## _sdn1604

## Source details

**Canonical URL:** [_sdn1604](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/sdn/2016/_sdn1604.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/sdn/2016/_sdn1604.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/sdn/2016/_sdn1604.pdf.json)

---

### Overview and scope
- The paper assesses the benefits of bank capital in terms of its ability to absorb losses and measures these benefits against how much capital would have been needed to avoid imposing losses on bank creditors or resorting to public recapitalizations in past banking crises.
- It reviews evidence on the welfare costs of tighter capital regulation through potential impacts on bank credit and lending rates.
- The analysis follows a macro approach and relies on country-level data on nonperforming loans and loss given default ratios in past banking crises, and on balance sheet parameters to convert losses into capital needs.
- The paper’s findings support the range of loss absorbency suggested by the Financial Stability Board (FSB) and the Basel Committee for systemically important banks.

### Methodology and caveats
- Macro-level analysis due to dearth of bank-level data on losses during crises.
- Reliance on the precision and comparability of available country-level data on nonperforming loans and loss given default ratios, and on other balance sheet parameters.
- Different exercises produce similar conclusions despite data limitations.
- The paper’s conclusions are described as tentative given the above caveats.
- Baseline four-step approach to translate past losses into required capital:
  - Compile peak NPL ratios during banking crises.
  - Convert NPLs into loan losses using LGD proxy (Mean LGD in U.S. over 1970–2003: 50 percent on average in normal times and up to 75 percent in downturns).
  - Net out loan loss provisions (historic loan loss provisioning averaged about 1.5 percent, used as baseline).
  - Add 1 percent of capital as a margin of safety and convert unweighted capital needs into risk-weighted capital by applying a Total assets / RWA ratio of 1.75 (average for U.S. banks).
- Baseline formula:
  - Bank capital = (NPL * LGD – Provisions + 1 percent) * (Total assets / RWA)
- Caveats noted:
  - Considerable uncertainty about coefficient magnitudes; discuss ranges not point estimates.
  - Conversion of leverage ratios into risk-weighted capital varies across countries; baseline uses conversion ratio of 1.75.
  - Analysis equates bank losses with loan losses and assumes uniform accrual across assets; may overestimate when losses concentrated on loans and underestimate when concentrated on securities.
  - Does not incorporate nonlinear multipliers from bank-sovereign linkages in fiscally vulnerable settings.
  - Based on average NPLs in banking system; losses at individual banks may differ significantly.

### Key empirical findings
- Main result (advanced economies):
  - Marginal benefits of increases in capital are high initially but decline rapidly once banks’ risk-weighted capital ratios reach the 15–23 percent level (depending on assumptions).
  - A loss-absorption capacity of 15–23 percent of risk-weighted assets (RWA) would have been sufficient to absorb losses in a large proportion of banking crises in advanced economies.
  - Protecting against the most extreme crises would have required substantially more capital; such extreme crises are very rare in advanced economies.
- Specific illustrative calculations and parameters (from baseline schedule):
  - Example: hypothetical 18 percent NPL ratio => 9 percent loan losses (with LGD = 50 percent).
  - Loan losses net of provisions = 7.5 percent of total assets (assuming 1.5 percent provisions).
  - Adding a 1 percent margin of safety gives an 8.5 percent leverage ratio.
  - With Total assets / RWA = 1.75, this corresponds to approximately a 15 percent risk-weighted capital ratio.
- Sensitivity:
  - With loss given default = 75 percent, risk-weighted capital necessary to absorb losses in 85 percent of historical crises in OECD countries rises from 15 percent to about 23 percent.
  - Higher conversion ratio example: conversion ratio = 2.5 leads to a capital ratio of 21 percent.
  - Table-derived baseline metrics (example columns):
    - NPLs during a Banking Crisis: 18.0 18.0 18.0 18.0
    - Loss Given Default: 50.0 75.0 50.0 50.0
    - Loan Losses (Mean point): 9.0 13.5 9.0 9.0
    - Absorbed by Prior Provisioning: 1.5 1.5 1.5 1.5
    - Loan Losses Net of Provisions (3 ‒ 4): 7.5 12.0 7.5 7.5
    - Margin of Safety (Residual Capital): 1.0 1.0 1.0 3.0
    - Capital to Assets Ratio or Leverage Ratio (5 + 6): 8.5 13.0 8.5 10.5
    - Total Assets/Risk-Weighted Assets: 175.0 175.0 250.0 175.0
    - Capital Ratio (percent of Risk-Weighted Assets) (7 * 8): 14.9 22.8 21.3 18.4

### Cross-country differences and OECD vs non-OECD
- Finding: Higher NPL ratios in banking crises in non-OECD countries drive higher capital-need estimates.
- Evidence and magnitudes:
  - Capital ratios in the 15–23 percent range would have been sufficient to absorb losses in only about half of all banking crises once non-OECD countries are considered.
  - In 2010, the minimum capital ratios in OECD countries were almost uniformly 8 percent.
  - The median minimum capital ratio in non-OECD countries was 10 percent in 2010.
  - Almost a quarter of non-OECD countries had a minimum capital ratio of 12–15 percent.
- Offsets and context:
  - Banks in non-OECD countries tend to hold fewer risky securities on their balance sheets than banks in OECD countries, making NPL losses a more comprehensive loss statistic and the capital-needs estimates more conservative for non-OECD countries.
  - Non-OECD countries tend to have much smaller banking systems relative to GDP; if bank losses exceeded capital absorption capacity, their fiscal impact is likely smaller.
  - Example: Had non-OECD countries imposed bank capital ratios in the 15–23 percent range, in 80 percent of banking crises, losses exceeding the absorption capacity of capital would have been within 3 percent of GDP.
- Policy implication: Complement higher capital ratios with institutional improvements to reduce potential NPLs; desirable bank capital levels in OECD and non-OECD countries might be closer than raw estimates suggest.

### Capital sufficient to avoid public recapitalizations
- Methodology summary:
  - Assume postcrisis public recapitalizations brought banks to the minimum capital level needed to restore viability.
  - Combine Bankscope average capital ratios in 2007 for crisis countries (2007–13) with Laeven and Valencia (2013) fiscal outlays for recapitalizations; express variables as percentages of total RWA.
  - Sum precrisis bank capital levels and public bank recapitalization injections to infer hypothetical precrisis capital that would have avoided recapitalizations.
- Findings and magnitudes:
  - Marginal benefit of additional capital in avoiding public recapitalizations is relatively high until 15–17 percent risk-weighted capital ratios, which help avoid public recapitalizations in 75 percent of banking crises.
  - Capacity to avoid public recapitalizations is lower in non-OECD countries.
  - Bank-level evidence (U.S. and Europe, available data):
    - A capital ratio of 15 percent in 2007 would have avoided the need for capital injection in almost 55 percent of cases in the United States (based on available sample) and 75 percent of cases in Europe.
    - A capital ratio of 23 percent would have eliminated the need for injection in virtually all observed cases.
- Data limitations: incompleteness of European bank injection data (covers November 2008–January 2010); U.S. injections data based on TARP.

### Benefits and costs of higher bank capital
- Benefits:
  - Capital absorbs losses, reduces probability of bank failure, protects creditors and taxpayers.
  - Improves incentives for better risk management by increasing shareholders’ skin in the game.
  - Helps attract funds, maintain long-term customer relationships, and carry essential lending risks.
- Costs:
  - Distinguish transition costs versus steady-state costs.
  - Transition costs:
    - Equity issuance underwriting fees usually of 5–7 percent.
    - Signaling discounts when issuing equity (Myers and Majluf 1984).
    - Rapid adjustment may reduce credit and harm macroeconomic performance.
  - Steady-state costs:
    - Tax treatment differences favoring debt (De Mooij 2011).
    - Loss of funding advantage inherent in safe and liquid bank debt; empirical work documents funding spreads (e.g., negative 50 to 70 basis points premium for Treasuries relative to other bonds).
    - Destruction of cash-investor surplus when replacing debt with equity (Song and Thakor 2007; DeAngelo and Stulz 2013).
  - Systemic considerations:
    - Higher capital reduces systemic spillovers from bank distress.
    - Distribution of capital across banks matters for aggregate credit and acquisition of weak institutions.
    - Excessively high capital requirements may push activities to less-regulated parts of the financial system (raising systemic risk).
- Empirical estimates on lending rates:
  - U.S. panel regressions (1996–2011): a 1 percentage point higher Tier 1 capital ratio is associated with loan rates that are 2.5 basis points higher.
  - Literature range for impact of a 1 percentage point increase in capital requirements on lending rates: 2 basis points to 20 basis points.
  - Historical evidence: spreads between reference and lending rates were not higher in periods when banks were much more highly capitalized (Miles, Yang, and Marcheggiano 2012).
- Interpretation qualifiers:
  - Modigliani-Miller offset may reduce steady-state cost if higher capital lowers firm risk and borrowing costs.
  - Empirical estimates often reflect small endogenous capital variations; exogenous regulatory increases could have larger effects.
  - Costs may increase nonlinearly with higher capital levels.
  - General equilibrium effects (role of bank debt as safe asset; migration to shadow banking) could matter materially.

### Recent regulatory changes and calibration context
- Basel III introduced:
  - a leverage ratio requirement (Leverage Ratio = 3.0).
  - higher quality of capital by requiring a larger fraction of tangible common equity and tightening eligibility of instruments that count as capital.
  - Capital Conservation Buffer: 2.5 percent.
  - Countercyclical Buffer: 0–2.5 percent.
  - Minimum Total Capital Plus Conservation Buffer: 10.5 percent.
  - Minimum Total Capital Plus Conservation Buffer, Countercyclical Buffer, and G-SIB Charge: 11.5–15.5 percent.
  - Minimum Common Equity Capital: 4.5 percent.
  - Minimum Tier 1 Capital: 6.0 percent.
  - Hybrid capital instruments with incentive to redeem: Not eligible under Basel III.
- Some jurisdictions adopted higher standards (example provided: Switzerland enforcing 19 percent capital ratios for its largest banks).
- Observed impact: capital ratios for large banks in advanced economies in America, Europe, and Asia have increased.

### Comparison with stress tests and robustness
- 2015 U.S. stress test (FRB 2015), “severely adverse scenario”: 4.5 percent loan losses and a loss of 5.5 percentage points of risk-weighted bank capital.
- 2014 European Banking Authority stress test (EBA 2014), “adverse scenario”: 2.3 percent loan losses that induced a loss of 4.4 percentage points of bank capital.
- Loan losses corresponding to the 85th percentile of historical banking crises (used in this analysis) are two to four times as high as losses in the U.S. and European stress tests.
- Analysis employs coefficient of 1.75 to convert loan losses into losses in risk-weighted capital (between coefficients used in referenced stress tests).

### Insights from general equilibrium models (Section IV.C)
- Calibrated models typically suggest optimal bank capital can range from 8 to 20 percent.
- The 15–23 percent bank capital ratio discussed in Sections III.A and B is near the top of the range of optimal levels identified in general equilibrium models but is consistent with their predictions.
- Sources of variation in model estimates include substitution between bank funding and direct lending, changes in liquidity provision quality, and “last bank standing” scarcity rents.
- Given model uncertainty and parameter sensitivity, the 15–23 percent figure should be seen as consistent with, but not uniquely implied by, general equilibrium calibrated results.

### Policy recommendations and transitional guidance
- New regulatory minima should be imposed gradually, when conditions allow, and over a relatively long period of time, to mitigate large short-term transitional costs.
- Supervisors should promote equity-based increases in loss-absorption capacity (new issuance or retained earnings) rather than asset contraction to avoid reduced credit availability.
- When regulators set new minimum capital requirements based on solid economic and financial foundations, markets tend to anticipate full compliance with new standards ahead of phase-in periods.
- Widen the perimeter of prudential and macroprudential regulation to address regulatory arbitrage and risks migrating to the shadow banking system.
- Complement higher capital requirements with institutional improvements in regulation, supervision, resolution, and governance to reduce expected bank losses.

*Source: IMF staff analysis as presented in the chapter "Benefits and Costs of Bank Capital" (SDN1604).*

### Executive Summary.......................................................................................................

### _sdn1604 - Executive Summary

### Overview and scope
- The paper assesses the benefits of bank capital in terms of its ability to absorb losses and measures these benefits against how much capital would have been needed to avoid imposing losses on bank creditors or resorting to public recapitalizations in past banking crises.
- It also reviews evidence on the welfare costs of tighter capital regulation through potential impacts on bank credit and lending rates.
- The analysis follows a macro approach and relies on country-level data on nonperforming loans and loss given default ratios in past banking crises, and on balance sheet parameters to convert losses into capital needs.
- The paper’s findings support the range of loss absorbency suggested by the Financial Stability Board (FSB) and the Basel Committee for systemically important banks.

### Methodology and caveats
- Macro-level analysis due to dearth of bank-level data on losses during crises.
- Reliance on the precision and comparability of available country-level data on nonperforming loans and loss given default ratios, and on other balance sheet parameters.
- Different exercises produce similar conclusions despite data limitations.
- The paper’s conclusions are described as tentative given the above caveats.

### Key findings
- Capital in the range of 15–23 percent of risk-weighted assets would have been sufficient to absorb losses in the majority of past banking crises (at least in advanced economies).
- Further capital increases beyond this range would have had only marginal effects on preventing additional crises, suggesting that this level of loss-absorption capacity is, on average, appropriate for advanced economies.
- From a regulatory standpoint, appropriate capital requirements may be below this range because:
  - Banks tend to hold capital in excess of regulatory minima.
  - Other bail-in-able instruments can contribute to loss-absorption capacity.
- Institutional and regional factors can lead to variations in appropriate levels of capital and loss-absorption capacity across jurisdictions:
  - Emerging markets have, on average, suffered greater bank losses (relative to bank assets but not to GDP) during crises and have imposed tighter requirements.
  - The extent to which institutional improvements (in regulation, supervision, resolution, and governance) can reduce required levels of loss absorption is an open question.
- Short-term costs of transitioning to higher capital standards may be substantial and much higher than long-term costs.
- When regulators set new minimum capital requirements based on solid economic and financial foundations, markets tend to anticipate full compliance with new standards ahead of phase-in periods.
- Supervisors should encourage banks to increase loss absorption by raising equity (through new issuance or retained earnings) rather than shrinking assets, to avoid reduced credit availability.
- Tighter bank requirements may increase incentives for regulatory arbitrage and migration of activities to unregulated or less regulated financial intermediaries (the so-called shadow banking system), highlighting the importance of widening the perimeter of prudential and macroprudential regulation.

### Policy recommendations and transitional guidance
- New regulatory minima should be imposed gradually, when conditions allow, and over a relatively long period of time, to mitigate large short-term transitional costs.
- Supervisors should promote equity-based increases in loss-absorption capacity (new issuance or retained earnings) rather than asset contraction.
- Widen the perimeter of prudential and macroprudential regulation to address regulatory arbitrage and risks migrating to the shadow banking system.

*International Monetary Fund — Executive Summary from the content unit "_sdn1604 - Executive Summary"*

### 15.5 percent of risk-weighted assets, when all surcharges are activated (see Table 1). It also

### _sdn1604 - 15.5 percent of risk-weighted assets, when all surcharges are activated (see Table 1). It also

### Overview of recent regulatory changes
- Basel III introduced:
  - a leverage ratio requirement (Leverage Ratio = 3.0).
  - higher quality of capital by requiring a larger fraction of tangible common equity and tightening eligibility of instruments that count as capital.
  - Capital Conservation Buffer: 2.5 percent.
  - Countercyclical Buffer: 0–2.5 percent.
  - Minimum Total Capital Plus Conservation Buffer: 10.5 percent.
  - Minimum Total Capital Plus Conservation Buffer, Countercyclical Buffer, and G-SIB Charge: 11.5–15.5 percent.
  - Minimum Common Equity Capital: 4.5 percent.
  - Minimum Tier 1 Capital: 6.0 percent.
  - Hybrid capital instruments with incentive to redeem: Not eligible under Basel III.
- Some jurisdictions adopted higher standards (example: Switzerland enforcing 19 percent capital ratios for its largest banks).
- Observed impact: capital ratios for large banks in advanced economies in America, Europe, and Asia have increased (Figure 1).

### Research focus and approach
- Primary question addressed: how much capital would have been sufficient to absorb all bank losses through equity in past banking crises (i.e., prevent bank creditors or taxpayers from losing money)?
- Rationale for approach:
  - Reduces number of assumptions relative to general equilibrium welfare-maximization exercises.
  - Conservative framing: focuses on equity as loss-absorption and abstracts from other bail-in-able instruments (TLAC reinterpretation possible).

### Empirical findings on sufficient capital levels
- Main result (advanced economies):
  - Marginal benefits of increases in capital are high initially but decline rapidly once banks’ risk-weighted capital ratios reach the 15–23 percent level (depending on assumptions).
  - A loss-absorption capacity of 15–23 percent of risk-weighted assets (RWA) would have been sufficient in a large proportion of banking crises in advanced economies.
  - Protecting against the most extreme crises would have required substantially more capital; such extreme crises are very rare in advanced economies.
- Asymmetry with emerging and developing economies:
  - The discontinuity in marginal benefits is less pronounced and occurs at a higher capital level for emerging and developing economies, where banking crises have more often been associated with large bank losses.

### Methodology to translate past losses into required capital
- Four-step baseline approach:
  1. Compile peak NPL ratios during banking crises (Laeven and Valencia 2013; Figure 2).
  2. Convert NPLs into loan losses using loss given default (LGD) proxy:
     - Mean LGD in U.S. over 1970–2003: 50 percent on average in normal times and up to 75 percent in downturns.
  3. Net out loan loss provisions:
     - Historic loan loss provisioning averaged about 1.5 percent (used as baseline provision).
  4. Add 1 percent of capital as a margin of safety and convert unweighted capital needs into risk-weighted capital by applying a Total assets / RWA ratio of 1.75 (average for U.S. banks).
- Baseline illustrative calculation:
  - Example: hypothetical 18 percent NPL ratio => 9 percent loan losses (with LGD = 50 percent).
  - Loan losses net of provisions = 7.5 percent of total assets (assuming 1.5 percent provisions).
  - Adding a 1 percent margin of safety gives an 8.5 percent leverage ratio.
  - With Total assets / RWA = 1.75, this corresponds to approximately a 15 percent risk-weighted capital ratio.
- Baseline formula:
  - Bank capital = (NPL * LGD – Provisions + 1 percent) * (Total assets / RWA)

### Interpretation and caveats regarding regulatory application
- Results relate to levels of bank capital rather than minimum capital requirements:
  - Banks typically maintain buffers over minimum requirements and can draw on these in stressed periods; thus regulatory minima can be below the range identified.
- Other bail-in-able instruments can contribute to loss-absorption capacity:
  - From a regulatory standpoint, appropriate capital requirements may be below the estimated capital range if other TLAC instruments are recognized.
- The paper abstracts from instrument composition debates; focusing on bank capital is the most conservative approach but results may be reinterpreted as applying to TLAC instruments.
- Estimates are conservative because they do not incorporate possible incentive effects (e.g., “skin-in-the-game” reducing risk-taking), which would raise the benefits of capital and could lower the required capital to absorb losses.

### Benefits and costs of higher bank capital (summary)
- Benefits:
  - Capital absorbs losses, reduces probability of bank failure, protects creditors and taxpayers.
  - Improves incentives for better risk management by increasing shareholders’ skin in the game.
  - Helps attract funds, maintain long-term customer relationships, and carry essential lending risks.
- Costs:
  - Distinguish transition costs versus steady-state costs.
  - Transition costs:
    - Equity issuance underwriting fees usually of 5–7 percent.
    - Signaling discounts when issuing equity (Myers and Majluf 1984).
    - Rapid adjustment may reduce credit and harm macroeconomic performance.
  - Steady-state costs:
    - Tax treatment differences favoring debt (De Mooij 2011).
    - Loss of funding advantage inherent in safe and liquid bank debt; estimated funding cost advantage captured in empirical work (e.g., negative 50 to 70 basis points premium for Treasuries relative to other bonds).
    - Destruction of cash-investor surplus when replacing debt with equity (Song and Thakor 2007; DeAngelo and Stulz 2013).
  - Systemic considerations:
    - Higher capital reduces systemic spillovers from bank distress.
    - Distribution of capital across banks matters for aggregate credit and acquisition of weak institutions.
    - Excessively high capital requirements may push activities to less-regulated parts of the financial system (raising systemic risk).

### Summary policy implications
- A pragmatic loss-absorption perspective suggests:
  - In advanced economies, targeting loss-absorption capacity in the 15–23 percent of RWA range would have sufficed for most historical crises.
  - For emerging and developing economies, higher capital buffers may be required due to larger and more frequent banking losses.
  - Complementarity of capital requirements with institutional improvements (regulation, supervision, resolution) is crucial to reduce expected bank losses.
  - Regulatory design should account for other TLAC instruments, banks’ buffer behavior over minima, transition costs, and potential migration of activities to shadow banking.

*Source: IMF staff analysis as presented in the chapter "Benefits and Costs of Bank Capital" (SDN1604).*

### 1. NPLs duri ng a Banki ng Cri si s18.018.018.018.0

### 1. NPLs duri ng a Banki ng Cri si s18.018.018.018.0

### Key parameters and computed metrics
- Loss Gi ve n De f aul t: 50.075.050.050.0
- Loan Losse s ( 1 * 2)  ( Me an poi nt): 9.013.59.09.0
- Absorbed by Prior Provisioning: 1.51.51.51.5
- Loan Losses Net of Provisions (3 ‒ 4): 7.512.07.57.5
- Margin of Safety (Residual Capital): 1.01.01.0 3.0
- Capital to Assets Ratio or Leverage Ratio  (5 + 6): 8.513.08.510.5
- Total  Asse ts/Ri sk-We i ghte d Asse ts: 175.0175.0 250.0 175.0
- Capi tal  Rati o ( pe rce nt of  Ri sk-We i ghte d Asse ts)  ( 7 * 8): 14.922.821.318.4

### Main findings from the baseline schedule and sensitivity
- Baseline (OECD countries): marginal benefit of additional capital from a loss-absorption point of view is relatively high until 15 percent risk-weighted capital ratio (which enables banks to absorb losses in 85 percent of banking crises, assuming a loss given default ratio of 50 percent).
- Marginal benefit of additional capital declines rapidly after 15 percent risk-weighted capital ratio.
- With loss given default = 75 percent (conceivable in systemic crises), the risk-weighted capital necessary to absorb losses in 85 percent of historical crises in OECD countries rises from 15 percent to about 23 percent (red line in Figure 3; second column of Table 2).
- Higher conversion ratio example: conversion ratio = 2.5 leads to a capital ratio of 21 percent (third column of Table 2).
- Higher-than-average recovery rates would lead to lower capital estimates; conversely, higher loss given default rates raise required capital.

### Uncertainty, assumptions, and methodological caveats
- Considerable uncertainty exists about the magnitude of some coefficients; “sufficient” loss-absorbing capital is better discussed as ranges than point estimates.
- Conversion of raw leverage ratios into risk-weighted capital ratios varies across countries (Le Lesle and Avramova 2012). Baseline uses conversion ratio of 1.75 consistent with U.S. data for loss given default estimates.
- If higher conversion ratios corresponded to safer bank portfolios, they would arguably have to be matched with lower loss given default estimates, partly compensating for direct effects.
- Strategic risk weighting by banks would affect the “true” conversion ratio.
- Choice of the safety margin is arbitrary; reasonable variations will not significantly affect the range of required bank capital.
- Analysis equates bank losses with loan losses and assumes uniform accrual across assets:
  - Tends to overestimate capital needs when losses are concentrated on loans.
  - Tends to underestimate capital needs when losses are concentrated on securities.
- Benchmark parameters for loan losses exceed losses on securities in the United States during the global financial crisis (Berrospide 2013) and losses assumed in the Dodd-Frank stress tests under the “severely adverse scenario.”
- Some past-crisis losses due to interbank exposure may cause methodology to overestimate capital buffer needs.
- Capital buffers would need to be higher in the presence of nonlinear multipliers associated with bank-sovereign linkages in fiscally vulnerable settings (these nonlinearities are abstracted from in this analysis).
- Analysis is based on average NPL ratios in the banking system; losses at individual banks may differ significantly from the country mean, so methodology may underestimate capital necessary to preserve positive equity across the entire system.
- Average level of capital needed to maintain positive equity remains informative for overall banking system capacity and authorities’ ability to confront crises, including facilitating takeovers of weaker banks by stronger ones.

### Comparison with stress tests and robustness context
- 2015 U.S. stress test (FRB 2015), “severely adverse scenario”: 4.5 percent loan losses and a loss of 5.5 percentage points of risk-weighted bank capital.
- 2014 European Banking Authority stress test (EBA 2014), “adverse scenario”: 2.3 percent loan losses that induced a loss of 4.4 percentage points of bank capital.
- Loan losses corresponding to the 85th percentile of historical banking crises (used in this analysis) are two to four times as high as losses in the U.S. and European stress tests.
- Analysis employs coefficient of 1.75 to convert loan losses into losses in risk-weighted capital (between coefficients used in referenced stress tests).

*Source: IMF staff calculations. BENEFITS AND COSTS OF BANK CAPITAL, INTERNATIONAL MONETARY FUND*

### 1.25 and 1.9 implied by U.S. and European stress tests. (A smaller coefficient in U.S. stress tests is a product of low

### _sdn1604 - 1.25 and 1.9 implied by U.S. and European stress tests. (A smaller coefficient in U.S. stress tests is a product of low

### Cross-country differences in capital needs (OECD vs. non-OECD)
- Finding: Higher NPL ratios in banking crises in non-OECD countries drive higher capital-need estimates.
- Evidence and magnitudes:
  - Capital ratios in the 15–23 percent range would have been sufficient to absorb losses in only about half of all banking crises once non-OECD countries are considered.
  - In 2010, the minimum capital ratios in OECD countries were almost uniformly 8 percent.
  - The median minimum capital ratio in non-OECD countries was 10 percent in 2010.
  - Almost a quarter of non-OECD countries had a minimum capital ratio of 12–15 percent.
- Caveats and offsets:
  - Banks in non-OECD countries tend to hold fewer risky securities on their balance sheets than banks in OECD countries (e.g., during the 2008 crisis), making NPL losses a more comprehensive loss statistic and the capital-needs estimates more conservative for non-OECD countries.
  - Estimation parameters (loss given default, provisions, conversion rates) for non-OECD countries are subject to greater uncertainty and heterogeneity.
  - Non-OECD countries tend to have much smaller banking systems relative to GDP than OECD countries; when bank losses exceed capital absorption capacity, their impact on the economy and fiscal accounts is likely smaller.
    - Example: Had non-OECD countries imposed bank capital ratios in the 15–23 percent range, in 80 percent of banking crises, losses exceeding the absorption capacity of capital would have been within 3 percent of GDP.
  - Policy implication: Complement higher capital ratios with institutional improvements (regulation, supervision, resolution) to reduce potential NPLs; desirable bank capital levels in OECD and non-OECD countries might be closer than raw estimates suggest.

### Capital sufficient to avoid public recapitalizations
- Methodology summary:
  - Working assumption: postcrisis public recapitalizations brought banks to the minimum capital level needed to restore viability.
  - Data combined: Bankscope average capital ratios in 2007 for crisis countries (2007–13) and Laeven and Valencia (2013) fiscal outlays for recapitalizations; variables expressed as percentages of total risk-weighted assets (RWA).
  - The sum of precrisis bank capital levels and public bank recapitalization injections used to infer hypothetical precrisis capital that would have avoided recapitalizations.
- Findings and magnitudes:
  - Marginal benefit of additional capital in avoiding public recapitalizations is relatively high until 15–17 percent risk-weighted capital ratios, which help avoid public recapitalizations in 75 percent of banking crises.
  - Capacity of bank capital to avoid public recapitalizations is lower in non-OECD countries.
  - Bank-level evidence (U.S. and Europe, available data):
    - A capital ratio of 15 percent in 2007 would have avoided the need for capital injection in almost 55 percent of cases in the United States (based on available sample) and 75 percent of cases in Europe.
    - A capital ratio of 23 percent would have eliminated the need for injection in virtually all observed cases.
  - Data limitations: incompleteness of European bank injection data; European injection data cover injections only between November 2008 and January 2010; U.S. injections data based on TARP.

### How costly are higher capital requirements?
- Conceptual separation:
  - Transitional effects (short-term) versus steady-state consequences (long-term).
  - Transitional-cost literature often uses bank-level shocks (losses or idiosyncratic regulatory actions) and may overstate steady-state costs because shocks typically affect distressed banks.
- Steady-state evidence (empirical estimates and magnitudes):
  - U.S. panel regressions (1996–2011) on loan-level data: a 1 percentage point higher Tier 1 capital ratio is associated with loan rates that are 2.5 basis points higher (Table 3).
    - Note: these regressions do not identify causality and largely reflect small variations in bank capital not associated with regulatory actions.
  - Literature range for impact of a 1 percentage point increase in capital requirements on lending rates: 2 basis points to 20 basis points (Table 4A).
  - Historical prima facie evidence: spreads between reference and lending rates were not higher in periods when banks were much more highly capitalized (Miles, Yang, and Marcheggiano 2012).
- Interpretation and caveats:
  - The Modigliani-Miller (MM) theorem implies a potential offset: higher capital reduces firm risk and cost of borrowing, possibly offsetting higher funding costs from replacing debt with equity.
  - Key uncertainty: the degree to which MM holds (the MM offset) determines ultimate steady-state cost to lending rates.
  - Three important qualifications before extrapolating low observed costs to policy-imposed higher capital:
    - Most empirical estimates rely on small endogenous variations in capital; exogenous regulatory increases could have larger effects.
    - Costs of bank capital may increase nonlinearly with the level of capital, so extrapolation to substantially higher levels warrants caution.
    - Hard-to-measure general equilibrium effects could matter materially, including:
      - The role of bank debt as a safe asset for the economy.
      - Potential migration of lending to the shadow banking sector.

### Summary policy-relevant points
- Empirical central ranges:
  - Capital ratios of 15–23 percent appear to substantially reduce the likelihood of banking crises and public recapitalizations in advanced economies.
  - A 15–17 percent risk-weighted capital ratio helps avoid public recapitalizations in about 75 percent of banking crises.
  - A 1 percentage point increase in capital requirements has been estimated to raise lending rates by between 2 basis points and 20 basis points in the literature; U.S. loan-level evidence indicates 2.5 basis points for a 1 percentage point Tier 1 increase.
- Policy considerations:
  - Advanced economies: marginal benefits of bank capital decline substantially after the 15–23 percent range; weigh these benefits against transition and steady-state costs.
  - Emerging markets/non-OECD: higher observed NPLs suggest higher capital needs, but smaller banking-system sizes relative to GDP and institutional improvements may reduce required capital differentials.
  - Complement higher capital requirements with institutional reforms (regulation, supervision, resolution) to reduce NPL formation and improve comparability and effectiveness.

*Sources: Laeven and Valencia 2013; IMF Financial Soundness Indicators; Bankscope; Fratianni and Marchionne 2013; SNL Financial; IMF staff calculations.*

### Section IV.C.

### _sdn1604 - Section IV.C.

### Insights from General Equilibrium Models
- The general equilibrium literature nests opposing channels: capital reduces bank risk taking and crisis probability (lowering consumption volatility) but can crowd out deposits and reduce credit and output due to liquidity preferences of cash investors.
- Calibrated models typically suggest optimal bank capital can range from 8 to 20 percent (Van den Heuvel 2008; Nguyen 2013; Begneau 2014; Martinez-Miera and Suarez 2014; Mendicino and others 2015).
- The calibrated welfare effects of varying bank capital within that range—measured via higher lending rates and lost consumption—are relatively small.
- The 15–23 percent bank capital ratio discussed in Sections III.A and B is near the top of the range of optimal levels identified in general equilibrium models but is consistent with their predictions.

### Sources of Variation in Model Estimates
- The relatively wide range (8 to 20 percent) reflects many degrees of freedom in model design and parameter calibration.
- Model features highlighted in the literature that affect optimal capital estimates:
  - Substitution between bank funding and direct lending (effects of migration of activities to shadow banking).
  - An increase in bank capital may reduce the quantity but increase the quality of bank liquidity provision (making deposits safer and cheaper).
  - “Last bank standing” effects: well-capitalized banks can survive systemic shocks and earn scarcity rents.

### Key Implications
- The trade-off captured by these models supports the view that a 15–23 percent loss-absorption capacity is plausible and consistent with general equilibrium analysis.
- Given model uncertainty and parameter sensitivity, the 15–23 percent figure should be seen as consistent with, but not uniquely implied by, general equilibrium calibrated results.

*Italic: Source: _sdn1604 - Section IV.C.*

### 33. Barth, J. R., C. Lin, Y. Ma, J. Seade,  and F. M. Song. 2013. “Do Bank Regulation,

### 33. Barth, J. R., C. Lin, Y. Ma, J. Seade, and F. M. Song. 2013. “Do Bank Regulation, Supervision and Monitoring Enhance or Impede Bank Efficiency?” Journal of Banking & Finance 37 (8): 2879–92.

### Major themes covered in the citations
- Bank regulation, supervision, monitoring, and their effects on bank efficiency and risk-taking.
- Capital requirements, liquidity requirements, and their macroeconomic and bank-level impacts.
- Interaction of competition, market structure, and regulatory capital on bank behavior.
- Macroprudential policy tools, stress testing, contingency planning (bail-in, living wills), and resolution frameworks.
- Empirical evaluations of bank lending, capital ratios, and crisis-era behaviors (liquidity hoarding, credit flows).
- Theoretical models of bank runs, deposit insurance, demandable debt, and optimal capital structure.
- Shadow banking, wholesale funding risks, and systemic risk/GSIB loss-absorbing capacity.

### Key policy-relevant subjects reflected across the citations
- Assessments of the macroeconomic impact of transitioning to stronger capital and liquidity requirements (BIS, Macroeconomic Assessment Group).
- Calibration and adequacy of loss-absorbing capacity for Global Systemically Important Banks (FSB, BIS).
- Empirical estimates of the effect of capital requirements on bank lending, credit supply, and cost of capital (Bank of England, Bridges et al. 2015; Elliott 2009; Noss and Toffano 2014).
- Design and evaluation of stress-testing methodologies and supervisory stress tests (FRB Dodd-Frank Act Stress Test 2015).
- Policy tools for mitigating financial system vulnerabilities: macroprudential instruments, contingency capital, bail-in frameworks, and living wills.
- Trade-offs between financial stability (higher capital buffers) and potential short-run impacts on lending and economic activity.

*Source: Bibliographic and reference list excerpt from "BENEFITS AND COSTS OF BANK CAPITAL", International Monetary Fund.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/sdn/2016/_sdn1604.pdf_