## IMF Working Paper — Bank Profits and Bank Taxes in Europe (excerpt: 1. Introduction; 4. Some Trade-Offs in the Design of Bank Taxes; 6. Conclusions)

## Source details

**Canonical URL:** [IMF Working Paper — Bank Profits and Bank Taxes in Europe (excerpt: 1. Introduction; 4. Some Trade-Offs in the Design of Bank Taxes; 6. Conclusions)](https://www.imf.org/-/media/files/publications/wp/2024/english/wpiea2024143-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2024/english/wpiea2024143-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2024/english/wpiea2024143-print-pdf.pdf.json)

---

### Introduction — context, purpose, and headline findings
- Context and motivation
  - Since 2022, EU banks saw a significant increase in profits as economies emerged from the pandemic, inflation rose following the invasion of Ukraine, and monetary policy interest rates increased sharply.
  - Many European governments introduced new taxes on banks in response to large fiscal deficits and political-economy pressures.
  - Literature concern: bank taxes may adversely affect financial stability if banks do not retain earnings and are valued at a discount, making it challenging to accumulate buffers or raise new equity when needed.
- Purpose and scope
  - Document new bank taxes in Europe and highlight diversity in design (tax base, rate, duration, and burden).
  - Discuss trade-offs in bank tax design and argue that an alternative or complementary policy is to lock temporarily high bank profits into usable bank capital through an increase in countercyclical capital buffers (CCyB), where scope exists.
- Key analytical claims and findings
  - Recent high bank profits are predominantly related to delayed pass-through of policy interest rates to deposit interest rates and are likely transitory.
  - Deposit betas were around 25-30 percent during the 2022-23 ECB tightening cycle but at the 50 percent level during the 2005-07 tightening cycle.
  - Back-of-the-envelope estimate: a 4 percentage points increase in the ECB’s policy interest rates with 20 percent lower transmission to deposit rates and deposit financing representing about half of EU banks’ liabilities yields a (4 x 0.2 x 0.5) = 0.4 percentage point benefit for the cost of funding of EU banks from low deposit betas.
  - Absent this benefit, EU bank ROA over 2022-23 would have been in the range of 0.1-0.4 percent rather than 0.5-0.8 percent, and would be below or close to the 0.3 percent “normal times”-average over 2015-2019.
  - IMF analysis and other literature suggest low deposit betas reflect delayed pass-through rather than permanently low pass-through; as deposit markets adjust, deposit betas can catch up, increasing bank funding costs and reversing profitability.
  - Chen et al. (2024) estimate that by 2026 interest rate margins of European banks are likely to revert to close to 2015-2019 averages.
- Heterogeneity and risks
  - Aggregate NPL ratio in EU banks was about 1.8 percent in 2023 Q3, down from a peak of 6.8 percent in 2015; NPLs tend to peak about three years after a shock.
  - Bank profitability in 2022-23 is highly heterogeneous across EU countries:
    - CESEE countries tend to have above-average profitability; French and German banks have below-average profitability.
    - Statistically significant cross-sectional associations: banks more profitable where (4A) more depository funding; (4B) more adjustable-rate and fewer fixed-rate mortgages; (4C and 4D) higher concentration (HHI or S5); (4E) more cost-efficient (lower number of bank offices per capita); (4F) more conducive macroeconomic environment (higher nominal GDP growth).

### Trade-offs in the design of bank taxes
- Taxes on stocks (liabilities/assets) versus flows (profits/net revenue/NII)
  - Stocks (liabilities or assets) are relatively inert; flows (profits, net revenue, NII) are cyclical.
  - Taxes on liabilities or assets:
    - Provide fiscal revenue that is more stable over time.
    - May be particularly burdensome during downturns when profits are lower.
    - Tend to be used as longer-term or permanent measures.
  - Taxes on flows:
    - Are easier to manipulate via provisioning or non-distributable reserves and internal intragroup pricing.
    - May incentivize greater pass-through of policy rates to deposit rates (because higher interest expense lowers taxable profits/revenue), especially if temporary.
    - Could be offset by banks increasing margins or reducing deposit rates; net effects are ambiguous.
- Taxes on profits vs. net revenue vs. NII — definitions and incentive effects
  - Profits ≈ net revenue less operational and provisioning expenses.
  - Net revenue = NII + net non-interest income.
  - NII focuses most directly on the effects of monetary policy tightening (e.g., low deposit betas).
  - Incentive implications:
    - Tax on profits:
      - Reduces—but does not eliminate—incentives to invest in operational cost-efficiency.
      - Maintains incentives to make provisioning expenses because provisions reduce the tax base.
      - Is not more burdensome for banks during downturns (provisions deducted).
    - Tax on net revenue or NII:
      - Maintains bank incentives for cost-efficiency investments.
      - Tax base cannot be affected by provisioning and operational expense choices.
      - Tax on NII alone may induce substitution toward higher fees; net revenue tax is immune to substitution between interest and non-interest income.
- Taxes on “excess” vs. regular profits/net revenue/NII
  - Theoretical appeal: taxing excess profits (economic rent above normal return) can be efficient.
  - Practical limitations for banking:
    - Difficult to define “normal” vs. “excess” profits because bank profits are highly cyclical and structural profitability is changing.
    - Most estimates of the normal (or required) rate of return on bank capital are broadly in the range of 8 to 15 percent.
    - Aggregate ROE of EU banks has been below consistently 8% since the GFC.
    - If the normal-return threshold is unmet, the excess-profit tax rate will be zero; losses relative to the normal rate may be carried forward, producing frequent zero taxation and political controversy.
    - Some implementations used historical averages that included pandemic years, effectively lowering the normal benchmark.
  - Fiscal revenue from an excess-profits tax would be cyclical and potentially more unpredictable than regular profit taxes.
- Bank taxes versus higher CCyB rates
  - Concern: bank taxes draw on retained earnings that could have been allocated to capital.
  - Schemes allowing optional allocation of tax funds into non-available reserves (part of Tier 1 capital), as implemented in Italy, risk fungibility: non-available reserves prevent short-term payouts but other capital forms or new earnings can be paid as dividends later.
  - CCyB as alternative:
    - Increasing CCyB rates during a phase of temporarily high profits can lock profits into usable bank capital, increasing resilience to shocks.
    - If a tax on banks would not compromise banks’ ability to extend credit, an increase in CCyB of similar magnitude drawing on bank resources would similarly not necessarily compromise lending.
  - Empirical note: multiple countries that introduced new taxes on banks still have zero or low CCyB rates (rates notified to ECB for 2024), suggesting a lost opportunity to allocate temporarily high bank profits into high usable capital.
- Unremunerated reserve requirements (URR) as a de-facto tax
  - In the euro area:
    - Reserve requirements are defined as 1% of deposits and other debt liabilities with duration under 2 years (includes overnight deposits; excludes repo financing).
    - Total required reserves are around €170B, representing about 0.5% of bank assets and 1.5% of bank RWA.
    - A hypothetical increase in URR by 1pp when DFR is 4% would cost euro area banks €19B = 0.06% of RWA.
    - The €19B hypothetical revenue would more than offset the €8B loss recorded by the Eurosystem in 2023.
  - Considerations:
    - URR revenue accrues to the central bank (Eurosystem) rather than fiscal authorities.
    - Using URR as a fiscal tool is controversial: it may have unintended effects on monetary and financial stability objectives and could affect central bank independence.
    - Ad hoc increases in URR are a blunt instrument that do not differentiate between banks with different profitability and may raise regulatory-stability concerns akin to ad hoc new taxes.
- Bank taxes, profitability, capital, and cross-country heterogeneity
  - Observations from 2022 cross-country data:
    - All countries that introduced new bank taxes had above-average bank profitability as captured by ROA.
    - Several countries that introduced new bank taxes had below-average bank capital (Tier 1 capital), implying authorities could have aimed to convert temporarily high profits into bank capital instead.
  - Pre-existing bank taxes:
    - Substantial heterogeneity exists across EU countries in pre-existing bank taxes as a share of RWA and as a share of bank profits.
    - Many countries with new bank taxes already had above-average pre-existing bank taxes by the share-of-RWA metric.
    - Many countries with new bank taxes had above-average bank profitability, implying their pre-existing taxes as a share of bank profits were below-average; new taxes may have temporarily evened out revenues raised from bank taxes across the EU.
  - Implication: heterogeneity in bank taxation may contribute to an uneven playing field within the single market and the incomplete banking union.
- Macroeconomic effects of bank taxes
  - First-order and documented effects:
    - Loan rates increase and loan volumes decline (evidence from multiple European and other countries and DSGE models).
    - Lower lending reduces corporate investment and suppresses banks’ financial market activities including interbank lending and market-making.
    - Effects on bank risk-taking are ambiguous across studies.
  - Effects on depositors and market power:
    - Bank taxes may lower interest rates and raise fees for depositors; households may bear a disproportionate share as they are less price-sensitive.
    - Pass-through to depositors is more pronounced in concentrated markets.
    - If taxes target liabilities excluding deposits, deposit rates may instead increase because deposit funding becomes relatively more attractive.
  - Effects on shareholders and equity funding:
    - Bank taxes penalize shareholders and tend to induce negative stock market responses, reducing bank market value.
    - Given low price-to-book ratios in the EU, such effects may impede banks’ access to equity funding markets.
  - Cross-border spillovers:
    - New taxes weakening affected banks’ competitive positions can allow non-affected banks to increase margins, supporting arguments for more coordinated bank taxation within the Banking Union.

### Conclusions — implications and policy considerations
- Scope and limits
  - The paper documents recent trends in EU bank profits and new bank taxes, discusses trade-offs in design, and reviews literature on macroeconomic effects of bank taxes.
  - It does not reach welfare- or economic efficiency-related conclusions on the optimal extent of bank taxation.
- Key considerations for approach to bank taxation
  - Ad hoc taxes introduced in response to a surge in profits may hamper predictability of the business environment.
  - In countries where CCyB rates are low, bank taxes can be substituted or complemented by raising CCyB rates to lock unusual bank profits into releasable capital buffers.
  - Governments need to consider effects of bank taxes on monetary policy stance and transmission, as well as on financial stability.
- Summary trade-offs reiterated
  - Taxes on assets or liabilities:
    - Offer relatively stable fiscal revenue, maintain incentives for cost-efficiency investments, and are difficult to evade.
  - Taxes on profits, net revenue, or NII:
    - Are less burdensome during downturns (profits tax deducts loan-loss provisions).
    - Taxes on net revenue and NII maintain incentives for cost-efficiency investments.
    - All three may incentivize pass-through of policy rates to deposits, all else being equal.

*IMF Working Paper — excerpted chapters: 1. Introduction; 4. Some Trade-Offs in the Design of Bank Taxes; 6. Conclusions (wpiea2024143-print-pdf).*

### 1. Introduction ........................................................................................................

### 1. Introduction

### Context and motivation
- Since 2022, EU banks have experienced a significant increase in profits as economies emerged from the pandemic, inflation rose following the invasion of Ukraine, and monetary policy interest rates increased sharply.
- Large fiscal deficits and political-economy pressures have prompted many European governments to introduce new taxes on banks.
- Concerns noted in the literature: bank taxes may adversely affect financial stability if banks do not retain earnings and are valued at a discount, making it challenging to accumulate buffers or raise new equity when needed (Bochmann et al., 2023).

### Purpose and scope of the paper
- Document new bank taxes in Europe and highlight significant diversity in their design (tax base, rate, duration, and burden).
- Discuss trade-offs in bank tax design and argue that an alternative or complementary policy is to lock temporarily high bank profits into usable bank capital through an increase in countercyclical capital buffers (CCyB), where scope exists.
- Relates to preexisting literature on bank profitability in Europe (Pagano et al., 2014; Langfield and Pagano, 2016) and on bank taxation (International Monetary Fund, 2010).

### Key analytical claims and findings summarized
- Recent high bank profits are predominantly related to delayed pass-through of policy interest rates to deposit interest rates and are likely transitory.
- Deposit betas (ratio of the increase in deposit rates to the increase in policy rates) were around 25-30 percent during the 2022-23 ECB tightening cycle but at the 50 percent level during the 2005-07 tightening cycle (Adalid et al., 2023).
- Back-of-the-envelope estimate: a 4 percentage points increase in the ECB’s policy interest rates with 20 percent lower transmission to deposit rates and deposit financing representing about half of EU banks’ liabilities yields a (4 x 0.2 x 0.5) = 0.4 percentage point benefit for the cost of funding of EU banks from low deposit betas.
- Absent this benefit, EU bank ROA over 2022-23 would have been in the range of 0.1-0.4 percent rather than 0.5-0.8 percent, and would be below or close to the 0.3 percent “normal times”-average over 2015-2019.
- IMF analysis (Beyer et al., 2024) and other literature suggest the low deposit betas reflect delayed pass-through rather than permanently low pass-through; as deposit markets adjust, deposit betas can catch up, increasing bank funding costs and reversing profitability.
- Chen et al. (2024) estimate that by 2026 interest rate margins of European banks are likely to revert to close to 2015-2019 averages.

### Heterogeneity and risks highlighted
- Aggregate NPL ratio in EU banks was about 1.8 percent in 2023 Q3, down from a peak of 6.8 percent in 2015; low NPLs reduce provisioning and increase profits, but NPLs may increase with lagged impacts of the pandemic, war-driven cost shocks, and monetary tightening (NPLs tend to peak about three years after a shock; Ari et al., 2021).
- Bank profitability is highly heterogeneous across EU countries in 2022-23:
  - CESEE countries tend to have above-average profitability; French and German banks have below-average profitability.
  - Cross-sectional country associations (statistically significant) indicate banks are more profitable where:
    - (4A) banks have more depository funding;
    - (4B) there are more adjustable-rate and fewer fixed-rate mortgages;
    - (4C and 4D) the banking system is more concentrated (as captured by the HHI or the S5 measures);
    - (4E) banks are more cost-efficient (lower number of bank offices per capita);
    - (4F) the macroeconomic environment is more conducive to bank profitability (higher nominal GDP growth).

### Structure of the paper
- Section 2: Recent developments in bank profits in the EU (argues high profits are largely due to delayed deposit-rate pass-through and likely transitory).
- Section 3: Documents new bank taxes across EU countries, highlighting heterogeneity in tax base, rate, duration, and burden.
- Section 4: Discusses trade-offs in bank tax design.
- Section 5: Reviews literature on macroeconomic effects of bank taxes.
- Section 6: Conclusions.

*IMF Working Papers — Bank Profits and Bank Taxes in Europe (excerpt: 1. Introduction).*

### 4. Some Trade-Offs in the Design of Bank Taxes

### 4. Some Trade-Offs in the Design of Bank Taxes

### Trade-offs across tax bases: liabilities/assets vs. profits/revenue/NII
- Stocks (liabilities or assets) are relatively inert; flows (profits, net revenue, NII) are cyclical.
- Taxes on liabilities or assets:
  - Provide fiscal revenue that is more stable over time.
  - May be particularly burdensome during downturns when profits (and capacity to pay) are lower.
  - Tend to be used as longer-term or permanent measures; taxes on profits/revenue have been used more often as temporary (windfall) taxes.
- Taxes on flows (profits, net revenue, NII):
  - Are easier to manipulate via provisioning or non-distributable reserves and internal intragroup pricing.
  - May incentivize greater pass-through of policy rates to deposit rates (because higher interest expense lowers taxable profits/revenue), especially if temporary.
  - Could be offset by banks increasing margins or reducing deposit rates; net effects relative to no or lower bank taxes are ambiguous.

### Taxes on profits vs. net revenue vs. NII
- Definitions and incentive effects:
  - Profits ≈ net revenue less operational and provisioning expenses.
  - Net revenue = NII + net non-interest income.
  - NII focuses most directly on the effects of monetary policy tightening (e.g., low deposit betas).
- Incentive implications:
  - Tax on profits:
    - Reduces—but does not eliminate—incentives to invest in operational cost-efficiency (part of the efficiency gains are taxed away).
    - Maintains incentives to make provisioning expenses because provisions reduce the tax base, supporting timely recognition/resolution of NPLs.
    - Is not more burdensome for banks during downturns (provisions deducted).
  - Tax on net revenue or NII:
    - Maintains bank incentives for cost-efficiency investments.
    - Tax base cannot be affected by provisioning and operational expense choices.
    - Tax on NII alone may induce substitution toward higher fees (narrowing the tax base); net revenue tax is immune to substitution between interest and non-interest income streams.
- Summary characterization (as used in the source): Table 2 uses “XX” for strong effect and “X” for milder effect to compare these features across bases.

### Taxes on “excess” vs. regular profits, net revenue, and NII
- Theoretical appeal:
  - Taxing excess profits (economic rent above normal return to capital) can be efficient, targeting monopoly rents or super-normal profits and discouraging risk-taking behind unusually high profits.
- Practical limitations for banking:
  - Difficult to define “normal” vs. “excess” profits because bank profits are highly cyclical and the sector’s structural profitability is changing.
  - EU banks’ aggregate ROE has been below consistently 8% since the GFC, implying profits frequently fall short of a normal return threshold.
  - If the normal-return threshold is unmet, the excess-profit tax rate will be zero; losses relative to the normal rate may be carried forward, producing frequent zero taxation and potential political-economy controversy.
  - Fiscal revenue from an excess-profits tax would be cyclical and potentially more unpredictable than regular profit taxes.
- Practical observation:
  - In some European implementations, historical averages used to define “normal” profits included pandemic years (very low profits), effectively lowering the normal benchmark and appearing driven by desired revenue or political objectives rather than an economic normal-return concept.
- Context on normal return estimates:
  - Most estimates of the normal (or required) rate of return on bank capital are broadly in the range of 8 to 15 percent.
  - Aggregate ROE of EU banks has been below consistently 8% since the GFC.

### Bank taxes versus higher CCyB (countercyclical capital buffer) rates
- Concern: bank taxes draw on retained earnings that could have been allocated to capital.
  - Partial counterfactual: retained earnings could instead have been paid as shareholder payouts.
  - Schemes allowing optional allocation of tax funds into non-available reserves (part of Tier 1 capital), as implemented in Italy, risk fungibility: non-available reserves prevent short-term payouts but other capital forms or new earnings can be paid as dividends later, leaving eventual capital ratios unchanged.
- CCyB as alternative:
  - Increasing CCyB rates during a phase of temporarily high profits can lock profits into usable bank capital, increasing resilience to shocks.
  - If a tax on banks would not compromise banks’ ability to extend credit, an increase in CCyB of similar magnitude drawing on bank resources would similarly not necessarily compromise lending.
- Empirical note from the source:
  - Multiple countries that introduced new taxes on banks still have zero or low CCyB rates (rates notified to ECB for 2024), suggesting a lost opportunity to allocate temporarily high bank profits into high usable capital.

### Unremunerated reserve requirements (URR) as a de-facto tax
- Reserve requirements are primarily a systemic liquidity management tool and may have prudential and monetary effects; when unremunerated they function as a quasi-tax on banks.
- In the euro area:
  - Reserve requirements are defined as 1% of deposits and other debt liabilities with duration under 2 years (includes overnight deposits; excludes repo financing).
  - Total required reserves are around €170B, representing about 0.5% of bank assets and 1.5% of bank RWA.
  - A hypothetical increase in URR by 1pp when DFR is 4% would cost euro area banks €19B = 0.06% of RWA.
  - The €19B hypothetical revenue would more than offset the €8B loss recorded by the Eurosystem in 2023.
- Considerations:
  - URR revenue accrues to the central bank (Eurosystem) rather than fiscal authorities.
  - Using URR as a fiscal tool is controversial: it may have unintended effects on monetary and financial stability objectives and could affect central bank independence if used to strengthen its financial position.
  - Ad hoc increases in URR are a blunt instrument that do not differentiate between banks with different profitability and may raise concerns about regulatory stability akin to ad hoc new taxes.

### Bank taxes, profitability, capital, and heterogeneity across countries
- Observations from 2022 cross-country data:
  - All countries that introduced new bank taxes had above-average bank profitability as captured by ROA.
  - Several countries that introduced new bank taxes had below-average bank capital (Tier 1 capital), implying authorities could have aimed to convert temporarily high profits into bank capital instead.
- Pre-existing bank taxes:
  - Substantial heterogeneity exists across EU countries in pre-existing bank taxes as a share of RWA and as a share of bank profits.
  - Many countries with new bank taxes already had above-average pre-existing bank taxes by the share-of-RWA metric.
  - Many countries with new bank taxes had above-average bank profitability, implying their pre-existing taxes as a share of bank profits were below-average; new taxes may have temporarily evened out revenues raised from bank taxes across the EU.
- Implication: heterogeneity in bank taxation may contribute to an uneven playing field within the single market and the incomplete banking union.

### Macroeconomic effects of bank taxes
- First-order and documented effects:
  - Loan rates increase and loan volumes decline (evidence from multiple European and other countries and DSGE models).
  - Lower lending reduces corporate investment and suppresses banks’ financial market activities including interbank lending and market-making.
  - Effects on bank risk-taking are ambiguous: some studies find reduced risk-taking; others find increased risk-taking.
- Effects on depositors and market power:
  - Bank taxes may lower interest rates and raise fees for depositors; households may bear a disproportionate share as they are less price-sensitive.
  - Pass-through to depositors is more pronounced in concentrated markets where banks can pass costs on customers.
  - If taxes target liabilities excluding deposits, deposit rates may instead increase because deposit funding becomes relatively more attractive.
- Effects on shareholders and equity funding:
  - Bank taxes penalize shareholders and tend to induce negative stock market responses, reducing bank market value.
  - Given low price-to-book ratios in the EU, such effects may impede banks’ access to equity funding markets.
- Cross-border spillovers:
  - New taxes weakening affected banks’ competitive positions can allow non-affected banks to increase margins (cross-border spillovers), supporting arguments for more coordinated bank taxation within the Banking Union.

*Source: IMF Working Paper chapter "4. Some Trade-Offs in the Design of Bank Taxes" (wpiea2024143-print-pdf).*

### 6. Conclusions

### 6. Conclusions

### Summary of scope and limits
- Documented recent trends in the profits of EU banks and the new bank taxes in the EU.
- Discussed several trade-offs in the design of bank taxes.
- Reviewed the literature on the potential macroeconomic effects of bank taxes.
- Did not attempt to arrive at welfare- or economic efficiency-related conclusions relating to the optimal extent of bank taxation; such conclusions remain outside the scope of this analysis.

### Key considerations for overall approach to bank taxation
- New taxes introduced in an ad hoc manner in response to a surge in profits may be undesirable because they may hamper the predictability of the business environment.
- In countries where CCyB rates are low, bank taxes can be substituted or complemented by raising CCyB rates to lock in unusual bank profits into releasable capital buffers.
- Governments need to consider the effects of bank taxes on monetary policy stance and transmission, as well as on financial stability (see European Central Bank, 2022, and references therein).

### Trade-offs in the design of bank taxes
- Taxes on assets or liabilities:
  - Offer relatively stable fiscal revenue.
  - Maintain bank incentives for cost-efficiency investments.
  - Are difficult to evade.
- Taxes on profits, net revenue, or NII:
  - Are less burdensome for banks during downturns (especially the tax on profits as it deducts loan-loss provisions from tax base).
  - Taxes on net revenue and NII maintain bank incentives for cost-efficiency investments.
  - All three (tax on profits, net revenue, NII) may incentivize the pass-through of policy rates to deposits, all else being equal.

*IMF Working Paper — 6. Conclusions*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2024/english/wpiea2024143-print-pdf.pdf_