## wpiea2026033

## Source details

**Canonical URL:** [wpiea2026033](https://www.imf.org/-/media/files/publications/wp/2026/english/wpiea2026033.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2026/english/wpiea2026033.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2026/english/wpiea2026033.pdf.json)

---

### 4.1 Baseline Analysis — Calibration and key parameters
- Model parameters broadly calibrated in line with Covi et al. (2021). Key calibrated parameters and definitions:
  - Loss-given-default rate (λijk): calibrated as the ratio of net exposure to gross exposure. Net exposure = exposure after exemptions and credit risk mitigation instruments; gross exposure = exposure after exemptions but before credit risk mitigation instruments (templates C.27–C.28).
  - Funding shortfall rate (ρik): assumed to be 100 percent, implying that all interbank funding matures within the 30-day horizon.
  - Net liquidity position (γi): defined as the bank’s high-quality liquid assets (HQLAi) (template C.72.00.a.10) in excess of its net liquidity outflows (template C.76.00.a.20).
  - Fire-saleable asset pool (θi): total stock of unencumbered, non-central bank eligible marketable assets held by the bank (template F.32.01).
  - Fire sale discount rate (δi): assumed to be 50 percent, corresponding to the highest haircut applied to central bank eligible assets.
  - Capital surplus (ci,t − ci,min): defined as the excess of Common Equity Tier 1 (CET1) capital over the minimum capital requirement (hurdle rate). Minimum capital requirement = 4.5 percent minimum Pillar 1 requirement + bank-specific Pillar 2 Requirement (P2R) set under SREP.
- Note on insolvency trigger: the trigger in the euro area corresponds to breaching overall capital requirements (4.5 percent CET1 plus P2R), although banks could also be resolved for cash-flow insolvency or confidence-driven deposit outflows even if regulatory capital exceeds requirements.

### 4.1 Baseline Analysis — Baseline statistics (Table 1, Panel A; Billion euros; 2024Q2)
- Loss-given-default rate (λijk):
  - Gross Exposure: Mean 0.7; SD 0.9; p25 0.1; p50 0.3; p75 0.9
  - Credit Risk Mitigation: Mean 0.1; SD 0.4; p25 0.0; p50 0.0; p75 0.1
  - Net Exposures: Mean 0.5; SD 0.8; p25 0.1; p50 0.2; p75 0.6
- Net liquidity position (γi):
  - High-Quality Liquid Assets: Mean 61.6; SD 81.0; p25 10.4; p50 26.3; p75 91.8
  - Net Liquidity Outflows: Mean 38.9; SD 56.8; p25 4.5; p50 11.6; p75 56.6
- Fire-saleable asset pool (θi):
  - Unencumbered non-HQLA assets: Mean 215; SD 346; p25 39; p50 66; p75 204
- Capital surplus (ci,t − ci,min):
  - CET1: Mean 16.2; SD 20.6; p25 3.2; p50 7.4; p75 19.0
  - CET1 Minimum: Mean 6.1; SD 8.3; p25 1.1; p50 2.6; p75 7.0

### 4.1 Baseline Analysis — Simulation methodology and systemic indices
- Interbank simulation: iterative series of hypothetical defaults; two key outputs:
  - (i) number of additional defaults caused by the initial trigger event;
  - (ii) losses incurred by each entity in the network during each simulation round.
- Indices used to rank banks (following Covi et al. (2021)):
  - Contagion Index (CI):
    - CI_i = 100· (Σ_{j≠i} L_ji) / (Σ_{j≠i} k_ji)
  - Vulnerability Index (VI):
    - VI_i = 100· (Σ_{j≠i} L_ij) / (Σ_{j≠i} k_i)
  - Interpretation:
    - High CI: failure of bank i generates disproportionately large losses system‑wide (excluding i).
    - High VI: bank i experiences large average losses across simulations, expressed as a percentage of its own capital.
  - Joint distribution used to identify banks that are both systemically important and vulnerable.

### 4.1 Baseline Analysis — Baseline simulation results (72 hypothetical default scenarios)
- Aggregate outcomes:
  - Among the 72 hypothetical default scenarios analyzed, the ten most impactful events (Table 2) generate, on average, capital losses equivalent to 1.3 percent of the aggregate capital in the euro area banking system.
  - Transmission occurs exclusively through credit losses.
  - No cascading defaults occur under the baseline parametrization.
  - Comparison: Covi et al. (2021) reported average system-wide losses of 2.5 percent based on 2017Q3 data.
- Table 2 (top ranks; Contagion and Vulnerability scores shown):
  - Rank 1: Business Model G-SIB — Contagion Index 2.5; Vulnerability Index 0.2; Total Defaults 0
  - Rank 2: G-SIB — Contagion Index 2.2; Vulnerability Index 0.4; Total Defaults 0
  - Rank 3: G-SIB — Contagion Index 1.8; Vulnerability Index 0.2; Total Defaults 0
  - Rank 4: G-SIB — Contagion Index 1.8; Vulnerability Index 0.3; Total Defaults 0
  - Rank 5: G-SIB — Contagion Index 1.2; Vulnerability Index 0.0; Total Defaults 0
  - Rank 6: Universal — Contagion Index 1.0; Vulnerability Index 0.4; Total Defaults 0
  - Rank 7: Universal — Contagion Index 0.8; Vulnerability Index 0.6; Total Defaults 0
  - Rank 8: Universal — Contagion Index 0.6; Vulnerability Index 0.9; Total Defaults 0
  - Rank 9: Lender — Contagion Index 0.6; Vulnerability Index 1.0; Total Defaults 0
  - Rank 10: Universal — Contagion Index 0.5; Vulnerability Index 0.8; Total Defaults 0
- Key qualitative findings:
  - Two institutions identified with characteristics that may amplify systemic spillovers under stress.
  - Six banks exhibit elevated contagion scores (failure could generate outsized system-wide losses even if those banks are relatively resilient).
  - Three institutions rank high on the vulnerability index (particularly exposed to counterparties despite limited outward spillover potential).
  - Heterogeneity across bank types and jurisdictions:
    - G-SIBs: pronounced capacity to generate system-wide losses (high CI) but relatively low inward vulnerability under current conditions.
    - Universal and lender-type banks: tend to display higher vulnerability scores.
    - Jurisdictional concentration matters: jurisdictions with more G-SIBs account for institutions classified as highly contagious; jurisdictions with more universal or lender-type banks account for more highly vulnerable institutions.
- Policy relevance:
  - Results can inform identification and calibration of the capital buffer for systemically important banks (linked to the Contagion Index) and systemic risk buffers (linked to the Vulnerability Index).

### 5.3 Market Risk Scenario — Scenario description
- Combined shock: market risk compounded with NBFI-associated risks leads to materially larger contagion through interbank exposures.
- Exogenous defaults in the scenario: seven bank default events—specifically those ranked 1 through 8, excluding event 4—are imposed.
- Transmission channel: contagion is transmitted exclusively through the credit channel; losses arise solely from materialization of credit risk despite liquidity risks associated with NBFIs and market volatility.

### 5.3 Market Risk Scenario — Quantitative findings and system-wide outcomes
- Total cascading defaults resulting from the combined shock: eleven cascading defaults within the network.
- Average system-wide capital losses under the combined market-risk and NBFI stress scenario: 1.9 percent.
- Incremental effect versus NBFI-only scenario: an additional 1.3 percentage points of system-wide capital losses.
- Most severe event under the combined shock:
  - Additional defaults in that event: three additional defaults.
  - Capital losses in that event: 4.4 percent of the euro area banking system’s aggregate capital.
- Change in frequency of cascading defaults: two additional cascading defaults under the combined scenario relative to NBFI-only.

### 5.3 Market Risk Scenario — Contagion and vulnerability by ranked banks (Table 4)
- Rank 1: Business Model = G-SIB, Contagion Index = 4.40, Vulnerability Index = 0.23
- Rank 2: Business Model = G-SIB, Contagion Index = 3.40, Vulnerability Index = 0.52
- Rank 3: Business Model = G-SIB, Contagion Index = 2.30, Vulnerability Index = 0.21
- Rank 4: Business Model = G-SIB, Contagion Index = 2.20, Vulnerability Index = 0.40
- Rank 5: Business Model = G-SIB, Contagion Index = 1.70, Vulnerability Index = 0.01
- Rank 6: Business Model = Universal, Contagion Index = 1.30, Vulnerability Index = 0.81
- Rank 7: Business Model = Universal, Contagion Index = 1.31, Vulnerability Index = 1.31
- Rank 8: Business Model = Universal, Contagion Index = 1.11, Vulnerability Index = 1.42
- Rank 9: Business Model = Universal, Contagion Index = 0.70, Vulnerability Index = 0.00
- Rank 10: Business Model = Lender, Contagion Index = 0.63, Vulnerability Index = 3.30

### 5.3 Market Risk Scenario — Bank business-model breakdown and simulation distribution
- Disaggregated results by bank business model remain broadly consistent with baseline findings, reinforcing robustness of the systemic risk map across stress scenarios.
- Distributional shift: the distribution of contagion and vulnerability losses across simulations shifts markedly to the right—more so than in the NBFI scenario—highlighting amplification from significant exposures to NBFIs combined with the market shock.
- Increased cascading defaults under the combined shock underscore the importance of incorporating multi-dimensional stress scenarios into macroprudential surveillance.

### 5.3 Market Risk Scenario — Policy implications and recommendations
- Enhanced monitoring of bank–NBFI exposures is needed.
- Improved integration of market risk into systemic risk frameworks is required.
- Continued vigilance regarding the structural liquidity position of euro area banks is necessary.
- Microprudential policies focused solely on individual institution resilience may underestimate systemic contributions; macroprudential policies should encompass the entire financial ecosystem, including NBFIs and market vulnerabilities.
- Consideration for identifying systemically important institutions (O-SIIs) may incorporate contagion and vulnerability indices as additional quantitative factors for buffer calibration.
- Authorities should integrate granular bilateral exposures and contagion models into stress testing frameworks and foster cross-sectoral cooperation and data sharing among banking, securities, and NBFI regulators.

### 5.3 Market Risk Scenario — Caveats and suggested extensions
- Omitted transmission channels may understate contagion risk: banks’ funding exposures to NBFIs, liquidity-related borrowing costs, and asset valuation effects from market volatility are not included.
- The conclusion that contagion stems solely from insolvency rather than illiquidity should be interpreted with caution; no defaults are triggered by liquidity shortfalls across baseline and risk scenarios.
- Current ample excess liquidity in the euro area—largely attributable to the Eurosystem’s accommodative policy measures during and after the COVID-19 pandemic—may explain the lack of liquidity-triggered defaults; balance sheet normalization will erode this buffer and increase liquidity vulnerability.
- Further analysis could:
  - Incorporate counterparty credit risk from derivatives and securities financing transactions, including credit valuation adjustments.
  - Expand the network to include banks domiciled in foreign jurisdictions and include NBFIs, requiring cross-jurisdictional and cross-sectoral data sharing and joint risk assessments.

*Source: ECB and IMF staff calculations (from the provided content).*

### 4.1  Baseline Analysis

### 4.1  Baseline Analysis

### Calibration and key parameters
- Model parameters broadly calibrated in line with Covi et al. (2021). Key calibrated parameters and definitions:
  - Loss-given-default rate (λijk): calibrated as the ratio of net exposure to gross exposure. Net exposure = exposure after exemptions and credit risk mitigation instruments; gross exposure = exposure after exemptions but before credit risk mitigation instruments (templates C.27–C.28).
  - Funding shortfall rate (ρik): assumed to be 100 percent, implying that all interbank funding matures within the 30-day horizon.
  - Net liquidity position (γi): defined as the bank’s high-quality liquid assets (HQLAi) (template C.72.00.a.10) in excess of its net liquidity outflows (template C.76.00.a.20).
  - Fire-saleable asset pool (θi): total stock of unencumbered, non-central bank eligible marketable assets held by the bank (template F.32.01).
  - Fire sale discount rate (δi): assumed to be 50 percent, corresponding to the highest haircut applied to central bank eligible assets.
  - Capital surplus (ci,t − ci,min): defined as the excess of Common Equity Tier 1 (CET1) capital over the minimum capital requirement (hurdle rate). Minimum capital requirement = 4.5 percent minimum Pillar 1 requirement + bank-specific Pillar 2 Requirement (P2R) set under SREP.
- Note on insolvency trigger: the trigger in the euro area corresponds to breaching overall capital requirements (4.5 percent CET1 plus P2R), although banks could also be resolved for cash-flow insolvency or confidence-driven deposit outflows even if regulatory capital exceeds requirements.

### Baseline statistics (Table 1, Panel A; Billion euros; 2024Q2)
- Loss-given-default rate (λijk):
  - Gross Exposure: Mean 0.7; SD 0.9; p25 0.1; p50 0.3; p75 0.9
  - Credit Risk Mitigation: Mean 0.1; SD 0.4; p25 0.0; p50 0.0; p75 0.1
  - Net Exposures: Mean 0.5; SD 0.8; p25 0.1; p50 0.2; p75 0.6
- Net liquidity position (γi):
  - High-Quality Liquid Assets: Mean 61.6; SD 81.0; p25 10.4; p50 26.3; p75 91.8
  - Net Liquidity Outflows: Mean 38.9; SD 56.8; p25 4.5; p50 11.6; p75 56.6
- Fire-saleable asset pool (θi):
  - Unencumbered non-HQLA assets: Mean 215; SD 346; p25 39; p50 66; p75 204
- Capital surplus (ci,t − ci,min):
  - CET1: Mean 16.2; SD 20.6; p25 3.2; p50 7.4; p75 19.0
  - CET1 Minimum: Mean 6.1; SD 8.3; p25 1.1; p50 2.6; p75 7.0

### Simulation methodology and systemic indices
- Interbank simulation: iterative series of hypothetical defaults; two key outputs:
  - (i) number of additional defaults caused by the initial trigger event;
  - (ii) losses incurred by each entity in the network during each simulation round.
- Indices used to rank banks (following Covi et al. (2021)):
  - Contagion Index (CI):
    - CI_i = 100· (Σ_{j≠i} L_ji) / (Σ_{j≠i} k_ji)
  - Vulnerability Index (VI):
    - VI_i = 100· (Σ_{j≠i} L_ij) / (Σ_{j≠i} k_i)
  - Interpretation:
    - High CI: failure of bank i generates disproportionately large losses system‑wide (excluding i).
    - High VI: bank i experiences large average losses across simulations, expressed as a percentage of its own capital.
  - Joint distribution used to identify banks that are both systemically important and vulnerable.

### Baseline simulation results (72 hypothetical default scenarios)
- Aggregate outcomes:
  - Among the 72 hypothetical default scenarios analyzed, the ten most impactful events (Table 2) generate, on average, capital losses equivalent to 1.3 percent of the aggregate capital in the euro area banking system.
  - Transmission occurs exclusively through credit losses.
  - No cascading defaults occur under the baseline parametrization.
  - Comparison: Covi et al. (2021) reported average system-wide losses of 2.5 percent based on 2017Q3 data.
- Table 2 (top ranks; Contagion and Vulnerability scores shown):
  - Rank 1: Business Model G-SIB — Contagion Index 2.5; Vulnerability Index 0.2; Total Defaults 0
  - Rank 2: G-SIB — Contagion Index 2.2; Vulnerability Index 0.4; Total Defaults 0
  - Rank 3: G-SIB — Contagion Index 1.8; Vulnerability Index 0.2; Total Defaults 0
  - Rank 4: G-SIB — Contagion Index 1.8; Vulnerability Index 0.3; Total Defaults 0
  - Rank 5: G-SIB — Contagion Index 1.2; Vulnerability Index 0.0; Total Defaults 0
  - Rank 6: Universal — Contagion Index 1.0; Vulnerability Index 0.4; Total Defaults 0
  - Rank 7: Universal — Contagion Index 0.8; Vulnerability Index 0.6; Total Defaults 0
  - Rank 8: Universal — Contagion Index 0.6; Vulnerability Index 0.9; Total Defaults 0
  - Rank 9: Lender — Contagion Index 0.6; Vulnerability Index 1.0; Total Defaults 0
  - Rank 10: Universal — Contagion Index 0.5; Vulnerability Index 0.8; Total Defaults 0
- Key qualitative findings:
  - Two institutions identified with characteristics that may amplify systemic spillovers under stress.
  - Six banks exhibit elevated contagion scores (failure could generate outsized system-wide losses even if those banks are relatively resilient).
  - Three institutions rank high on the vulnerability index (particularly exposed to counterparties despite limited outward spillover potential).
  - Heterogeneity across bank types and jurisdictions:
    - G-SIBs: pronounced capacity to generate system-wide losses (high CI) but relatively low inward vulnerability under current conditions.
    - Universal and lender-type banks: tend to display higher vulnerability scores.
    - Jurisdictional concentration matters: jurisdictions with more G-SIBs account for institutions classified as highly contagious; jurisdictions with more universal or lender-type banks account for more highly vulnerable institutions.
- Policy relevance:
  - Results can inform identification and calibration of the capital buffer for systemically important banks (linked to the Contagion Index) and systemic risk buffers (linked to the Vulnerability Index).

*Source: ECB and IMF staff calculations (from the provided content).*

### 5.3  Market Risk Scenario

### 5.3  Market Risk Scenario

### Scenario description
- Combined shock: market risk compounded with NBFI-associated risks leads to materially larger contagion through interbank exposures.
- Exogenous defaults in the scenario: seven bank default events—specifically those ranked 1 through 8, excluding event 4—are imposed.
- Transmission channel: contagion is transmitted exclusively through the credit channel; losses arise solely from materialization of credit risk despite liquidity risks associated with NBFIs and market volatility.

### Quantitative findings and system-wide outcomes
- Total cascading defaults resulting from the combined shock: eleven cascading defaults within the network.
- Average system-wide capital losses under the combined market-risk and NBFI stress scenario: 1.9 percent.
- Incremental effect versus NBFI-only scenario: an additional 1.3 percentage points of system-wide capital losses.
- Most severe event under the combined shock:
  - Additional defaults in that event: three additional defaults.
  - Capital losses in that event: 4.4 percent of the euro area banking system’s aggregate capital.
- Change in frequency of cascading defaults: two additional cascading defaults under the combined scenario relative to NBFI-only.

### Contagion and vulnerability by ranked banks (Table 4)
- Rank 1: Business Model = G-SIB, Contagion Index = 4.40, Vulnerability Index = 0.23
- Rank 2: Business Model = G-SIB, Contagion Index = 3.40, Vulnerability Index = 0.52
- Rank 3: Business Model = G-SIB, Contagion Index = 2.30, Vulnerability Index = 0.21
- Rank 4: Business Model = G-SIB, Contagion Index = 2.20, Vulnerability Index = 0.40
- Rank 5: Business Model = G-SIB, Contagion Index = 1.70, Vulnerability Index = 0.01
- Rank 6: Business Model = Universal, Contagion Index = 1.30, Vulnerability Index = 0.81
- Rank 7: Business Model = Universal, Contagion Index = 1.31, Vulnerability Index = 1.31
- Rank 8: Business Model = Universal, Contagion Index = 1.11, Vulnerability Index = 1.42
- Rank 9: Business Model = Universal, Contagion Index = 0.70, Vulnerability Index = 0.00
- Rank 10: Business Model = Lender, Contagion Index = 0.63, Vulnerability Index = 3.30

### Bank business-model breakdown and simulation distribution
- Disaggregated results by bank business model remain broadly consistent with baseline findings, reinforcing robustness of the systemic risk map across stress scenarios.
- Distributional shift: the distribution of contagion and vulnerability losses across simulations shifts markedly to the right—more so than in the NBFI scenario—highlighting amplification from significant exposures to NBFIs combined with the market shock.
- Increased cascading defaults under the combined shock underscore the importance of incorporating multi-dimensional stress scenarios into macroprudential surveillance.

### Policy implications and recommendations
- Enhanced monitoring of bank–NBFI exposures is needed.
- Improved integration of market risk into systemic risk frameworks is required.
- Continued vigilance regarding the structural liquidity position of euro area banks is necessary.
- Microprudential policies focused solely on individual institution resilience may underestimate systemic contributions; macroprudential policies should encompass the entire financial ecosystem, including NBFIs and market vulnerabilities.
- Consideration for identifying systemically important institutions (O-SIIs) may incorporate contagion and vulnerability indices as additional quantitative factors for buffer calibration.
- Authorities should integrate granular bilateral exposures and contagion models into stress testing frameworks and foster cross-sectoral cooperation and data sharing among banking, securities, and NBFI regulators.

### Caveats and suggested extensions
- Omitted transmission channels may understate contagion risk: banks’ funding exposures to NBFIs, liquidity-related borrowing costs, and asset valuation effects from market volatility are not included.
- The conclusion that contagion stems solely from insolvency rather than illiquidity should be interpreted with caution; no defaults are triggered by liquidity shortfalls across baseline and risk scenarios.
- Current ample excess liquidity in the euro area—largely attributable to the Eurosystem’s accommodative policy measures during and after the COVID-19 pandemic—may explain the lack of liquidity-triggered defaults; balance sheet normalization will erode this buffer and increase liquidity vulnerability.
- Further analysis could:
  - Incorporate counterparty credit risk from derivatives and securities financing transactions, including credit valuation adjustments.
  - Expand the network to include banks domiciled in foreign jurisdictions and include NBFIs, requiring cross-jurisdictional and cross-sectoral data sharing and joint risk assessments.

*Source: ECB and IMF staff calculations as presented in the chapter "5.3  Market Risk Scenario" of the provided document.*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2026/english/wpiea2026033.pdf_