general risk provisions = €16.4 billion).
  - Banque de France: capital (€1 billion) + statutory reserves (€2 billion) = €3 billion.
  - Banca d'Italia: capital = €7.5 billion; ordinary reserve = €6.5 billion.
  - Banco de España: capital = €1 billion.
  - De Nederlandsche Bank: capital = €8.5 billion.

### Alternative Scenarios (selected quantitative impacts)
- Scenario 1: 50 basis point parallel upward shift in rate path vs baseline.
  - Eurosystem cumulative profits 2022−31 increase by €19 billion.
  - Bundesbank cumulative profits 2022−31 fall by €6 billion.
  - Banca d'Italia gains ≈ €13 billion; Banque de France, Banco de España, De Nederlandsche Bank make small gains.
- Scenario 2: Faster QT (cessation of PEPP reinvestment in Jan 2025 + higher 10-year yields).
  - Eurosystem cumulative net income 2022−31 falls by €93 billion.
  - Savings from faster reserve reduction exceeded by lost income from lower reinvestment; Banca d’Italia hit largest reduction among top-five; De Nederlandsche Bank smallest hit.
  - Assumed yield elasticity: every €100 billion of net asset purchases reduces yields by 4.5 basis points.
- Scenario 3: Currency in circulation growth half as fast (3.3 percent vs baseline 6.6 percent).
  - Eurosystem cumulative net income 2022−31 reduced by €118 billion.
  - Lower banknote demand reduces seigniorage and holdings of interest-earning assets.
- Scenario 4: Setting pooling reference rate = DFR instead of MRO.
  - Modest distributional impacts 2022−31: Banque de France ≈ €5 billion higher net income; Banco de España ≈ €4 billion higher; Bundesbank, Banca d’Italia, De Nederlandsche Bank modest profit reductions.

### Policy Discussion and Recommendations
- Projected fiscal interruptions:
  - Long interruptions to tax and profit payments for some NCBs:
    - Bundesbank: unlikely to make payments for a total of 11 years (includes 2020−21).
    - De Nederlandsche Bank: gap = 4 years.
    - Banque de France: gap = 3 years.
    - Banco de España: gap = 2 years.
    - Banca d'Italia: corporate tax payments and profit transfers dip to minimal levels in 2023 but avoid interruption.
  - Annual taxes and transfers of 0.1−0.2 percent of GDP may be interrupted, in one case for 11 years.
- Approaches to addressing central bank loss-making (models):
  - Swedish model: capital calls on the state, government transfers interest-earning securities to central bank in exchange for equity; suitable if negative equity is on an explosive path.
  - UK model: standing indemnity from the state; links central bank net income to the national budget.
  - US model: cumulate losses as an accounting entry to be wound down by future profits; suited to temporary loss-making.
- Recommendation:
  - The paper recommends NCBs, in consultation with governments and the ECB, consider adopting the US approach (looking through temporary negative capital by recording losses to be offset by future profits), consistent with ECB guidance noting remaining losses may be recorded on the ECB balance sheet to be offset against future net income.
- Operational guidance:
  - More-conservative profit distribution norms in future steady state to smooth transfers.
  - Maintain provisioning practices and buffers to mitigate NCB-level fiscal exposures.
  - Ensure loss-making remains orthogonal to monetary policy decisions; QT and other liquidity-reducing tools should be used for monetary policy reasons alone, not to manage balance-sheet valuation outcomes.

### Conclusion and Final Assessment
- Main findings:
  - Losses will be large and persistent at some NCBs but temporary in this cycle and not so large as to necessitate capital support from member states under the baseline.
  - Parallel upward shifts in yield curves slightly improve ten-year net income; faster QT modestly reduces net income; lower banknote demand has negative seigniorage effects.
  - Coupon income from home sovereigns is the key driver of projected return to profits; home-bias rules generate dispersion across NCBs.
- Institutional judgment:
  - The EMU design and safeguards produce heterogeneous outcomes across NCBs; absent ECB balance-sheet policies, a weaker macroeconomic environment might have had larger negative public finance impacts.
  - ECB credibility depends on delivering price stability; loss-making should not influence monetary policy choices.
- Phasing for NCBs (conceptual three-phase projection):
  - Phase 1: Loss-making (including 2022) — years of annual losses vary by NCB (Germany ≈ five years, France and Netherlands ≈ three years, Spain ≈ two years, Italy avoids).
  - Phase 2: Buffer replenishment — tax payments paused until buffers restored (Spain by 2025, Netherlands by 2026, France by 2028, Germany by 2031).
  - Phase 3: New steady state — historical payout ratios applied; retained earnings flow into provisions when profits not distributed.

*IMF Working Paper: Raising Rates with a Large Balance Sheet: The Eurosystem’s Net Income and its Fiscal Implications (excerpted projections, scenarios, and policy discussion).*

### References .............................................................................................................

### wpiea2023145-print-pdf - References .............................................................................................................

### Summary
- The Eurosystem, having grown its footprint vastly in recent years, now faces a period of loss-making as rising policy rates lift the remuneration of bank reserves while assets churn more slowly.
- Quantitative easing (QE) removed duration risk from the private sector’s balance sheet and created bank reserves, effectively executing a fixed-for-floating rate swap that leaves the Eurosystem exposed to interest rate risk during tightening.
- The paper projects net income for the Eurosystem and its “top-five” national central banks (NCBs) over a ten-year horizon (2022−31), finding losses that are large but temporary and recoupable under the baseline.
- Asset churn into higher-yielding securities materially improves net income; faster quantitative tightening (QT) can reduce cumulative net income because savings from faster reserve reduction are exceeded by lost interest income from lower reinvestments.
- Parallel upward shifts in yield curves are profit enhancing; lower banknote demand reduces profits.
- Policy conclusions:
  - Temporary and recoupable loss-making obviates the need for capital contributions or indemnities from the state; losses can be offset against future net income following the US Federal Reserve’s approach.
  - Fiscal impacts will be material even without capital support, with annual taxes and transfers of 0.1−0.2 percent of GDP giving way to long interruptions, in one case stretching to 11 years.
  - More-conservative profit distributions in the future steady state could mitigate the on-off pattern of dividends.
  - Loss-making must remain orthogonal to monetary policy decision-making; credibility rests on delivering the price stability mandate.

### Projections and Scenarios (2022−31)
- Baseline scenario:
  - Uses market rate paths and announced policies.
  - Indicates two years of loss-making for the Eurosystem.
  - Indicates 0−5 years of loss-making for the five NCBs.
- Sensitivities:
  - Faster QT reduces cumulative net income due to lost interest income from lower reinvestments despite faster reserve reduction.
  - Parallel upward shifts in yield curves increase profits.
  - Lower banknote demand reduces net income.
- Valuation effects:
  - The authors treat unrealized valuation losses as of limited relevance for the Eurosystem context and emphasize income-sharing rules to derive NCB-level P&L projections.

### Balance Sheet Evolution and Key Statistics
- End-2000 total assets: €836 billion.
- End-2000 liabilities included:
  - Currency in circulation: €371 billion.
  - Bank reserves: €125 billion, of which €116 billion satisfied a minimum reserve requirement set at 2 percent.
- By mid-2012 excess liquidity: almost €1 trillion.
- June 2014: DFR cut to −0.1 percent; first TLTRO announced.
- March 2015: APP began, with APP monthly net purchases initially set at €60 billion.
- Pandemic period (March 2020 − mid-2022):
  - Euro area national public sector net debt issuance: some €1.7 trillion.
  - Eurosystem net purchases on the secondary market: €1.9 trillion.
  - Excess liquidity grew by €2.5 trillion.
- By end-2021:
  - Eurosystem balance sheet: €8.6 trillion.
  - Monetary policy assets: equivalent of 56 percent of euro area GDP.
  - QE book at amortized cost: €4.7 trillion.
  - Bank reserves: €4.3 trillion.
- Reserve remuneration and policy implementation:
  - Reserve tiering introduced with current account holdings exempt in an amount equivalent to a multiple of minimum required reserves, multiple initially set at six.
  - Minimum required reserves were remunerated at the MRO rate until December 20, 2022, and thereafter at the DFR.

### Eurosystem Structure, Income Sharing, and NCB Heterogeneity
- The Eurosystem comprises multiple reserve banks (now 20 NCBs); NCBs serve sovereign states rather than banking districts.
- The ECB Governing Council sets base money volume; issuance is largely delegated to NCBs.
- Decentralized structure implies fiscal spillovers of Eurosystem profits and losses differ by country; disaggregated NCB-level net income projections are necessary to assess fiscal impacts.
- Home bias in NCBs’ bondholdings—an outcome of mutually agreed Eurosystem safeguards—means NCBs whose sovereigns enjoy lower market funding costs generally experience larger and more persistent losses.
- Projections are made and shown for:
  - The Eurosystem, the ECB, Deutsche Bundesbank, Banque de France, Banca d’Italia, Banco de España, and De Nederlandsche Bank, with emphasis on the five NCBs.

### Policy Implications and Recommendations
- Capital and indemnities:
  - Because losses are temporary and recoupable, there is no need for capital contributions or state indemnities; losses can be offset against future net income.
- Fiscal impacts:
  - Even without state capital support, annual taxes and transfers of 0.1−0.2 percent of GDP may be interrupted, with interruptions potentially long (one case stretching to 11 years).
  - A weaker macroeconomic environment might have larger negative public finance impacts absent ECB balance sheet policies.
- Profit distribution and buffers:
  - More conservative profit distributions in the future steady state could smooth the on-off pattern of dividends.
  - Maintaining financial buffers and provisioning practices matters for NCB-level fiscal exposures.
- Institutional safeguards:
  - Loss-making must remain orthogonal to monetary policy decisions; credibility depends on delivering on the price stability mandate.
  - The paper explains Eurosystem income-sharing rules and applies them with transparent assumptions and public data to generate NCB-level P&L projections.

### Paper Scope and Structure
- Focus: Eurosystem profits and losses in 2022−31 and net income projections for the top-five NCBs, with policy discussion on central bank negative capital and fiscal spillovers.
- Organization:
  - Section II: Evolution of the Eurosystem’s balance sheet in its first 25 years.
  - Section III: Special Eurosystem rules that materially impact NCB net income.
  - Section IV: Assumptions used for projections.
  - Section V: Ten-year profit and loss projections in a baseline and alternative scenarios.
  - Section VI: Policy discussion on negative capital and fiscal spillovers.
  - Section VII: Conclusion.

*IMF Working Paper: Raising Rates with a Large Balance Sheet: The Eurosystem’s Net Income and its Fiscal Implications (References and selected sections).*

### 18. Projections at the NCB level are complex to produce, with a web of intra-Eurosystem claims and

### 18. Projections at the NCB level are complex to produce, with a web of intra-Eurosystem claims and

### Eurosystem rules shaping NCB-specific results
- Three interlocking rule sets shape NCB-specific outcomes:
  - Parametric limits to ensure QE delivers generalized monetary policy stimulus and preserves price-based incentives for fiscal prudence:
    - Require national public debt purchases to generally mirror the ECB’s capital key.
    - Per-issue and per-issuer limits set at 33 percent in most cases (PEPP holdings not consolidated into PSPP holdings).
  - Bond allocation and risk-sharing rules:
    - Of every €100 of public sector debt acquired: €80 must be domestic national debt bought by NCBs, €10 supranational debt bought by NCBs, and €10 national debt bought by the ECB.
    - Credit losses on national public sector debt held by NCBs remain each NCB’s responsibility.
    - Credit losses on supranational debt held by NCBs, national debt held by the ECB, and all private sector securities owned are fully shared across the Eurosystem.
    - As of end-2022, roughly two-thirds of the Eurosystem’s QE portfolio comprised NCBs’ holdings of debt securities issued by their own home sovereigns.
  - Pooling rules for equitable sharing of monetary income (Decision (EU) 2016/2248):
    - The ECB collects net monetary income from NCBs annually, aggregates it, and allocates it back to NCBs based on capital key shares.
    - Monetary income is income deriving from “earmarkable” assets and is measured at the actual rate of return, for certain assets at a “reference rate” currently set at the MRO rate, or in the case of gold at zero.
    - Expenses on bank reserves are shared fully.
    - Example: Banca d’Italia retains actual income on its Italian sovereign bondholding less an amount calculated at the MRO rate that is remitted to the ECB for pooling.

### Key balance-sheet magnitudes and amortized cost accounting
- Amortized cost accounting: adjusts book value from market value at purchase to face value at maturity based on contractual cash flows less realized impairments.
  - As of end-2021:
    - ECB’s QE portfolio: €445 billion at amortized cost and €453 billion at market value.
    - Top-five NCBs’ QE portfolios: €3.46 trillion at amortized cost and €3.53 trillion at market value.
- Eurosystem consolidated balance sheet (Table 1, € billion):
  - Securities of euro area residents denominated in euro: 2021 = 4,886; 2022 = 5,102
  - Banknotes in circulation: 2021 = 1,544; 2022 = 1,572
  - of which held for monetary policy purposes: 2021 = 4,713; 2022 = 4,937
  - Liabilities to euro area credit institutions: 2021 = 4,371; 2022 = 4,077
  - TLTRO and other claims on euro area credit institutions: 2021 = 2,229; 2022 = 1,355
  - Liabilities to other euro area residents: 2021 = 771; 2022 = 576
  - Gold and gold receivables: 2021 = 559; 2022 = 593
  - Liabilities to non-euro area residents: 2021 = 713; 2022 = 545
  - Claims on non-euro area residents: 2021 = 513; 2022 = 537
  - Revaluation accounts: 2021 = 555; 2022 = 586
  - Other assets: 2021 = 377; 2022 = 369
  - Capital and reserves: 2021 = 113; 2022 = 115
  - Other liabilities: 2021 = 498; 2022 = 485
  - Total assets: 2021 = 8,564; 2022 = 7,956
  - Total liabilities: 2021 = 8,564; 2022 = 7,956
- Return of the DFR to positive territory in September 2022 led to a large sweep of bank reserves from the current account (interest capped at zero) to the deposit facility to earn the positive DFR.

### Pooling rules and remuneration itemization (Table 2 summary)
- Pooling/remuneration principles (selected earmarkable assets and remuneration rules):
  - Amount of gold holdings in proportion to each NCB's capital key share: 0 percent
  - Claims equivalent to transfer of foreign reserves to the ECB (except gold): MRO rate
  - CBPP, CBPP2, PSPP (government/agency), and PEPP (government): MRO rate
  - SMP, CBPP3, PSPP (supranational), CSPP, and PEPP (excluding government/agency): Rate of return
  - Liabilities to euro area credit institutions related to monetary policy operations denominated in euro: Rate of return
  - Lending to euro area credit institutions related to monetary policy operations denominated in euro: Rate of return
  - Net liabilities arising from balances of TARGET accounts: MRO rate
  - Claims on non-Eurosystem central banks relating to liquidity-providing operations: Actual interest income (including accruals)
  - Net claims related to allocation of euro banknotes within the Eurosystem: MRO rate
  - Accrued coupon interest-related balances on impaired securities with full income/risk sharing applies: Not applicable
  - Accrued interest relating to regular monetary policy operations with maturity > 1 year: Not applicable

### Specific pooling items and their distributional effects
- Net lending to euro area banks related to monetary policy operations:
  - Claims on banks include MROs (at the MRO rate) and TLTROs (indexed to applicable ECB policy rates).
  - Liabilities to banks include required reserves (remunerated at the MRO rate until Dec 21, 2022, then at the DFR), banks’ deposits in current accounts (unremunerated), and deposit facility balances (remunerated at the DFR).
  - Interest on these items is fully pooled on an actual basis; each NCB shoulders a share of aggregate net expenses proportional to its capital key.
- Securities held for monetary policy purposes:
  - NCBs retain income on national public debt holdings to the extent actual rate of return exceeds the reference rate (MRO).
  - An amount calculated at the reference rate is remitted to the ECB for pooling irrespective of actual outperformance or underperformance.
  - Among top-five NCBs, Banca d’Italia and Banco de España have the largest excesses of actual returns over the reference rate; the Bundesbank retains essentially nothing.
  - Income on supranational, institutional, and most private sector debt is fully pooled.
- Net intra-Eurosystem TARGET2 (T2) claims:
  - T2 is the conduit for base money flows; QE transactions produced large Bundesbank net T2 claims and Banca d’Italia and Banco de España net T2 liabilities.
  - Historically little attention to remuneration because the reference rate (MRO) was zero during much of QE.
- Net intra-Eurosystem claims on euro banknotes:
  - Each NCB receives a banknote allocation proportional to its capital key after deducting an 8 percent allocation for the ECB.
  - Country-level demand differences (higher in Germany and the Netherlands, lower in Italy and Spain) create net intra-Eurosystem liability positions charged at the reference rate for pooling and redistribution.

### Methodology for projections and key assumptions
- Projection mechanics:
  - Monthly iteration to end-2031: earnings and expenses for each month are applied to relevant balances at end of previous month.
  - When pre-provision net income is negative, the full amount is subtracted from “other assets” and “other liabilities”; TLTRO repayments, QT, and banknote growth feed into the balance sheet construction.
  - When pre-provision net income reverts to positive, payouts assumed zero until capital and reserves plus general provisions are replenished to end-2021 level; thereafter observed average payout ratios for 2008−21 applied.
  - Starting point for projections: end-2021.
- Seven-block balance-sheet view:
  - Asset side: QE book (projected at amortized cost), TLTROs (assumed to roll off at maturity), and “other assets” (securities held for non-monetary policy purposes, foreign assets, gold, real estate, etc.; these are revalued monthly).
  - Liability side: banknotes, bank reserves, and “other liabilities” (government and non-EU official deposits, revaluation accounts, general provisions, capital and reserves).
  - Seventh block: net T2 position (asset or liability).
  - Valuation change on non-monetary policy securities is added to “other assets” and “other liabilities.”
- Key assumptions:
  - Policy rate path:
    - Baseline DFR trajectory derived from market pricing of €STR forwards as of end-February 2023.
    - Vintage indicates a peak DFR of 3.5 percent around mid-2023, falling to 2.6 percent three years later.
    - Thereafter assumed to converge linearly to an assumed nominal neutral rate of 2.3 percent by end-2028 and remain at that level thereafter.
    - The 50 basis point policy margin of the MRO rate over the DFR is assumed to remain constant.
  - Yields on QE portfolios:
    - Starting points are reported period-average rates of return for 2021 on portfolios of the five NCBs and the ECB.
    - From January 2022 through February 2023, rates of return assumed to increase in lockstep with observed changes in 10-year sovereign bond yields.
    - Thereafter, the 10-year bund yield is projected to move in lockstep with the baseline DFR.
    - Synthetic 10-year yields for other sovereigns constructed by applying observed average historical spread over bunds for 2015−22:
      - Netherlands: 16 basis points
      - France: 36 basis points
      - Spain: 102 basis points
      - Italy: 165 basis points
    - An average of the rates of return at the top-five NCBs is used as the yield for the Eurosystem as a whole, the “rest of the Eurosystem,” and for own funds and other non-monetary policy securities holdings.
    - New bonds acquired through reinvestment assumed to yield the projected interest rate at time of purchase.
  - TLTRO repayments, QT, and other balances:
    - Stock of TLTROs incorporates actual repayments through March 2023 and normal maturity schedules thereafter; no replacement of maturing TLTROs by MROs or 3-month LTROs is assumed.
    - Baseline QT path: APP portfolio shrinking by €15 billion per month during March−June 2023, followed by zero reinvestment thereafter.
    - Baseline assumption for PEPP: full reinvestment throughout.
    - Foreign assets and gold assumed constant; foreign assets earn rate of return on 1-year US treasury bills (observed or from forwards).
    - Other assets assumed to have uniform underlying growth rate (before revaluation) equal to that of banknotes: 6.6 percent per year (euro area-wide average in 2012−21).
  - Impact of TLTRO repayments and QT on reserves and net T2 positions:
    - Each €1 of TLTRO repayment assumed matched by a €0.70 reduction of bank reserves and a €0.30 reduction of government and non-EU deposits at the Eurosystem (pattern observed since November 2022).
    - Each €1 of QT drains reserves one-for-one until each jurisdiction’s bank reserves reach a floor set at 1 percent of total bank assets (pre-global financial crisis average).
    - Where reserves hit the floor, further QT reduces net T2 liabilities.
    - Government and non-EU official deposits assumed to remain constant after TLTRO repayments; remuneration assumptions set per ECB announcements.
  - Operating expenses and net fees and commissions (including securities lending) grow in line with the inflation target; both are relatively small items.

### Baseline results and historical context
- Historical averages and distributional patterns:
  - In 2008−21, the Eurosystem’s pre-provision, pre-tax net income averaged 0.3 percent of euro area GDP annually.
  - QE and negative DFR effects:
    - Negative DFR led banks to pay interest to the Eurosystem on a rising stock of reserves.
    - QE compressed long-term yields and dampened returns on the growing bond portfolio.
  - Top-five NCB performance in 2008−21:
    - Banca d’Italia and Banco de España: about 0.4 percent of GDP annual profits (above-average), reflecting larger holdings of higher-yielding debt.
    - Bundesbank and De Nederlandsche Bank: about 0.15 percent of GDP annual profits (below-average).
    - Banque de France: in between.
- Payout and provisioning patterns:
  - 2008−21 average: 40 percent of pre-provision, pre-tax net income of the top-five NCBs flowed to provisions and retained earnings.
  - Corporate tax payments on average 2008−21: Banque de France paid out 46 percent of post-provision net income; Banca d’Italia 22 percent.
  - QE years (2015−21) profit transfers to the state:
    - Banca d'Italia averaged 0.3 percent of GDP annually.
    - Banco de España averaged 0.2 percent of GDP annually.
  - Payouts peaked in 2019 before several top-five NCBs (notably the Bundesbank and De Nederlandsche Bank) prioritized provisioning.

*Source: Excerpt from the provided IMF Working Paper content.*

### 26. At end-2021, the Eurosystem’s general provisions stood at €128 billion, mostly held at the top-five

### 26. At end-2021, the Eurosystem’s general provisions stood at €128 billion, mostly held at the top-five

### Balance sheet buffers and provisions
- General provisions at end-2021: €128 billion.
- Capital and reserves at end-2021: €55 billion.
- Revaluation accounts at end-2021: €555 billion, mostly linked to exchange rate-related effects on holdings of foreign currency denominated assets and to movements in the value of gold holdings.
- Excluding revaluation accounts, buffers varied widely in size, reflecting different approaches to provisioning (Annex III).
- Among the top-five NCBs, Banca d’Italia held the largest buffers relative to unadjusted total assets.
- Note: Provisions and reserves here include only funds allocated to offset general risks, excluding items such as pension buffers.

### Baseline projections and aggregate losses
- After roughly breaking even in 2022, the Eurosystem is projected to incur large losses in 2023−24 in the baseline, before returning to profit in 2025 and fully recouping the losses by 2027.
- Cumulative losses over 2023−24: about €55 billion, worth 0.5 percent of euro area GDP.
- These cumulative losses will exceed half of the Eurosystem’s general provisions and, in the case of the Bundesbank and Banque de France, begin to erode capital and reserves.
- Main drivers of the projected turnaround: TLTRO runoff helps hasten return to profit, but the main driver is higher income on new assets acquired through reinvestment.
- For the Eurosystem as a whole, losses are fully recouped by 2027.

### Outcomes for the top-five NCBs
- Bundesbank:
  - Cumulative losses projected to peak at almost 1.2 percent of GDP in 2025.
  - These losses would exceed the sum of €20 billion of general provisions and €6 billion of capital and reserves set aside as of end-2021.
  - Capital and reserves projected to remain negative for seven years.
- Banque de France:
  - Cumulative losses projected to peak at 0.7 percent of GDP by 2024.
  - Provisions projected to be fully consumed, leaving capital and reserves positive but below the statutory national target.
  - Capital and reserves projected to remain negative for four years.
- Banco de España:
  - Cumulative losses peak at 0.2 percent of GDP in 2024.
  - General provisions reduced but capital and reserves left intact.
- De Nederlandsche Bank:
  - Cumulative losses peak at 0.3 percent of GDP in 2024.
  - General provisions reduced but capital and reserves left intact.
- Banca d'Italia:
  - Just avoids loss-making.

### Forecast accuracy and observed 2022 results
- The paper treats 2022 as the first “forecast” year, starting projections from 2021 data.
- Projections for 2022 come close to top-five NCBs’ reported results, with the noteworthy exception of Banque de France.
  - Close to De Nederlandsche Bank.
  - Differ but not widely from Banca d’Italia and the Bundesbank.
  - Miss by a wider margin on Banco de España.
- Main source of 2022 forecast errors: underestimation of gross interest income, especially for Banque de France, likely due to portfolio composition (shorter maturities, shares of higher-yielding securities, inflation-linked sovereign bonds, higher-yielding corporate bonds in QE books).
- Other potential accuracy risks going forward:
  - Overestimation of valuation changes on NCBs’ non-monetary policy portfolios if some portion is held at amortized cost.
  - Overestimation of interest on reserves if declining liquidity results in a gap opening up between €STR and the DFR (i.e., a return to the corridor system).
  - General information asymmetry between the Eurosystem and the public domain.

### Notional valuation effects on the QE book (aside)
- Under applicable accounting norms, unrealized gains or losses on the QE portfolio do not feed through the income statement or reflect on the balance sheet; valuation effects are realized only if securities are sold outright.
- Static QE-portfolio valuation relative to book value at end-2021 (assumes orderly market conditions, no fire sale discounts):
  - Estimated valuation losses at end-2022: €758 billion, equivalent to 5.7 percent of euro area GDP, exceeding the sum of general provisions, capital and reserves, and revaluation accounts as of end-2021.
  - Valuation losses peak in 2023 at €1,015 billion, or 7.2 percent of GDP, as interest rates peak.
  - Largest impacts on Banco de España and Banca d’Italia due to larger repricing on higher-yielding QE portfolios.
- The paper notes marking the QE book to market would wipe out all buffers including revaluation accounts for both the Eurosystem as a whole and its top-five NCBs.
- The paper treats these valuation figures as a moot aside given amortized cost accounting and limited relevance to balance sheet reporting, while noting reporting alternatives (e.g., US Federal Reserve memorandum items).

### Alternative scenarios and their impacts
- Scenario 1: 50 basis point parallel upward shift in the rate path relative to the baseline.
  - Eurosystem’s cumulative profits in 2022−31 increase by €19 billion.
  - Initial deterioration on account of reserve remuneration is more than fully recouped later by higher returns on the QE portfolio.
  - Bundesbank ends up worse off than in the baseline: profits in 2022−31 fall by €6 billion due to low bond returns and large stock of bank reserves relative to its capital key share.
  - Banque de France, Banco de España, and De Nederlandsche Bank make a small gain.
  - Banca d'Italia gains about €13 billion.
  - A scenario combining significantly higher near-term rates with lower rates farther out would be more challenging for profits.
- Scenario 2: Faster QT (cessation of PEPP reinvestment in January 2025 and higher 10-year sovereign yields due to faster QT).
  - Eurosystem’s cumulative net income in 2022−31 falls by €93 billion.
  - Savings from faster reduction of bank reserves are exceeded by lost income from lower reinvestment.
  - Banca d’Italia sees the largest cumulative net income reduction among the top-five NCBs; De Nederlandsche Bank records the smallest hit.
  - Assumed yield elasticity: every €100 billion of net asset purchases reduces yields by 4.5 basis points (mid-point estimate from QE literature cited).
- Scenario 3: Currency in circulation growth half as fast as in baseline (uniform annual banknote growth of 3.3 percent instead of 6.6 percent).
  - Eurosystem’s cumulative net income in 2022−31 reduced by €118 billion.
  - Lower demand for currency reduces seigniorage and holdings of interest-earning assets (e.g., potential effect of a digital euro).
- Scenario 4: Setting the DFR as the pooling reference rate instead of the MRO rate.
  - Modest distributional impact on cumulative P&Ls in 2022−31:
    - Banque de France: cumulates about €5 billion higher net income.
    - Banco de España: cumulates about €4 billion higher net income.
    - Bundesbank, Banca d’Italia, and De Nederlandsche Bank: similarly modest profit reductions.

### Policy discussion and recommendations
- Key projected outcomes:
  - Some Eurosystem NCBs will see general provisions fully depleted and capital and reserves turning negative as early as in 2024.
  - Capital and reserves are projected to remain negative at the Bundesbank for seven years and at Banque de France for four years.
  - In more adverse scenarios, a larger number of NCBs could experience sustained negative capital and reserves.
- Institutional and policy considerations:
  - The ECB’s Convergence Report (June 2022) takes a strict view: an NCB’s net equity below statutory capital or negative should be avoided; such a situation would require the respective member state to provide appropriate capital within a reasonable period to comply with the principle of financial independence.
  - However, central banks can function credibly with negative equity for prolonged periods if negative equity is not on an explosive path and is not allowed or seen to unduly influence monetary policy decisions.
  - Examples of central banks with negative equity histories include the Czech National Bank, the National Bank of Slovakia, and the Central Bank of Chile.
  - The paper emphasizes that delivering on the primary mandate is central; for the Eurosystem, equity including revaluation accounts will remain positive, instances of negative capital and reserves are temporary, losses will be fully recouped from subsequent net income, and the Governing Council structure insulates monetary policy from NCB-specific considerations.
- Approaches to addressing central bank loss-making (three models):
  - Swedish model (Sveriges Riksbank): contemplates capital calls on the state in some situations; government transfers interest-earning securities to central bank and takes equity as consideration. Desirable if negative equity is on an explosive path.
  - UK model (Bank of England): standing indemnity from the state eliminates need for one-off support but tightens links between central bank net income and the national budget.
  - US model (Federal Reserve): cumulate losses as an accounting entry to be wound down by future profits; suited to temporary loss-making.
- ECB stance and recommendation:
  - ECB appears to be leaning toward the US approach: in an “explainer” updated in February 2023, it noted any remaining losses after indemnities from NCBs “may be recorded on the ECB’s balance sheet, to be offset against any net income received in the future.”
  - The paper recommends that NCBs, in consultation with their governments and the ECB, should consider adopting a similar approach (i.e., looking through temporary negative capital by recording losses to be offset by future profits).
- Fiscal links and profit transfers:
  - Taxes and transfers already declined for all top-five NCBs in 2020−21, to zero for the Bundesbank and near zero for De Nederlandsche Bank.
  - In 2020, Banque de France set aside additional provisions and channeled retained earnings to its “fund for general risks,” pushing payments to the state to 0.02 percent of GDP in 2021.
  - Banco de España and Banca d’Italia continued provisioning at annual average rates of about 0.3 percent and 0.1 percent of GDP, respectively, in 2020−21, while continuing payments to the state.
- Conceptual three-phase projection period for NCBs:
  - Phase 1: Loss-making (including 2022).
    - Annual losses projected for five years in Germany, three years in France and the Netherlands, and two years in Spain; only Italy avoids loss-making.
  - Phase 2: Buffer replenishment (profits resume and tax payments remain paused until buffers restored to end-2021 nominal levels).
    - Full replenishment timing: 2025 in Spain, 2026 in the Netherlands, 2028 in France, and 2031 in Germany.
  - Phase 3: New steady state (historical payout ratios applied; capital and reserves assumed constant; profits not distributed flow to provisions).

*IMF WORKING PAPERS Raising Rates with a Large Balance Sheet: The Eurosystem’s Net Income and its Fiscal Implications*

### 43. In sum, it becomes clear that the current QE−QT cycle will likely result in long interruptions to

### VII. Conclusion

### Main findings on net income, losses, and fiscal implications
- Losses will be large and persistent at some NCBs, yet not so large and persistent as to necessitate capital support from member states.
- Losses are temporary in this cycle; alternative scenarios find:
  - Parallel upward shifts in yield curves improve net income slightly on a ten-year basis.
  - Faster QT modestly reduces net income improvement.
  - Lower demand for banknotes has a predictably negative impact on seigniorage.
  - Switching the reference rate in the Eurosystem’s income-pooling formulae from the MRO rate to the DFR would have a modest distributional effect across NCBs.
- With coupon income from the home sovereign as the key driver of the projected return to profits, NCBs in lower-debt countries with lower funding costs will, barring other factors, tend to see lower net income.
- The Eurosystem’s mutually agreed safeguards generate wide dispersion of financial results across NCBs through:
  - Capital key constraints keeping QE focused on the monetary policy stance.
  - Per-issue and per-issuer limits preserving fiscal incentives.
  - Pooling rules ensuring equitable sharing of monetary income.
  - Bond allocation and risk-sharing rules dictating a home bias in NCBs’ sovereign debt holdings, producing a home bias in NCBs’ interest earnings even after pooling.

### Fiscal interruptions to tax and profit payments (selected NCBs)
- Current QE−QT cycle will likely result in long interruptions to some NCBs’ tax and profit payments to national budgets:
  - Bundesbank: looks unlikely to make any payments to the German government for a total of 11 years (this includes 2020−21).
  - De Nederlandsche Bank: gap would be four years.
  - Banque de France: gap would be three years.
  - Banco de España: gap would be two years.
  - Banca d'Italia: corporate tax payments and profit transfers would dip to minimal levels in 2023 but avoid interruption.
- Fiscal impacts quantified:
  - Annual taxes and transfers of 0.1−0.2 percent of GDP facing interruptions, in one case for as long as 11 years.
- Table 3 selected averages (2008–31) as presented:
  - Bundesbank: 0.08, 0.00, 0.03; No. of yrs. old steady state loss-making & resumption without tax or transfer: 11
  - Banque de France: 0.17, 0.00, 0.19; No. of yrs.: 3
  - Banca d'Italia: 0.23, 0.12, 0.41; No. of yrs.: 0
  - Banco de España: 0.21, 0.00, 0.26; No. of yrs.: 2
  - De Nederlandsche Bank: 0.09, 0.00, 0.18; No. of yrs.: 4

### Policy suggestions and recommendations
- Suggestion to NCBs: consider adopting more-conservative profit distribution policies in the future steady state to smooth transfers to the state.
  - Paying more-moderate dividends when net income is positive can help NCBs maintain some dividends when pre-provision net income is negative.
  - Example: Banca d’Italia, Banco España, and De Nederlandsche Bank, having set aside relatively generous financial buffers as a ratio to total assets, are better placed to smooth their transfers to the state than some other NCBs.
  - Adopting more-conservative profit distribution norms would assist in the formulation of medium-term national budgetary plans.
- Four explicit policy conclusions:
  1. The temporary nature of losses in this cycle suggests the ECB should encourage NCBs to look through negative capital.
  2. Even without capital support from the state, fiscal impacts will be material, with annual taxes and transfers of 0.1−0.2 percent of GDP facing interruptions, in one case for as long as 11 years.
  3. More-conservative profit distribution norms in the future steady state could help mitigate the on-off pattern of dividends.
  4. Loss-making must remain orthogonal to monetary policy decision-making. If policymakers favor a faster pace of QT (to tighten the stance, adjust the operational framework, or support transmission), the specter of realizing valuation losses in outright bond sales should not be a consideration. Similarly, if other tools to reduce excess liquidity were to come into focus, it should be for monetary policy reasons alone.

### Methodology and key quantitative inputs (selected annex highlights)
- QE portfolio maturity:
  - Remaining weighted average maturity (WAM) on securities held for monetary policy purposes calculated at 7.1 years (assumed constant going forward and applied uniformly to the top-five NCBs and the Eurosystem).
  - APP and PEPP maturity distribution (December 2021) mid-point maturity assumptions: < 1 year = 0.5; 1–5 years = 2.5; 5–10 years = 7.5; > 10 years = 18.0.
- QE portfolio valuation and income assumptions:
  - Stock of bonds held by the Eurosystem for monetary policy purposes as of December 2021: €4.7 trillion (assumed to represent one unique portfolio; a 5 percent correction applied to obtain its par value).
  - Yearly interest income reported in the 2021 ECB Annual Report amounted to €1.006 billion (assumed to reflect the coupon payment of the aggregate representative bond).
  - ECB held €445.5 billion in securities for monetary policy purposes; implied interest rate on the representative bond equals 0.23 percent.
- TLTRO III assumptions and balances (May 2023, Table 5 totals):
  - Total drawings: 2,339 (€ bn.)
  - Repayments to date: 1,121 (€ bn.)
  - Amount outstanding: 1,218 (€ bn.)
  - TLTRO III operations carry a three-year maturity; interest-rate features described include special interest rate periods with rates reduced by 50 basis points and a −1 percent floor for banks meeting lending thresholds during special periods; participants could repay early from September 2021; conditions recalibrated October 27, 2022 to index remaining operations to average applicable key ECB interest rates effective November 23, 2022.
- Provisioning, capital, and profit distribution norms (selected figures as of end-2021):
  - ECB: GRF balance stood at zero at end-2021 while general risk provisions stood at €8.2 billion. Capital (€8.2 billion) + general risk provisions (€8.2 billion) = €16.4 billion.
  - Banque de France: Capital (€1 billion) + statutory reserves (€2 billion) = €3 billion.
  - Banca d'Italia: Capital = €7.5 billion; ordinary reserve standing at €6.5 billion.
  - Banco de España: Capital = €1 billion.
  - De Nederlandsche Bank: Capital = €8.5 billion.

### Final assessment
- The EMU institutional design and safeguards produce heterogeneous outcomes across NCBs, with home bias in holdings and earnings contributing to dispersion.
- While projected budgetary impacts from temporary Eurosystem loss-making are material in the years ahead, absent those policies a weaker macroeconomic environment might well have had a larger negative impact on the public finances.
- The ECB’s credibility will rest on its ability to deliver on its primary objective of price stability.

*IMF staff estimates, projections, and calculations as presented in the source document.*

### References

### References

### Major themes covered in the references
- Central bank finances, capital, and financial strength (for example: “Does Central Bank Capital Matter for Monetary Policy?”, “Central bank finances”, “Central bank financial strength and inflation: Is there a robust link?”, “Do central banks need capital?”).
- Profit, losses, and fiscal/quasi‑fiscal implications of central bank operations (for example: “Why are central banks reporting losses? Does it matter?”, “Annual loss for Swiss National Bank precludes profit distribution”, “Quasi-Fiscal Implications of Central Bank Crisis Interventions”).
- Asset purchase programs and balance sheet effects (for example: “Asset Purchase Programs and Financial Markets: Lessons from the Euro Area”, “Tracing the impact of the ECB’s asset purchase programme on the yield curve”, “Quantifying The Costs And Benefits Of Quantitative Easing”).
- Eurosystem / ECB governance, rules, and decisions (for example: “Decision (EU) 2016/2248... allocation of monetary income”, “Decision (EU) 2019/1743... remuneration of holdings of excess reserves”, “Guideline ECB/2019/7”, “Convergence report”).
- National central bank statutes, annual accounts, and official communications (examples include Banca d’Italia statutes and Annual Report 2021; Deutsche Bundesbank Annual Report 2021 and Annual Report 2022; De Nederlandsche Bank statutes and annual reports; Banco de España Annual Accounts 2021; Sveriges Riksbank Act).
- Market and public commentary on central bank losses and political/fiscal implications (examples include Financial Times pieces and Project Syndicate, Peterson Institute commentary).

### Key contributors and institutions cited
- International Monetary Fund: Stella (multiple working papers), IMF Working Paper WP/23/114.
- European Central Bank: multiple Decisions, Guidelines, Annual Accounts, Explainers, and press releases (2010, 2016, 2019a, 2019b, 2020, 2022a–2022g, 2023a–2023b).
- Bank for International Settlements: “Issues in the Governance of Central Banks”, “Central bank finances”, BIS Bulletins (2009, 2013, 2023a, 2023b).
- National central banks and authorities: Banca d’Italia, Banco de España, De Nederlandsche Bank, Deutsche Bundesbank, National Bank of Belgium, Swiss National Bank, Czech National Bank, Sveriges Riksbank.
- Academic and policy researchers and think tanks: Adler, Castro, and Tovar; Altavilla, Carbonia, and Motto; Levin, Lu, and Nelson; Gros and Shamsfakhr; Honohan; Perera, Ralston, and Wickramanayake; Sinn and Wollmershäuser.
- Media and public commentary: Financial Times (Arnold; Arnold and Chazan), Project Syndicate (De Grauwe and Ji), Jackson Hole Economic Symposium speech (Villeroy de Galhau), Central Banking Newsdesk.

### Document types, serial identifiers, and dates present
- IMF Working Papers and Notes: WP/97/83, WP/02/137, WP/08/37, WP/23/114, WP/2023/145.
- BIS Papers and Bulletins: BIS Papers No. 71; BIS Bulletin No. 68.
- ECB and Eurosystem legal instruments and publications: Decision (EU) 2016/2248; Decision (EU) 2019/1743; ECB/2019/31; Guideline ECB/2019/7; ECB Working Paper Series No. 2293.
- National legal instruments and reports with exact dates: Real Decreto 2059/2008, de 12 de diciembre; Bundesbank Act, July 2013; DNB Statutes, July 2019; Sveriges Riksbank Act, December 1, 2022.
- Specific dated items and press pieces: FEDS Notes, July 15, 2022; Financial Times, December 11, 2022; Financial Times, March 1, 2023; Swiss National Bank ad hoc announcement, January 9, 2023; Project Syndicate, February 27, 2023; Central Banking Newsdesk, March 1, 2023.

### Observations on coverage and emphasis
- Strong emphasis on euro-area and Eurosystem topics: multiple ECB decisions, Eurosystem income allocation rules, national central bank statutes and annual accounts.
- Recurrent focus on the interaction between central bank balance sheets, rising interest rates, and fiscal or distributive consequences (loss reporting, profit distribution, recapitalisation, and potential political pressure).
- Combination of academic, policy, institutional, and media sources spanning working papers, official statutes and accounts, central bank explainer pieces, and news commentary.

*Raising Rates with a Large Balance Sheet: The Eurosystem’s Net Income and its Fiscal Implications, Working Paper No. WP/2023/145*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2023/english/wpiea2023145-print-pdf.pdf_