## EXECUTIVE SUMMARY

## Source details

**Canonical URL:** [EXECUTIVE SUMMARY](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16315.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16315.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16315.pdf.json)

---

### Introduction and context
- FSAP stress testing took place amid a rapid Irish economic rebound and banking profitability return in 2014, but with a very large stock of non-performing loans.
- Banking sector structure and size:
  - Total banking sector assets amount to 294 percent of GDP.
  - Top five banks control 61 percent of system assets.
  - One large privately-owned bank controls 24 percent of total assets.
  - Two state-owned banks make up 27 percent of the banking system’s assets.
  - Subsidiaries of two foreign banks control 10 percent of system assets.
- Stress testing horizon and data:
  - Three-year horizon (2016–2018).
  - Data available through June 2015, with capital figures updated based on December 2015 data.

### Design of the stress testing framework
- Coverage and approaches:
  - Tests covered solvency, liquidity, and contagion risks using top-down (TD) and bottom-up (BU) approaches.
  - FSAP TD and BU focused on five systemically important institutions representing 61 percent of banking sector assets; Central Bank of Ireland TD included six additional non-systemically important institutions.
- Macroeconomic shocks and domestic risks incorporated:
  - External: protracted weak growth in advanced economies (especially the euro area); surge in global financial market volatility; planned U.K. exit from the EU.
  - Domestic: domestic confidence shock, house price decline reverting towards crisis-era levels, commercial real estate correction risk.
- Methodologies:
  - Satellite models and methodologies from Central Bank of Ireland, ECB, and Fund staff for profitability and capitalization.
  - TD liquidity tests used maturity ladder analysis and supervisory information, aggregated and by currency.
  - Contagion tests covered interbank exposures among the three largest domestic banks, domestic interlinkages, and cross-border exposures using BIS International Banking Statistics and market data.
- Caveats:
  - Scenarios are extreme but plausible and not forecasts; adverse scenarios reflect events with very low ex ante probability (could possibly materialize once every 20–40 years).
  - Satellite models estimated on historical data and subject to estimation uncertainty.

### Macroeconomic scenarios (selected projections)
- Severe (Ireland-specific) scenario (Real GDP growth):
  - 2016: 1.1
  - 2017: 0.3
  - 2018: 0.1
  - Cumulative decline of GDP relative to baseline over three years: about 9.5 percentage points.
- Moderate (euro area wide) scenario (Real GDP growth):
  - 2016: 1.3
  - 2017: 1.2
  - 2018: 3.1
  - Cumulative decline of GDP relative to baseline over three years: 6.5 percentage points.
- Selected macro variables (Est. 2015; Baseline; Severe; Moderate):
  - Real GDP growth: Est. 2015: 5.8; Baseline: 4.7, 4.0, 3.0; Severe: 1.1, 0.3, 0.1; Moderate: 1.3, 1.2, 3.1.
  - Unemployment rate: Est. 2015: 9.5; Baseline: 8.5, 7.8, 7.2; Severe: 9.6, 10.5, 11.2; Moderate: 9.5, 9.4, 8.4.
  - Residential real estate price annual change: Est. 2015: 6.5; Baseline: 7.5, 6.7, 5.1; Severe: -5.0, -5.0, 0.0; Moderate: 1.2, 0.8, 2.6.
  - Commercial real estate price annual change: Est. 2015: 21.8; Baseline: 10.3, 4.2, 1.9; Severe: -4.6, -16.1, -11.2; Moderate: 2.9, -5.9, -4.7.
  - Nominal annual interest rate (10-year government bond rate): Est. 2015: 1.3; Baseline: 1.7, 2.3, 3.3; Severe: 3.2, 5.3, 7.9; Moderate: 3.6, 4.6, 4.8.
  - 10-year sovereign spread (bp vs German): Est. 2015: 75; Baseline: 102, 142, 226; Severe: 255, 447, 683; Moderate: 251, 328, 339.

### SOLVENCY STRESS TESTS

### Framework and regulatory hurdle rates
- Applied framework: Basel III, EU and national frameworks as defined by CRD IV, national law and Central Bank regulation.
- Hurdle rates (fully-loaded Basel III definitions):
  - Total capital adequacy: 8 percent.
  - Tier 1 capital: 6 percent.
  - Common Equity Tier 1 capital (CET1): 4.5 percent.
  - Capital conservation buffer: 2.5 percent (capital shortfall calculations did not take this buffer into account).
  - Additional hurdle rate for leverage ratio: 3 percent from 2018 onwards.
- RWAs: For Basel II IRB banks, risk weights projected using Basel II formula depending on PD, LGD and asset correlation; RWAs exogenously constrained not to decrease over the stress period.

### PD modeling and credit-loss projections
- PD segmentation: eight segments including Irish household mortgages, U.K. household mortgages, Irish NFCs, U.K. NFCs, Irish financials, U.K. financials, Irish construction and CRE, sovereign exposures.
- IMF Top-Down model key estimated effects (logit transform of PD):
  - Irish mortgage loans:
    - +1 percentage point unemployment → PD rises by 2 percentage points.
    - +1 percentage point real interest rate → PD rises by 0.5 percentage points.
    - +1 percentage point bank market share → PD decreases by 0.3 percentage points one year later.
  - NFC loans:
    - -1 percentage point YoY real GDP growth → PD rises by 2 percentage points.
    - Euro depreciation by 0.1 → PD rises by 0.5 percentage points.
- System-wide PD projections (IMF model), from 1.9 percent in 2015:
  - Baseline: PDs remain flat at 1.9 percent.
  - Moderate stress: peak PD of 2.5 percent in 2018.
  - Severe stress: PD rises to 4.3 percent in 2018.
- Credit losses (loan book):
  - Severe stress: credit losses amount to EUR 7.6 bn, equivalent to 2.2 percent of total banking system assets.
  - Baseline: flow of new provisions EUR 2.2 bn, equivalent to 0.6 percent of total banking system assets.

### Solvency test results and drivers (fully-loaded Basel III)
- Aggregate CET1 ratio impact:
  - Severe scenario: aggregate CET1 would drop by 8 percentage points compared with the starting point and be 7.2 percentage points below the baseline in 2018.
- Main contributors to CET1 decline (in percentage points of RWAs):
  - Funding costs: -4.8 percentage points.
  - Loss provisions (credit risk): -4.4 percentage points.
  - Other comprehensive income linked to AFS valuation losses: -2 percentage points over the whole stress horizon.
  - Basel III capital adjustments: -3.7 percentage points.
- Bank-level outcomes (severe scenario):
  - Two banks would be undercapitalized with regard to total CAR and Tier 1 hurdle rates of 8 percent and 6 percent, respectively (capital shortfalls of 0.2 and 0.1 percent of GDP).
  - Four banks would not meet CET1 level of 7 percent (combined minimum CET1 and capital conservation buffer).
  - Three banks would have leverage ratio below hurdle rate of 3 percent in 2018 (leverage shortfall equivalent to 0.1 percent of GDP).
  - System-level leverage ratio (Tier 1/total assets): 2015 initial level: 9.6 percent; Baseline 2018: 7.4 percent; Moderate stress 2018: 6.5 percent; Severe stress 2018: 5.2 percent.
- Transitional arrangements:
  - Results similar under EU phase-in rules, with reduced capital shortfall (aggregate shortfall limited to 0.1 percent of GDP under transitional).

### Comparison TD vs BU and vs Central Bank of Ireland TD
- FSAP Top-Down results more severe than banks’ Bottom-Up submissions:
  - Bottom-Up: no bank undercapitalized under adverse scenario; only one could not meet conservation buffer.
- Central Bank of Ireland TD applied more conservative adjustments (credit benchmarks, funding shock, interest income, RWA projection), lowering CET1 by 1 to 3 percentage points of RWAs and producing one undercapitalized bank under the adverse scenario.
- Key methodological differences:
  - Net interest income on NPLs: FSAP assumed no accrued interest; CBI allowed discount unwind per EBA.
  - AFS shock: FSAP point-in-time shock; CBI dynamic approach.
  - Pension fund balances: allowed by CBI, absent in IMF methodology.
  - Coverage of LSIs: included in CBI TD; not covered by IMF TD.

### SENSITIVITY, MARKET AND CONCENTRATION RISKS

### Selected sensitivity test outcomes (static as of June 2015)
- Unemployment shock (to 12.7 percent):
  - PD increases by 1.6 percentage points (1.9 → 3.5 percent).
  - Credit loss provisions equal to EUR 1.5 bn.
  - Decline in total CAR by 0.9 percentage points.
  - No undercapitalization among the five banks.
- Domestic sovereign securities price shock (500 bp increase):
  - CAR decline of 2 percentage points taken in isolation.
  - Initial capitalization on transitional basis sufficient to avoid undercapitalization.
- Interest rate shock (500 bp increase) — net interest income:
  - System net interest income loss EUR 403 mn.
  - Aggregate Tier 1 ratio falls by 0.2 percentage points.
  - No undercapitalization.
- Interest rate shock (500 bp) — credit quality:
  - PD increases by 1.4 percentage points (1.9 → 3.3 percent).
  - Credit loss provisions EUR 1.5 bn; total CAR decline 0.9 percentage points.
  - No undercapitalization.
- Real estate price decline (30 percent):
  - System LGD rises by 16 percentage points (from 16½ percent to 32½ percent).
  - Aggregate loss provisions EUR 1.7 bn.
  - CAR loss equivalent to 1 percentage point.
- Concentration risk — failure of large corporate exposures:
  - Average net largest exposure = 4.9 percent of total regulatory capital.
  - Default of largest exposure for each bank → no undercapitalization.
  - Simultaneous default of five to ten largest exposures → undercapitalization in one bank.
  - Default of five largest exposures → Tier 1 shortfall EUR 167 mn (0.1 percent of GDP).
  - Default of ten largest exposures → Tier 1 shortfall EUR 639 mn (0.3 percent of GDP).

### Market risks assessed
- Repricing gap for maturities below one year (5 banks, June 2015): EUR -8.1 bn.
- Severe scenario: aggregated loss due to change in interest rates EUR 82 mn, no material effect on CAR.
- Valuation losses on domestic sovereign securities in AFS and HFT portfolios under severe scenario: EUR 2.7 bn, contributing 1.6 percentage points to CAR decline.
  - Average AFS exposure to own sovereign: 4.3 percent of total assets (individual bank ratios between 2 percent and 8.5 percent).
  - Average duration of sovereign AFS portfolio: 5 years.
  - Large increase in Irish 10-year government bond rate under adverse scenario: 6.6 percentage points.
- Net open FX position for the banking system (June 2015): EUR 330 mn, equivalent to 0.1 percent of assets and 0.4 percent of capital.
  - Severe scenario aggregated loss on net FX positions: EUR 2 mn.

### LIQUIDITY STRESS TESTS

### Funding structure and regulatory timelines (June 2015)
- Funding structure (excluding capital and contingent credit lines):
  - 77 percent from customer deposits.
  - 19 percent wholesale funding (Debt Capital Markets 11 percent, Interbank deposits and Repo 8 percent).
  - 4 percent from central banks.
- LCR timeline and national transposition:
  - LCR above 60 percent from October 2015 and 100 percent from 2018.
  - National transposition includes a larger range of HQLA with high haircuts and accelerated phase-in to 100 percent by 2018.
- NSFR: requirement for a stable funding profile at a one-year horizon from 2018.

### LCR-based stress test scenarios and results
- Aggregate LCR at June 2015 (European Commission Regulation parameters): 124 percent; every bank passed 60 percent hurdle; four of five banks above 100 percent hurdle.
- Adverse scenario 1 (severe retail deposit shock):
  - Aggregate LCR would fall to 87 percent.
  - Liquidity shortfall EUR 7.1 bn, equivalent to 3.3 percent of GDP.
- Adverse scenario 2 (dry-up of unsecured wholesale funding):
  - Aggregate LCR would fall to 87 percent.
  - Total liquidity shortfall EUR 7.4 bn, equivalent to 2.8 percent of these banks’ assets.
- Pound sterling liquidity stress:
  - Aggregate shortfall in GBP ranges between EUR 1.7 bn and EUR 6.0 bn, i.e., 0.8 percent and 2.8 percent of GDP across LCR scenarios.

### NSFR and outflow analysis
- NSFR (5 largest banks, June 2015):
  - Available stable funding: EUR 211.3 bn.
  - Required stable funding: EUR 196.4 bn.
  - Aggregate NSFR: 107.6 percent.
  - Some banks have NSFR < 100 percent → aggregate shortfall EUR 7.4 bn (3.4 percent of GDP).
- Outflow analysis (maturity buckets up to >10 years):
  - With expanded counterbalancing capacity (including securitized, stressed assets), banks could counterbalance net outflows for a wide range of maturity buckets and survive up to 10 years.
  - Using only standard HQLA, liquidity shortfalls arise for maturities beyond 3 months.

### Liquidity vulnerabilities identified
- Some banks exposed to liquidity risks under large deposit withdrawals or extreme wholesale funding drying up (scenarios more severe than Basel III LCR variants).
- Material exposure to funding risks in pound sterling for some banks.
- Additional counterbalancing capacity (expanded liquidity buffers) would allow banks to cope with net outflows in every maturity bucket in the tests.

### CONTAGION, INTERCONNECTEDNESS, AND CROSS-BORDER RISKS

### Domestic interbank contagion
- Bilateral interbank contagion risks among the three largest domestic banks are limited.
  - Interbank positions between the three largest banks are small relative to capitalization; gross exposures to the other two banks smaller than regulatory capital.
  - No single failure of a domestic bank would trigger another domestic bank’s failure via this three-bank market.
  - One bank’s vulnerability index is higher (still below 9 percent) but overall contagion via domestic interbank exposures is very limited.

### Cross-border exposures and simulation findings
- Composition of foreign claims (2015Q3):
  - Claims on non-resident private sectors: 76 percent of total foreign claims (CBS-URB).
  - Claims on foreign banks: 11 percent.
  - Claims on public sectors: 13 percent.
  - Local claims constituted 70 percent of total foreign claims at 2015Q3, mainly held by U.K. subsidiaries of Irish banks.
- BIS-based network simulations (17-country matrix, CBS-IBB, end-2014; baseline: ߣ = 1, ߩ = 0.5, ߜ = 0.5):
  - Credit shock only:
    - Hypothetical failure of entire U.K. banking system → total meltdown of global banking system after three contagion rounds.
    - Hypothetical failure of entire U.S. banking system → total meltdown after five contagion rounds.
    - Ireland directly hit by U.K. failure due to concentrated exposure; U.S. failure impacts Ireland via the U.K.
    - Ireland’s index of vulnerability: 17 percent (drops to 5 percent excluding U.K. and U.S.).
    - A banking crisis in Ireland would deprive only 7 percent of total regulatory capital buffers in the global banking network and would not generate domino effects.
  - Joint credit and funding shocks:
    - Germany and France become additional sources of negative spillovers toward Ireland.
    - Germany’s affected countries rise from one to sixteen when funding shocks included.
    - A French failure would not induce Irish failure but could deprive Ireland of 90 percent of its capital buffers via spillovers to Belgium and the Netherlands.
    - Ireland’s index of contagion stays below 10 percent under joint shocks.
- Market-price network (equity returns) analysis:
  - Irish financial institutions significantly affected by U.K. and French banks, and Spanish and Italian financial systems.
  - Individual contributions of these four sectors to Irish inward spillovers each exceed 5 percent; combined contribution 28½ percent.
  - Net inward connectedness of Irish financial system within the network: 7.7 percent (Ireland is a net recipient of spillovers).

### Historical and supervisory interpretation
- Deleveraging has reduced cross-border linkages since the pre-crisis period.
- Both foreign banks’ exposures to Ireland and Irish banks’ claims on non-residents have declined dramatically.
- Tight linkages with the U.K. financial system warrant ongoing attention due to large exposures relative to Irish banks’ balance sheets.

### KEY VULNERABILITIES, CONCLUSIONS, AND POLICY RECOMMENDATIONS

### Primary vulnerabilities
- Funding risk: funding costs increase sharply under adverse scenario.
- Credit risk: loan quality highly sensitive to unemployment and GDP growth; also sensitive to real interest rates.
- Concentration risk: high concentration of loan portfolios can exacerbate vulnerability; failure of the five largest exposures can cause undercapitalization of the affected bank.
- Liquidity risk: some banks vulnerable to large deposit withdrawals or a dry-up of unsecured wholesale funding; material GBP funding exposure for some banks.

### Main conclusions
- Several solvency and liquidity vulnerabilities remain but have declined markedly since the financial crisis and are manageable at the macro level.
- Under severe stress on a fully-loaded Basel III basis:
  - Two banks become undercapitalized with regard to total CAR and Tier 1 hurdle rates of 8 percent and 6 percent.
  - Three banks would have leverage ratio below 3 percent in 2018.
  - Four banks would not meet CET1 level of 7 percent (combined minimum CET1 and capital conservation buffer level).
- Contagion:
  - Direct domestic interbank contagion limited.
  - Cross-border spillovers significant from U.K. and French banks; interconnectedness has decreased since crisis.

### Policy recommendations (selected, with timing)
- Agencies: Central Bank of Ireland (Central Bank), European Central Bank (ECB).
- Near-term (NT — up to 2 years):
  - Further develop the Central Bank of Ireland Top-Down credit risk satellite model to increase ability to challenge banks’ submissions operationally.
  - Update Central Bank of Ireland loan-loss forecasting model calibration to ensure out-of-sample forecasting power remains adequate.
  - Carry out liquidity stress tests currency-by-currency.
  - Improve surveillance by closing data gaps on cross-border bilateral financial exposures (Central Bank and ECB).
  - Examine thoroughly banks’ liquidity positions in pound Sterling in a stressed environment (Central Bank and ECB).
- Medium-term (MT — 2–5 years):
  - Continue to review the link between sovereign and banking sector risks by reviewing the amount of banks’ AFS reserves for unrealized gains on exposures to central governments (Central Bank and ECB).
  - Continue to monitor banks’ plans to meet Basel III new capital definitions and ratios throughout the transitional period (Central Bank and ECB).
- Additional supervisory actions:
  - Close data gaps on foreign bilateral liability positions; enhance cross-border monitoring with European counterparts.
  - Continue monitoring U.K. exposures of Irish banks and develop loan-loss forecasting models for U.K. mortgage and corporate exposures.

*Source: IMF staff/FSAP technical note content as provided.*

### EXECUTIVE SUMMARY ___________________________________________________________________________ 5

### EXECUTIVE SUMMARY

### Introduction and context
- The FSAP stress testing exercise took place at a turning point for the Irish financial system: the Irish economy is quickly rebounding, the banking system returned to profitability in 2014, and banks’ exposures to the volatile commercial real estate market have declined significantly; however, the banking system still has a very large stock of non-performing loans.
- Ireland’s banking sector is open and concentrated:
  - Total banking sector assets amount to 294 percent of GDP.
  - The top five banks control 61 percent of system assets.
  - One large privately-owned bank controls 24 percent of total assets.
  - Two state-owned banks make up 27 percent of the banking system’s assets.
  - Subsidiaries of two foreign banks control 10 percent of system assets.
- The FSAP stress tests are designed to assess capacity to withstand extreme but plausible macroeconomic shocks over a three-year horizon (2016–2018), based on data available through June 2015, with capital figures updated based on December 2015 data.

### Design of the stress testing framework
- Tests covered solvency, liquidity, and contagion risks, using both top-down (TD) and bottom-up (BU) approaches.
- Macroeconomic scenarios incorporated main external risks:
  - Protracted weak growth in advanced economies, particularly the euro area, affecting Ireland via lower investment and direct investment inflows.
  - A surge in global financial market volatility that could increase interest rates and raise funding costs.
  - Planned U.K. exit from the EU, with possible disruption to trade, labor mobility, and financial interaction, particularly between Ireland and the U.K.
- Domestic risks incorporated included:
  - A domestic confidence shock leading to consumption and investment collapse.
  - A house price decline reverting towards crisis-era levels.
  - Potential overvaluation and correction risks in commercial real estate after protracted low interest rates.
- Methodologies and data:
  - Effects on individual banks’ profitability and capitalization assessed using satellite models and methodologies developed by the Central Bank of Ireland, the ECB, and Fund staff.
  - TD liquidity tests used maturity ladder analysis and supervisory information, aggregated and by currency.
  - Contagion tests covered interbank exposures among the three largest domestic banks, domestic interlinkages, and cross-border exposures using BIS International Banking Statistics and market data.
- Caveats:
  - Scenarios are extreme but plausible and should not be interpreted as forecasts; adverse scenarios typically reflect events with very low ex ante probability (could possibly materialize once every 20–40 years).
  - Satellite models are estimated on historical data and subject to estimation uncertainty.

### Solvency stress test outcomes
- Under the severe stress scenario on a fully-loaded Basel III basis:
  - Two banks become undercapitalized with regard to the total CAR and Tier 1 capital ratio hurdle rates of 8 percent and 6 percent, respectively.
  - Three banks would have a leverage ratio below the hurdle rate of 3 percent in 2018.
  - Four banks would not meet the Common Equity Tier 1 (CET1) level of 7 percent (combined minimum CET1 ratio and capital conservation buffer level).
- Results under transitional arrangements would be similar, with a reduced capital shortfall.
- Drivers of higher vulnerability for the two undercapitalized banks include lower initial capitalization, asset quality and profitability issues, higher exposure to funding risks, and sensitivity to Basel III capital adjustments.

### Key vulnerabilities and sensitivity findings
- Primary vulnerabilities:
  - Funding risk: funding costs increase sharply under the adverse scenario.
  - Credit risk: loan quality is highly sensitive to unemployment and GDP growth rates, and to a lesser extent to real interest rates.
- In the severe stress scenario:
  - Loan loss provisions rise in parallel with higher probability of default (PD) and loss given default (LGD), negatively impacting profitability.
- Sensitivity analyses:
  - Confirm predominance of credit risks and exposure to sovereign and real estate market risks.
  - Show that high concentration of loan portfolios can exacerbate vulnerability: failure of the five largest exposures causes undercapitalization of the affected bank.

### Liquidity stress test results
- Global liquidity stress tests indicate:
  - Some banks would be exposed to liquidity risks if large deposit withdrawals occur, under scenarios more severe than Basel III LCR variants, or if unsecured wholesale funding dries up.
  - Some banks have material exposure to funding risks in pound sterling.
  - Additional counterbalancing capacity would allow banks to cope with net outflows in every maturity bucket.
- Recommendations arising from liquidity findings (Table 1):
  - Carry out liquidity stress tests currency-by-currency.
  - Examine banks’ liquidity positions in pound Sterling in a stressed environment.

### Contagion and interconnectedness findings
- Bilateral interbank contagion risks among the three largest domestic banks are limited.
- Cross-border linkages:
  - Both foreign banks’ exposures to Ireland and Irish domestic banks’ claims to non-residents have declined dramatically since the crisis.
  - Given deleveraging, cross-border bank linkages are less of a concern than pre-crisis, though tight linkages with the U.K. financial system warrant ongoing attention.
- Market-data-based spillovers:
  - The Irish financial system can receive significant spillovers from U.K. and French banks, and from Italian and Spanish financial systems.
  - The level of interconnectedness has decreased significantly since the onset of the financial crisis.

### Policy recommendations (selected from Table 1)
- Agencies: Central Bank of Ireland (Central Bank), European Central Bank (ECB).
- Near-term (NT — up to 2 years) recommendations:
  - Further develop the Central Bank of Ireland Top-Down credit risk satellite model to increase ability to challenge banks’ submissions operationally.
  - Update Central Bank of Ireland loan-loss forecasting model calibration to ensure out-of-sample forecasting power remains adequate.
  - Carry out liquidity stress tests currency-by-currency.
  - Improve surveillance by closing data gaps on cross-border bilateral financial exposures (Central Bank and ECB).
  - Examine thoroughly banks’ liquidity positions in pound Sterling in a stressed environment (Central Bank and ECB).
- Medium-term (MT — 2–5 years) recommendations:
  - Continue to review the link between sovereign and banking sector risks by reviewing the amount of banks’ AFS reserves for unrealized gains on exposures to central governments (Central Bank and ECB).
  - Continue to monitor banks’ plans to meet Basel III new capital definitions and ratios throughout the transitional period (Central Bank and ECB).
- Note: “NT-near-term” denotes up to 2 years; “MT-medium-term” denotes 2–5 years.

*Source: IMF staff*

### 7.      The TD stress test conducted by the FSAP team had the same coverage as the BU stress

### 7.      The TD stress test conducted by the FSAP team had the same coverage as the BU stress test but a smaller coverage than the TD test conducted by the Central Bank of Ireland

### Overview
- The FSAP team conducted Top-Down (TD) and Bottom-Up (BU) stress tests; TD had the same coverage as the BU test conducted by the FSAP team but smaller coverage than the TD test conducted by the Central Bank of Ireland.
- Due to restricted access to supervisory data and reliance on public information, the FSAP TD and BU stress tests focused on the five systemically important institutions making up 61 percent of the banking sector assets and including two subsidiaries of foreign banks.
- The Central Bank of Ireland’s TD test included, in addition, six non-systemically important institutions.

### Solvency stress tests: design, regulatory framework, and implementation
- Horizon and data:
  - Scenario horizon: three-year horizon (2016–2018).
  - Data available through June 2015, with capital figures updated based on December 2015 data.
- Coverage:
  - FSAP TD test coverage: five systemically important institutions representing 61 percent of banking sector assets, including two subsidiaries of foreign banks.
  - BU stress test: same coverage as FSAP TD.
  - Central Bank of Ireland TD test: included six additional non-systemically important institutions.
- Regulatory framework and hurdle rates:
  - Applied framework: Basel III, the European Union and national frameworks as defined by the Fourth Capital Requirements Directive (CRD IV), national law and Central Bank of Ireland regulation.
  - Hurdle rates:
    - Total capital adequacy: 8 percent
    - Tier 1 capital: 6 percent
    - Common Equity Tier 1 capital (CET1): 4.5 percent
    - Capital conservation buffer: 2.5 percent (note: capital shortfall calculations did not take this buffer into account)
    - Additional hurdle rate for leverage ratio: 3 percent from 2018 onwards
  - Every ratio was based on Basel III fully-loaded definitions.
  - Alternative results considered EU phase-in rules for deductions from CET1 and phase-out portions of instruments that no longer qualify as additional Tier 1 or Tier 2.
- Methodologies:
  - Effects of shocks on profitability and capitalization assessed using satellite models and methodologies developed by the Central Bank of Ireland and IMF staff.
  - Sensitivity stress tests assessed vulnerabilities to individual shocks.
  - Tests included TD and BU exercises based on macroeconomic scenarios and sensitivity analyses.

### Macro-financial risks and macroeconomic scenarios (A)
- Key external risks identified (see Risk Assessment Matrix in Appendix I):
  - Structurally weak growth in advanced economies, particularly in the euro area.
  - Sharp asset price decline and decompression of interest rate spreads.
  - Higher-than-expected fallout from the UK referendum result on EU membership affecting trade, labor mobility, and financial interaction with the EU and the U.K.
  - Financial imbalances from protracted low interest rates, including strong inflows into commercial real estate and potential over-building.
- Banking sector vulnerabilities:
  - NPLs have reduced by almost half since December 2015, but remaining stock remains a challenge; stock composed significantly of long-overdue mortgages, SME and CRE loans.
  - Property repossession has been used scarcely or very slowly despite efforts to accelerate resolution of problem loans.
  - Profitability fragile: returned to profitability in 2014 benefitting from large provision write-backs; profitability close to euro area banks averages in 2015 but impacted by low interest income due to prevalence of tracker loans indexed to the ECB or Euribor rate and reduced credit demand, and low non-interest income.
  - Tracker loans and SVR loans shares as of 2015Q3: tracker loans 50.4 percent; SVR loans 40.1 percent; fixed-rate (>1 year) loans 7.5 percent.
  - Capacity to finance recovery uncertain: capitalization currently adequate but less favorable on a fully-loaded Basel III basis.
- Scenario design:
  - Three-year projection chosen because forecast errors were too large beyond three years.
  - Baseline: combination of Central Bank of Ireland forecasts and October 2015 IMF WEO; forecasts favorable developments with very strong real GDP growth in 2015, slowdown from 2016 onwards, stabilization at 3 percent.
  - Two adverse scenarios developed:
    - First: developed by IMF team using the Global Macro-financial Model (disaggregated into forty national economies).
    - Second: jointly developed by FSAP team and Central Bank of Ireland using IMF Research Department’s Flexible System of Global Models (FSGM) and expert judgment.
  - Adverse scenarios take into account still negative output gap in 2015; milder GDP shocks than 2008/2009 but severe sovereign stress, higher funding costs and asset valuation impacts.
- Two adverse scenarios (text and Table 2 summaries):
  - Moderate (euro area wide) scenario:
    - Designed to reflect a revival in risk aversion affecting peripheral Europe and a sharp slowdown in emerging markets.
    - Irish growth: 1.3 percent in 2016 and 1.2 percent in 2017, pick up to 3.1 percent in 2018.
    - Cumulative decline of GDP relative to baseline over three years: 6.5 percentage points (equivalent to 0.6 standard deviations of the 3-year GDP growth using 1970–2014 benchmark).
    - Price deflation would become generalized; real interest rates would rise; bond and equity prices would fall.
  - Severe (Ireland-specific) scenario:
    - More severe stress with disruption in Ireland’s European partners and domestic shocks: sharp decline in FDI inflows, persistent recession, unemployment rise.
    - Real GDP growth would average 0.5 percent per annum over 2016–2018.
    - Cumulative decline of GDP relative to baseline over three years: about 9.5 percentage points (equivalent to approximately one standard deviation of the 3-year GDP growth using 1970–2014 benchmark).
    - Higher financial volatility, renewed sovereign stress, sharp rise in CDS spreads and banks’ funding costs, imperfect pass-through to lending rates.
    - Reversal in residential and commercial real estate price recovery towards trough levels; collapse in property prices triggers adverse wealth effects and deflationary feedback.
- Table 2: Selected macroeconomic and financial variable projections (Est. 2015; 2016–2018)
  - Real GDP growth:
    - Est. 2015: 5.8
    - Baseline: 4.7, 4.0, 3.0
    - Severe scenario: 1.1, 0.3, 0.1
    - Moderate scenario: 1.3, 1.2, 3.1
  - Unemployment rate:
    - Est. 2015: 9.5
    - Baseline: 8.5, 7.8, 7.2
    - Severe scenario: 9.6, 10.5, 11.2
    - Moderate scenario: 9.5, 9.4, 9.1, 8.4 [note: table shows 2016 9.5; 2017 9.4; 2018 8.4]
  - CPI Inflation rate:
    - Est. 2015: 0.3
    - Baseline: 1.5, 1.9, 1.6
    - Severe scenario: 0.5, 0.3, -0.7
    - Moderate scenario: 1.1, -0.1, -1.0
  - Residential real estate price annual change:
    - Est. 2015: 6.5
    - Baseline: 7.5, 6.7, 5.1
    - Severe scenario: -5.0, -5.0, 0.0
    - Moderate scenario: 1.2, 0.8, 2.6
  - Commercial real estate price annual change:
    - Est. 2015: 21.8
    - Baseline: 10.3, 4.2, 1.9
    - Severe scenario: -4.6, -16.1, -11.2
    - Moderate scenario: 2.9, -5.9, -4.7
  - Nominal annual interest rate (10-year government bond rate):
    - Est. 2015: 1.3
    - Baseline: 1.7, 2.3, 3.3
    - Severe scenario: 3.2, 5.3, 7.9
    - Moderate scenario: 3.6, 4.6, 4.8
  - 10-year sovereign spread (in bp, against German government bond rate):
    - Est. 2015: 75
    - Baseline: 102, 142, 226
    - Severe scenario: 255, 447, 683
    - Moderate scenario: 251, 328, 339

### Credit risks in the scenario analysis (B)
- Credit risk significance:
  - Credit risk in the loan book is the largest asset-side risk for the banking system.
  - Total loans represent 68 percent of total banking sector assets.
- Expected loss computation:
  - Basel II IRB approach formula used: expected losses = PD × LGD × EAD.
  - Times series and panel data models developed to project sectoral or bank-specific PDs.
  - LGD projections based on house and commercial real estate price projections; pre-shock LGD data taken from banks’ annual reports and the 2015 World Bank Doing Business report (global recovery rate for Ireland: 87.7 percent), used in expected losses computation.
- Data access limitations:
  - FSAP team had access to a limited set of supervisory data at individual bank level on solo and consolidated bases in a physical data room (Table 3 summary):
    - Data at the individual bank level: Fully available
    - Data aggregated along groups of banks: Partially available
    - Data aggregated at the banking system level: Fully available
  - Analysis completed with publicly-available data from commercial providers, CreditEdge database and the EBA transparency exercise.
- PD segmentation and modeling:
  - Eight segments for PD projection:
    - Irish household mortgage loans
    - U.K. household mortgage loans
    - Irish non-financial corporates
    - U.K. non-financial corporates
    - Irish financials
    - U.K. financials
    - Irish construction and commercial real estate
    - Irish sovereign exposures
  - Specific modeling approaches:
    - Irish household mortgage loans: panel data model with bank-by-bank supervisory data; dependent variable: annualized 3-month ahead PD; determinants: contemporaneous domestic unemployment rate, real 10-year Irish government bond rate lagged by 4 quarters, bank-specific growth in market share lagged by 4 quarters, bank fixed effects.
    - U.K. household mortgage loans: VAR with quarterly write-off rates and year-on-year house price growth as endogenous variables (lag order 2); exogenous variables: year-on-year real U.K. GDP growth and real 10-year U.K. government bond rate.
    - Irish non-financial corporate loans: year-on-year real GDP growth rate lagged by one period, real 10-year Irish government bond rate lagged by 6 quarters, contemporaneous euro/dollar exchange rate.
    - U.K. non-financial corporate loans: year-on-year real GDP growth rate, real 10-year U.K. government bond rate lagged by 6 quarters, quarterly change in pound/dollar exchange rate.
    - Irish financials loans: contemporaneous year-on-year real GDP growth rate, real 10-year Irish government bond rate lagged by 2 quarters, euro/dollar exchange rate level.
    - U.K. financials loans: contemporaneous year-on-year real GDP growth rate, real 10-year U.K. government bond rate lagged by 6 quarters, year-on-year growth rate of house prices lagged by 4 quarters.
    - Construction and real estate sector companies: contemporaneous year-on-year real GDP growth rate, real 10-year Irish government bond rate, euro/dollar exchange rate level, year-on-year growth in commercial real estate prices.
    - Sovereign exposures: VAR with quarterly EDFs and real 10-year Irish government bond rate as endogenous variables (lag order 3); exogenous variable: year-on-year real GDP growth.
- PD transformation and computation:
  - A logit transformation was applied to original PDs to ensure predictions between 0 and 1:
    - Y_it = ln(PD_it / (1 - PD_it))
  - Logit-transformed PDs modeled as linear functions of exogenous macroeconomic and financial factors and bank-specific variables (with fixed effects and lags as appropriate). Example model forms provided as equations (2) and (3).
  - PDs under stress for each borrower type in percent computed using the inverse logit function:
    - PD_stress, type_t = 100 * { exp(α + X_t β - 1) / [1 + exp(α + X_t β - 1)] }  [formula as provided in source]

*Source: IMF staff/FSAP technical note content as provided.*

### 20.      PDs were projected by banks using quarterly data over the period 2002Q1–2015Q4

### _cr16315 - 20.      PDs were projected by banks using quarterly data over the period 2002Q1–2015Q4

### PD projection models and key drivers
- PDs were projected by banks using quarterly data over the period 2002Q1–2015Q4 and estimating panel fixed-effect or time series OLS models.
- PD projections were based on the most severe results among four models: the banks’ bottom-up submissions, ECB benchmarks, the Central Bank of Ireland loan loss forecasting models, and the Top-Down model developed jointly by the FSAP team and Central Bank of Ireland staff.
- Model design differences:
  - ECB: suite of Bayesian VARs based on time series.
  - Central Bank of Ireland loan-level model: highly granular data with Markov-switching transitions.
  - FSAP team: Top-Down panel-data model.
- IMF Top-Down model key estimated effects (logit transform of the PD):
  - For Irish mortgage loans:
    - When the unemployment rate increases by 1 percentage point, PD rises by 2 percentage points.
    - When the real interest rate increases by 1 percentage point, PD rises by 0.5 percentage points.
    - When the bank’s market share increases by 1 percentage point, PD decreases by 0.3 percentage points one year later.
  - For non-financial corporate (NFC) loans:
    - A 1 percentage point decline in year-on-year real GDP growth increases PD by 2 percentage points.
    - A depreciation of the euro exchange rate against the dollar by a tenth of a point results in a rise in PD by 0.5 percentage points.

### PD and credit-loss projections under scenarios
- System-wide total PD projections (IMF model), from 1.9 percent in 2015:
  - Baseline scenario: PDs remain flat (1.9 percent in 2015 baseline).
  - Moderate stress scenario: peak PD of 2.5 percent in 2018.
  - Severe stress scenario: PD rises to 4.3 percent in 2018.
- Residential mortgage segment PD projection:
  - System-wide PD peak at 4.6 percent in 2018 in the severe stress scenario (Figure 4).
- Credit losses (loan book) under scenarios:
  - Severe stress scenario: credit losses amount to EUR 7.6 bn, equivalent to 2.2 percent of total banking system assets.
  - Baseline scenario: flow of new provisions limited to EUR 2.2 bn, equivalent to 0.6 percent of total banking system assets.

### IMF Credit Risk Satellite Model estimation results (Table 4)
- Dependent variable: logit transform of the PD.
- Selected coefficient estimates (as presented in Table 4):
  - Irish Mortgages:
    - Unemployment rate (in percent): 0.2015*** (t-statistic 11.24).
    - Real interest rate (in percent, lagged by 4 periods/6 periods): 0.0204** (t-statistic 3.18).
    - Bank's yoy growth in market share (in percent, lagged by 4 periods): -0.0113** (t-statistic -2.79).
    - Constant: -5.3984*** (t-statistic -32.24).
    - R-square: 0.71.
    - # of observations: 206.
  - Irish NFCs:
    - YoY real GDP growth (in log, lagged by 1 period): -8.0412*** (t-statistic -8.42).
    - Real interest rate (in percent, lagged by 4 periods/6 periods): 0.0151* (t-statistic 1.64).
    - EUR/USD exchange rate level: -1.6414*** (t-statistic 4.43). (Note: an increase denotes an appreciation of the euro against the dollar.)
    - Constant: -1.0298** (t-statistic 2.06).
    - R-square: 0.55.
    - # of observations: 54.
- Significance notation: * 10 percent; ** 5 percent; *** 1 percent.
- Notes: t-statistics in parentheses.

### Market risks assessed in scenario analysis
- Market risks included: interest rate risk and exchange rate risk. Equity investment risks were not included (negligible share of assets).
- Interest rate risk — time-to-repricing analysis:
  - Repricing buckets used: less than one month; 1 to 2 months; 2 to 3 months; 3 to 6 months; 6 to 12 months; and more than 12 months.
  - Aggregated repricing gap for the 5 banks as of June 2015 for maturities below one year: EUR -8.1 bn.
  - Severe stress scenario: aggregated loss directly due to change in interest rates amounts to EUR82 mn, with no material effect on the CAR over the entire stress horizon.
- Interest rate valuation effects on bond holdings:
  - Severe stress scenario: valuation losses on domestic sovereign securities in AFS and HFT portfolios amount to EUR2.7 bn, contributing 1.6 percentage points to the decline in the CAR over the entire stress horizon.
  - Contributing factors:
    - Average AFS exposure to own sovereign: 4.3 percent of total assets; individual bank ratios between 2 percent and 8.5 percent.
    - Average duration of own sovereign AFS bond portfolio: 5 years.
    - Large increase in Irish government bond rates under the adverse scenario: 6.6 percentage points for the 10-year rate.
- Foreign exchange rate risk:
  - Net open FX position for the banking system as of June 2015: EUR 330 mn, equivalent to 0.1 percent of assets and 0.4 percent of capital.
  - Severe stress scenario: aggregated loss on net FX positions of EUR 2 mn (euro expected to rise against the US dollar and pound sterling).

### Stress-test assumptions and projection mechanics
- Balance sheet growth: bank balance sheet size projected to grow in line with nominal GDP; assets remain constant relative to the economy.
- Risk-weighted assets (RWAs):
  - For Basel II IRB banks, risk weights projected using Basel II formula depending on PD, LGD and asset correlation.
  - RWAs are exogenously constrained not to decrease over the stress period.
- Profit evolution:
  - Non-interest profit items projected to grow in line with nominal GDP, with a floor of 0 for operational and administrative expenses growth.
  - Extraordinary items assumed not to recur during 2016-2018 in baseline and adverse scenarios.
  - Non-performing loans provide no accrued income.
  - Lending rates assumed to follow the path of the 3-month Euribor rate.
  - Banks’ funding costs projected using an econometric model (annual data 2000-2014) with explanatory variables: annual change in the 3-month money market interest rate; annual change in the VIX index; annual change in the amount of interbank deposits in percent. Coefficient on the change in the interbank rate: 0.4.
- Dividends:
  - Banks satisfying capital requirements during the whole stress period assumed to distribute 25 percent of their after-tax profits.
  - Undercapitalized banks in any year were not allowed to distribute dividends.
- Additional methodological notes:
  - Top-down stress test followed a static balance sheet approach (growth in line with nominal GDP).
  - Bottom-up submissions used a dynamic balance sheet approach and incorporated pension scheme balances (FSAP Top-Down did not).

### Solvency stress-test results and capital impacts
- Fully-loaded Basel III aggregate Common Equity Tier 1 (CET1) ratio:
  - Severe scenario: aggregate CET1 would drop by 8 percentage points compared with the starting point and be 7.2 percentage points below the baseline in 2018.
- Main drivers of change in capitalization in the severe scenario (contributions, in percentage points of RWAs):
  - Funding costs: -4.8 percentage points of RWAs.
  - Loss provisions (credit risk): -4.4 percentage points of RWAs.
  - Other comprehensive income linked to valuation losses in the AFS portfolio: -2 percentage points over the whole stress horizon.
  - Basel III capital adjustments: -3.7 percentage points.
- Bank-level outcomes:
  - Two banks would see total CAR and Tier 1 ratios fall below hurdle rates of 8 and 6 percent, translating into capital shortfalls of 0.2 and 0.1 percent of GDP, respectively.
  - Four banks would have CET1 ratios falling below 7 percent (combined minimum CET1 and capital conservation buffer).
  - If transitional arrangements towards Basel III were used, aggregate results similar but capital shortfall limited to 0.1 percent of GDP.
  - Baseline and moderate stress scenarios: every bank would stay above regulatory minima.
- Leverage ratio results:
  - Adverse scenario: three banks would see leverage ratio fall below the hurdle rate of 3 percent in 2018, translating into a capital shortfall of 0.1 percent of GDP.
  - System-level leverage ratio (Tier 1/total assets):
    - 2015 initial level: 9.6 percent.
    - Baseline 2018: 7.4 percent.
    - Moderate stress 2018: 6.5 percent.
    - Severe stress 2018: 5.2 percent.
  - Note: total assets used differ from EBA definition; using EBA definition only two banks would fall below 3 percent in 2018.
- Comparison with bottom-up results:
  - FSAP Top-Down solvency stress test results are more severe than banks’ Bottom-Up submissions.
  - Bottom-Up: no bank undercapitalized under the adverse scenario; only one could not meet conservation buffer.
  - Central Bank of Ireland applied more conservative adjustments to some banks’ results (credit benchmarks, funding shock estimate, interest income, RWA projection).

*Italic: Source — IMF staff calculations, _cr16315 - 20. PDs were projected by banks using quarterly data over the period 2002Q1–2015Q4*

### 37.      Differences in results between the Central Bank of Ireland TD model and the IMF’s

### Differences in results between the Central Bank of Ireland TD model and the IMF’s

### Methodological drivers of differences
- Central Bank of Ireland (CBI) based results on banks’ submissions adjusted by more conservative parameters.
- CBI adjustments lowered banks’ CET1 ratio projections by a range of 1 to 3 percentage points of RWAs and resulted in:
  - one bank being undercapitalized under the adverse scenario in terms of CET1 ratio, and
  - another bank under the capital conservation buffer level.
- Key methodological differences between FSAP/IMF and CBI approaches:
  - Net interest income projections: FSAP team conservatively assumed no accrued interest on non-performing loans; CBI allowed discount unwind in line with EBA methodology.
  - AFS portfolio shock: FSAP top-down model adopted a point in time shock; CBI used a dynamic approach.
  - Pension fund balances: CBI allowed changes; absent from IMF methodology.
  - Coverage of less systemically-important institutions (LSIs): LSIs were not covered by the IMF Top-Down model but were included in the CBI Top-Down test.
- Result for LSIs (CBI Top-Down test): LSIs’ CET1 ratio falls by 4 percentage points under the adverse scenario but remains largely above the minimum requirement and the capital conservation buffer, at 18.6 percent.

### Table 5: Results of the Bottom-Up and Top-Down Solvency Stress Tests (adverse scenario)
- Table contents as presented in source:
  - Banking system's CET1 ratio (in percent) | Number of banks with 4.5%<CET1<7% | Number of undercapitalized banks (CET1<4.5%) | Number of undercapitalized banks (leverage ratio<3%) | Max. capital shortfall in terms of CAR, T1, CET1 or leverage ratio (percent of GDP)
  - Fully-loaded Basel III
    - Bottom-Up Stress Test12.410-0.00
    - CBI Top-Down Stress Test10.111-0.05
    - IMF Top-Down Stress Test8.34030.20
  - European framework for transitional arrangements
    - Bottom-Up Stress Test14.310-0.00
    - CBI Top-Down Stress Test11.901-0.04
    - IMF Top-Down Stress Test9.44020.10
- Sources: Central Bank of Ireland and IMF staff calculations

### Sensitivity analyses: market and macro risks (static tests as of June 2015)
- Overview:
  - Sensitivity tests assessed instantaneous impacts on banks’ balance sheets as of June 2015; RWA assumed constant after shocks.
  - Shocks covered: unemployment, domestic sovereign prices, interest rates (NII and credit quality), house prices, nominal effective exchange rate depreciation/appreciation, and nominal depreciation effects on borrower credit quality.

- Unemployment shock
  - Single-factor test: unemployment rate up to 12.7 percent (i.e., 3.2 percentage points higher than in 2015, 1.5 percentage points higher than the peak of the macro scenario and in line with the 2016 EBA stress test scenario).
  - IMF credit risk model result: global PD increases by 1.6 percentage points (from 1.9 percent to 3.5 percent).
  - Consequences: credit loss provisions expenses equal to EUR1.5 bn and a decline in total CAR by 0.9 percentage points.
  - Impact on solvency: would not cause undercapitalization in any of the five banks.

- Domestic sovereign securities price shock
  - Test: 500 bp increase in interest rates applied to domestic bond AFS and HFT portfolios using modified duration.
  - Result: CAR in the system would decline by 2 percentage points taken in isolation.
  - Impact on solvency: comfortable initial capitalization on a transitional basis would provide a large enough buffer to avoid undercapitalization.

- Interest rate risk — net interest income effects
  - Gap analysis: 500 basis point increase in interest rates.
  - Result: banking system would lose EUR 403 mn euro in net interest income and aggregate Tier 1 capital ratio would fall by 0.2 percentage points.
  - Impact on solvency: would not cause undercapitalization in any of the five banks.

- Interest rate risk — effects on credit quality
  - Test: 500 basis point increase in domestic real interest rates.
  - IMF credit risk model result: global PD increases by 1.4 percentage points (from 1.9 percent to 3.3 percent).
  - Consequences: credit loss provisions expenses equal to EUR 1.5 bn and a decline in total CAR by 0.9 percentage points.
  - Impact on solvency: would not cause undercapitalization in any of the five banks.
  - Caveats: test assumes no pre-impairment profits and output unchanged; ignores second-round effects.

- Foreign exchange rate risk — direct effects on banks with net open FX positions
  - Net open FX position for banking system: EUR219 mn as of June 2015, equivalent to 0.05 percent of assets.
  - For the five largest banks: EUR330 mn; each has a net long position.
  - Test results:
    - 30 percent appreciation of the euro nominal effective exchange rate → loss of EUR99 mn and lower CAR by 0.1 percentage points; no undercapitalization.
    - 30 percent depreciation → gain of EUR99 mn and increase in CAR by 0.1 percentage points.

- Foreign exchange rate risk — indirect effects on credit quality from nominal depreciation
  - Test: 30 percent depreciation of the euro against the US dollar.
  - IMF credit model result: total PD increases by 0.7 percentage points (from 1.9 percent to 2.6 percent).
  - Consequences: loss of EUR1.1 bn and lower CAR by 0.7 percentage point.
  - Impact on solvency: would not cause undercapitalization in any bank.
  - Caveats: does not include potential positive macro effects of depreciation; assumes no pre-impairment profits and constant output.

- Real estate price decline
  - Test: 30 percent decline in domestic real estate (housing and CRE) prices.
  - Result: system LGD rises by 16 percentage points (from 16½ percent to 32½ percent).
  - Consequences: aggregate loss provisions of EUR1.7 bn; loss of CAR equivalent to 1 percentage point.
  - Impact on solvency: partial effect only; does not capture full macro feedbacks.

### Concentration risk: failure of large corporate exposures
- Method: hypothetical default of up to ten of the largest borrowers using supervisory data; excludes sovereign exposures.
- Findings:
  - Average size of the net largest exposure (after conversion factors) is 4.9 percent of total regulatory capital.
  - Default of the largest exposure of each of the five largest banks would cause no undercapitalization under national regulation.
  - Simultaneous default by five to ten largest exposures would cause undercapitalization in one bank.
  - Default of the five largest exposures → Tier 1 capital shortfall vs 6 percent hurdle = EUR 167 mn, equivalent to 0.1 percent of GDP.
  - Default of the ten largest exposures → Tier 1 capital shortfall = EUR639 mn, equivalent to 0.3 percent of GDP.

### Operational risks
- Historical-loss based estimate (CBI and FSAP sample of severe incidents across five largest institutions): losses would amount to 0.73 percent of total assets on aggregate.
- Limitations: large-scale operational incidents (e.g., technology-related or cyberattacks) are difficult to quantify from historic loss data and are not included in macro stress scenarios.

### Liquidity stress tests (framework and top-down implementation)
- Approaches used:
  - National transposition of the Liquidity Coverage Ratio (LCR), Net Stable Funding Ratio (NSFR), and a cash flow-based maturity bucket analysis.
- Key parameters and timelines:
  - LCR: banks must maintain an LCR above 60 percent from October 2015 and 100 percent from 2018.
  - National transposition differences from Basel III LCR:
    - Includes a larger range of high quality liquid assets but with high haircuts (e.g., equity haircut of 50 percent).
    - Greater granularity of deposits.
    - Accelerated phase-in to 100 percent by 2018 (vs 2019 under Basel III).
  - NSFR: will require a stable funding profile at a one-year horizon from 2018.
- Top-down cash-flow based liquidity stress tests:
  - Implemented jointly by CBI staff and FSAP team using supervisory maturity-structure data at June 2015.
  - Tests conducted at aggregate level combining currencies and separately on pound sterling (GBP) positions.
  - Scenarios included: standard LCR stress, more severe deposit withdrawal scenario, and a third scenario with a dry-up of unsecured wholesale funding calibrated to very severe stress conditions (e.g., 2008/2009 crisis experience).
  - Tests capture: (i) liquidity need from outflows; (ii) available standby liquidity from inflows; (iii) buffers to counterbalance liquidity gaps.
- Note on calibration: common FSAP practice assumes funding pressures are sizeable but limited to a number of banks (not systemic); generalized panics would require higher run-off rates and larger haircuts.

*Source: IMF staff and Central Bank of Ireland material as presented in the source content.*

### 53.      The LCR-based stress test was based on three scenarios, with various parameters in

### The LCR-based stress test was based on three scenarios, with various parameters in

### LCR-based stress test: setup and assumptions
- Funding structure of the banking system (excluding capital and including contingent credit lines) as of June 2015:
  - 77 percent of funding comes from customer deposits;
  - 19 percent is wholesale funding (Debt Capital Markets 11 percent , Interbank deposits and Repo 8 percent);
  - 4 percent of funding comes from central banks.
- Run-off and roll-off principles:
  - More informed and sophisticated depositors withdraw funding more rapidly than less informed ones → wholesale run-off rates higher than retail.
  - Secured funding sources have lower run-off rates than unsecured funding sources.
- Inflows:
  - Banks’ standby liquidity inflows stem mostly from maturing loans, deposits and credit facilities.
  - Assets that can generate cash inflows over one month include: maturing loans from retail counterparties, and level 1 assets.
  - Roll-off rates applied: 50 percent for inflows from retail and nonfinancial wholesale counterparties (performing loans to non-financial customers); 100 percent for maturing loans to financial institutions.
- Coverage and collateral:
  - In tests, banks were allowed to cover negative balances by using sovereign securities as collateral to obtain liquidity through the standard operations of the Eurosystem.
  - At the banking system level, liquid assets make up 20 percent of total assets (including both on- and off-balance sheet items).
- Table 6 (summary presentation in source) specifies run-off, roll-off and haircut parameters (presented in the source).

### LCR-based stress test: results and scenarios
- Baseline (using the European Commission Regulation parameters):
  - Aggregate LCR at June 2015: 124 percent.
  - Every bank passed the 60 percent hurdle rate (national regulation in 2015); four of the five banks are above the 100 percent hurdle rate (binding in 2018).
- Adverse scenario 1 (retail shock, more severe than 2008/2009 peak experience):
  - Assumptions: banks lose 10 to 15 percent of their retail deposits (sight deposits and term deposits with residual maturity below 30 days) and 10 to 60 percent of their non-financial corporate deposits in a month.
  - Aggregate LCR would fall to 87 percent.
  - Liquidity shortfall: EUR 7.1 bn, equivalent to 3.3 percent of GDP.
- Adverse scenario 2 (dry-up of unsecured wholesale funding):
  - Assumptions: 100 percent run-off rates on unsecured wholesale funding; operational outflow rates increase by 45 percentage points; non-operational outflows covered by DGS increase by 30 percentage points; non-covered by DGS increase by 60 percentage points.
  - Aggregate LCR would fall to 87 percent (same aggregate ratio as retail-shock scenario).
  - Total liquidity shortfall: EUR 7.4 bn, equivalent to 2.8 percent of these banks’ assets.
- Pound sterling liquidity stress:
  - Aggregate liquidity shortfall in pound sterling ranges between EUR1.7 bn and 6.0 bn, i.e. 0.8 percent and 2.8 percent of GDP, across the various LCR scenarios.

### NSFR-based stress test: results
- For the 5 largest banks as of June 2015:
  - Available stable funding: EUR211.3 bn.
  - Required stable funding: EUR196.4 bn.
  - Aggregate NSFR: 107.6 percent.
  - Nevertheless, some banks have NSFR below 100 percent, resulting in an aggregate liquidity shortfall of EUR7.4 bn, equivalent to 3.4 percent of GDP.

### Outflow analysis stress test: methodology and results
- Maturity buckets used: below 30 days, between 1 and 2 months, 2–3 months, 3–6 months, 6–9 months, 9–12 months, 1–2 years, 2–3 years, 3–5 years, 5–10 years, and above 10 years.
- Mapping and parameters:
  - Deposit outflows mapped by banks to LCR categorization.
  - LCR parameters applied, including a 50 percent allowance on inflows for retail and corporate obligations.
  - Inflows capped at 75 percent of outflows as in the LCR.
  - For each bucket, net outflows compared to sum of HQLA buffer and an expanded counterbalancing capacity (stressed buffer) comprising retained securitized assets.
  - Stress of counterbalancing capacity based on: sovereign-yield-derived haircuts from Irish sovereign stress test; haircuts on ABS and ACS based on downgrade using ECB collateral schedule; haircuts on retained ABS and ACS based on historic pricing; pool of securities stressed using increased PD and house price decrease from Irish solvency stress test.
- Findings:
  - The banking system as a whole would have enough buffers to counterbalance net outflows for a wide range of maturity buckets if an expanded range of counterbalancing capacity is included.
  - Irish banks would be able to survive the assumed stress conditions for maturities up to 10 years with counterbalancing capacity expanded to securitized, stressed assets.
  - Banks would experience liquidity shortfalls for maturities beyond 3 months if only standard high-quality liquid assets (HQLA) are considered.

### Interconnectedness and contagion risks
- Domestic interbank contagion (three largest domestic banks, end-June 2015 exposures):
  - Analysis uses network model based on Espinosa-Vega and Solé (2010) with gross bilateral interbank exposures.
  - Interbank exposures are mainly on the asset side and taken from the quarterly large exposures return.
  - Fire-sale assumptions: 50 percent haircut in fire sale of assets; 65 percent roll-over ratio of interbank debt.
  - Results:
    - Contagion risks stemming from purely domestic interbank exposures are very limited.
    - Interbank positions between the three largest domestic banks are small compared to these banks’ capitalization.
    - The sum of their gross exposures to the other two banks is smaller than their regulatory capital.
    - No single failure of a domestic bank would trigger the failure of another domestic bank; no cascade effect via this three-bank market.
    - As of end-June 2015, none of the three largest banks is undercapitalized relative to the regulatory minimum after shocks on interbank exposures to the other two.
    - One of the three largest banks has a higher vulnerability index (percentage loss at a single institution due to default of all other institutions) that is twice and four times as high as for the other two banks, respectively, but remains at a very low level, below 9 percent.

*Italic source: Excerpt from IMF staff report content unit _cr16315 - pages 53–75 (Ireland liquidity and interconnectedness analysis as presented in the source).*

### 73.      Exposures to non-bank private sectors, mostly in the form of local claims, accounted

### Exposures to non-bank private sectors and cross-border contagion risks

### Composition of foreign claims of Irish domestic banks
- Claims on non-resident private sectors accounted for 76 percent of total foreign claims (CBS-URB) in September 2015.
- Claims on foreign banks accounted for 11 percent, and claims on public sectors accounted for 13 percent of total foreign claims in September 2015.
- Local claims constituted 70 percent of total foreign claims at 2015Q3 and were mainly held by the U.K. subsidiaries of Irish banks.
- The claims are mostly in the form of mortgage loans, exposing them to the U.K. property cycle.

### BIS statistics and data scope
- The note uses BIS Consolidated Banking Statistics on a consolidated, head-office (CBS) basis rather than Locational Banking Statistics (LBS) to focus on risks that banks transfer in and out of the Irish economy.
- CBS data are available on an immediate borrower basis (CBS-IBB) and an ultimate risk basis (CBS-URB).
- Because local claims in foreign currency are minimal in Ireland, the size of cross-border claims under the CBS-URB is similar to international claims under the CBS-IBB.
- Caution: CBS on the immediate borrower basis include exposures to the International Financial Services Centre, which can overstate risks.

### Balance and share of foreign claims on non-residents (selected table figures)
- 2008Q1 balances and shares (US$ billion and Percent):
  - France: 20.5 (4.6 percent)
  - Germany: 21.8 (4.9 percent)
  - Poland: 17.7 (4.0 percent)
  - Spain: 16.3 (3.7 percent)
  - United Kingdom: 246.8 (55.9 percent)
  - United States: 39.5 (9.0 percent)
  - Rest of the World: 78.6 (17.8 percent)
  - Total foreign claims: 441.3 (100.0 percent)
- 2015Q3 balances and shares (US$ billion and Percent):
  - France: 5.0 (4.7 percent)
  - Germany: 0.9 (0.9 percent)
  - Poland: 0.4 (0.3 percent)
  - Spain: 2.8 (2.7 percent)
  - United Kingdom: 73.9 (70.6 percent)
  - United States: 6.1 (5.8 percent)
  - Rest of the World: 15.7 (15.0 percent)
  - Total foreign claims: 104.6 (100.0 percent)

### Cross-border network simulation framework and calibration
- Network represented by matrix of foreign claims for 17 countries (Australia, Austria, Belgium, Canada, Finland, France, Germany, Ireland, Italy, Japan, Netherlands, Portugal, Spain, Sweden, Switzerland, United Kingdom, United States) using BIS CBS-IBB at end-2014.
- Total regulatory capital taken from IMF Financial Soundness Indicator statistics; capital buffer figures as of end-2014.
- Two types of simulations under assumption that new capital cannot be raised:
  - Credit shock: failure of banking system A causes credit losses to banks with claims on A; failures propagate if capital insufficient.
  - Credit-funding shock: failure of system A creates funding shortfalls for systems with liabilities to A; forced asset sales at discounts may create additional losses.
- Baseline parameter values used: loss given default ratio (ߣ) = 1, loss of funding ratio (ߩ) = 0.5, discount rate (ߜ) = 0.5.
- In joint credit and funding simulations a 50 percent haircut in fire sales and a 50 percent rollover ratio of cross-border funding are assumed.

### Simulation findings — credit shock only
- Hypothetical failure of the entire U.K. banking system would result in a total meltdown of the global banking system after three contagion rounds.
- Hypothetical failure of the entire U.S. banking system would result in a total meltdown of the global banking system after five contagion rounds.
- Ireland is directly hit by a U.K. banking failure due to highly concentrated exposure across the Irish Sea; a U.S. failure would impact Ireland in a second round via contagion from the U.K.
- Ireland’s index of vulnerability (average capital impairment due to failure of all other banking systems) is 17 percent.
- Excluding the U.K. and the U.S., Ireland’s index of vulnerability drops to 5 percent.
- A banking crisis in Ireland would deprive only 7 percent of total regulatory capital buffers in the global banking network and would not generate domino effects.

### Simulation findings — joint credit and funding shocks
- When credit and funding shocks are considered jointly, Germany and France become additional sources of negative spillovers toward Ireland.
- For Germany, the number of affected countries increases from one to sixteen when funding shocks are included.
- Ireland would be affected by a German failure after three rounds of contagion (via the Netherlands in the first round, and Belgium, Italy, and Sweden in the second round).
- A failure of the French banking system would not induce failure of the Irish banking system but would deprive it of 90 percent of its capital buffers as the shock spills over to Belgium and the Netherlands.
- Ireland’s index of contagion (credit losses to other banking systems caused by an Irish banking system failure) stays below 10 percent under joint shocks and is not associated with any failures.

### Market-price network analysis (equity returns)
- Variance decomposition of weekly equity returns shows Irish financial institutions are significantly affected by banks in the U.K. and France, and financial systems in Spain and Italy.
- The individual contributions of these four sectors to Irish inward spillovers each exceed 5 percent; their combined contribution reaches 28½ percent.
- Net inward connectedness of the Irish financial system within the representative network is 7.7 percent, indicating Ireland is a net recipient of spillovers.

### Interpretation and channels
- The significant impact of U.K. banks reflects: presence of a large U.K. bank subsidiary in Ireland, a large Irish bank subsidiary in the U.K., and funding interlinkages between the two systems.
- The large effect of French banks may reflect their worldwide systemic importance (size and interconnectedness) and direct/indirect common exposures (including to the U.K. real estate market and exposures linked to exports and FDI in continental Europe’s industry sector).

*Source: IMF staff analysis using BIS CBS-IBB and CBS-URB, Central Bank of Ireland data, and IMF Financial Soundness Indicator statistics (as presented in the supplied content).*

### 85.      A historic comparison shows that cross-border linkages were higher before the global

### _cr16315 - 85.      A historic comparison shows that cross-border linkages were higher before the global

### Historical cross-border linkages and spillovers
- The spillover index, defined as the total amount of forecast error variance attributable to spillovers, remained broadly similar in the pre- and post-crisis periods, at above 70 percent.
- The Irish financial system’s share of spillovers to itself increased by 2.7 percentage points to 32 percent of total spillovers.
- Sectors with the largest falls in contribution: French insurers, the MSCI world real estate index, and Greek banks — reflecting post-crisis fragmentation of real estate markets, decline in banks’ exposures to commercial real estate markets, and the specific situation of Greek banks.
- Foreign banks sold subsidiaries and reduced interbank exposures to Greek banks, thereby reducing international financial linkages.

### Methodology limitations (caveats)
- Post-crisis market data informational content is low for some Irish banks due to large government ownership and government guarantees.
- The ISEQ financials index was used as best available data but:
  - It is dominated by one large bank which makes up the majority (90 percent) of the index market capitalization.
  - The index is not diversified; other constituents are mainly banks and include no real estate companies.
- Use of a generalized forecast-error variance decomposition:
  - Does not allow for inferring causality with certainty.
  - Granger-causality tests do not fully rule out common exposures.
  - Methodology is intended to complement supervisory and balance-sheet analysis; convergence between methods is reassuring.

### Summary of cross-border exposures and network analysis
- Deleveraging has led cross-border bank linkages to be less of a concern compared with the pre-crisis period.
- Both foreign banks’ exposures to Ireland and Irish domestic banks’ claims to non-residents have reduced dramatically.
- BIS Consolidated Banking Statistics indicate Ireland is exposed, through direct cross-border credit and funding exposures, to financial distress in the four largest banking systems: the U.K., the U.S., Germany, and France.
- Network-analysis findings:
  - The Irish banking system would be directly affected by distress in the U.K. banking system due to highly concentrated exposures.
  - Distress in the U.S., German, and French banking systems would generate domino effects through the U.K., Belgian, and Dutch banking systems.
  - A banking crisis in Ireland is not expected to cause failures in any other banking system.
- Market-data based interconnectedness analysis suggests spillovers from U.K. and French banks, as well as Spanish and Italian financial systems, towards the Irish financial system.
- Ireland’s exposure to international spillovers has declined significantly since the onset of the financial crisis.

### Ongoing concerns and supervisory actions
- Tight linkages with the U.K. financial system warrant ongoing attention.
  - Irish domestic banks have large exposures to the U.K. relative to the size of their balance sheets; severe U.K. distress would inflict large losses on Irish banks.
- Central Bank of Ireland actions welcomed:
  - Close monitoring of U.K. exposures of Irish banks.
  - Development of loan-loss forecasting models for stress-testing purposes, including models for U.K. mortgage exposures, corporate loans, and other consumer loans for the three largest domestic banks.

### Stress tests: design and quantitative results
- Stress-test framework:
  - Top-down stress tests by FSAP team and Central Bank of Ireland staff, supported by the ECB on selected issues.
  - Bottom-up stress tests by five banks using internal methodology and Central Bank guidance.
  - Quantitative analysis included macroeconomic scenario-based stress tests (one baseline and two adverse scenarios) and sensitivity analysis.
  - Risk coverage: credit risk in loan book; market risk effects on interest income and valuation of debt instrument holdings; exchange rate related risks; concentration risk; domestic interbank contagion; cross-border interconnectedness; liquidity stress tests including pound sterling liquidity.
- Main stress-test results:
  - Several sources of solvency vulnerability remain but have declined markedly since the financial crisis and are manageable at the macro level.
  - Under the severe stress scenario on a fully-loaded Basel III basis:
    - Two banks become undercapitalized with regard to the total CAR and Tier 1 capital ratio hurdle rates of 8 percent and 6 percent, respectively.
    - Three banks would have a leverage ratio below the hurdle rate of 3 percent in 2018.
    - Four banks would not meet the Common Equity Tier 1 (CET 1) level of 7 percent (combined minimum CET1 ratio and the capital conservation buffer level).
    - Results would be similar under transitional arrangements, with a reduced capital shortfall.
  - Sources of higher vulnerability for the two weaker banks include lower initial capitalization, asset quality and profitability, higher exposure to funding risks, and sensitivity to Basel III capital adjustments.
  - Funding and credit risks are the two main vulnerabilities:
    - Funding costs increase sharply under the adverse scenario.
    - Loan quality is very sensitive to changes in the unemployment rate and to GDP growth, and to a lesser extent to real interest rates.
    - In the severe stress scenario, loan loss provisions rise with higher PD and LGD, hurting profitability.
    - Sensitivity tests confirm predominance of credit risks, exposure to sovereign and real estate market risks; in one bank, high concentration means the failure of the five largest exposures would cause undercapitalization.
  - Global liquidity stress tests:
    - Some banks would be exposed to liquidity risks in the event of large deposit withdrawals, under a scenario more severe than Basel III LCR metrics, or a dry-up of unsecured wholesale funding.
    - Some banks display material exposure to funding risks in pound sterling.
  - Contagion analysis:
    - Banks are less vulnerable to direct contagion through bilateral exposures or cross-border contagion.
    - Interbank exposure contagion between the three largest domestic banks is very limited.
    - Both foreign banks’ exposures to Ireland and Irish domestic banks’ claims on non-residents have reduced dramatically since the financial crisis.
    - Irish financial system receives spillovers from U.K. and French banks, as well as Italian and Spanish systems, but interconnectedness has decreased significantly since the crisis.

### Conclusions and policy recommendations
- Systemic resilience and vulnerabilities:
  - The Irish banking system appears more resilient than in prior years but remains exposed to large vulnerabilities.
  - Two banks appear weaker than the other three systemically-important ones due to lower asset quality and exposure to solvency risks; one of these also faces liquidity risks; another bank is significantly exposed to sovereign risk.
  - The two weaker banks are not highly interconnected with other domestic banks, suggesting limited direct spillovers from potential losses; indirect contagion risks (e.g., reputational) were not assessed and could be significant.
- Data and cross-border monitoring:
  - Authorities need to work closely with European counterparts and agencies to close data gaps on cross-border bilateral financial exposures.
  - Current data: individual Irish banks’ cross-border asset positions vis-à-vis counterparties are available to the Central Bank of Ireland, but information on foreign bilateral liability positions is incomplete.
  - Future data collection enhancements will aid monitoring of systemic risk related to cross-border linkages.

### Risk Assessment Matrix: selected high-level entries
- I. Structurally weak growth in key advanced and emerging economies
  - Overall Level of Concern: High
  - Relative Likelihood: Medium (advanced economies downside risk); Medium (emerging economies downside risk)
  - Expected Impact: Ireland’s exports were equivalent to 110 percent of GDP in 2015; EA took 35 percent of total exports. Export-growth impact would be significant; weak EA could undermine domestic confidence, investment, and FDI inflows.
- II. Sharp asset price decline and decompression of interest rate spreads
  - Overall Level of Concern: Medium
  - Expected Impact: Ireland’s high level of private and public debt makes it susceptible to financial contagion; impact on deficits and debt limited by low financing needs and substantial cash buffers, but growth impact could be significant if CRE inflows reverse.
- III. Higher-than-expected fallout from the UK referendum result on EU membership
  - Relative Likelihood: Medium; High if process is volatile or leads to large increase in barriers.
  - Expected Impact: Sharper slowdown in the UK and rest of Europe, persistent uncertainty, and high market volatility would adversely affect the Irish economy; depreciation of pound Sterling would reduce exports to the UK; Irish banks’ profitability would decline; asset quality may deteriorate in Ireland and the UK. Some effects mitigated by potential relocation of firms from the U.K. to Ireland increasing FDI.
- IV. Financial imbalances from protracted low interest rates
  - Overall Level of Concern: Medium
  - Expected Impact: Low-Medium. Risk of CRE over-building from search for yield; potential residential real estate price bubble or consumer lending boom; high concentration in banking sector increases systemic similarity of strategies; low domestic credit growth limits current risk.
- V. Persistently lower energy prices
  - Overall Level of Concern: High
  - Expected Impact: Low. Lower oil prices reduce inflation and inflation expectations; could raise savings and lower investment; lasting low energy prices would reduce production costs and increase real incomes.
- VI. Domestic reform fatigue and political fragmentation
  - Overall Level of Concern: Medium
  - Expected Impact: Medium. Public pressure to reverse prudential measures could increase vulnerabilities.

### Stress Test Matrix (STeM) — key assumptions (solvency, liquidity, contagion)
- Institutional perimeter
  - Bottom-up by banks: 5 banks; market share: 61 percent of banking sector’s assets; baseline date: end-June 2015; bank proprietary data; bank consolidated level data for Irish-headquartered banks and sub-consolidated for foreign bank subsidiaries.
  - Top-down by Central Bank of Ireland: 11 banks; market share: 85 percent of banking sector’s assets; baseline date: end-June 2015; supervisory data; consolidated/sub-consolidated data.
  - Top-down by FSAP Team: 5 banks; market share: 61 percent of banking sector’s assets; baseline date: end-June 2015; publicly-available and supervisory data; market-data.
- Channels of risk propagation and methodology
  - Bottom-up: Guidelines from Central Bank of Ireland and FSAP team and banks’ internal stress-testing methodology; satellite models for credit losses, pre-impairment income, credit growth, pensions, valuation of marketable securities; expert judgment.
  - Top-down (Central Bank of Ireland): Satellite models developed by Central Bank of Ireland and the ECB; methodologies for AFS sovereign risk, idiosyncratic funding shock, income projections including accrued income on NPL loans, pension risk, RWA forecast.
  - Top-down (FSAP Team): Satellite models developed by FSAP team; balance sheet-based and market-data-based approaches; models to integrate solvency-funding interactions; methodologies to calculate sovereign risk and losses from bonds and money market instruments using a modified duration approach; net fee/commission and operating expenses projected based on nominal GDP growth; no accrued income on NPL loans.
- Stress test horizon
  - Bottom-up: 3-years (2016/2018)

*Source: IMF staff calculations and supporting analysis in the provided content.*

### 3. Tail shocks

### 3. Tail shocks

### Scenario analysis
- Scenario-based tests assess impacts on the entire portfolio including the loans and, if applicable, the trading book (TD exercise).
- Variables in scenarios include domestic macro-financial variables (e.g., GDP, inflation), GDP for key trading partners, interest rates, and real estate prices.
- Ireland-specific adverse scenario:
  - GDP growth rate declines to 1.1, 0.3 and 0.1 percent, in 2016, 2017 and 2018 respectively.
- A set of market shocks, including large and sudden changes in interest rates, is calibrated to magnitudes close to those observed in 2008/2009.

### Sensitivity analysis
- Sensitivity analyses are conducted in the BU and TD exercises.
- They evaluate domestic shocks including:
  - Direct effects of interest rate shocks.
  - Interest rate shock on credit quality.
  - Credit quality effect of unemployment rate.
  - Direct and indirect effects of exchange rate shocks.
  - A decline in the prices of sovereign bonds and real estate.
  - Failure of the largest to 10 largest corporate exposures.

### Risks and buffers (Credit and Market)
- Risks/factors assessed:
  - Credit risk on the banking book and trading book.
  - Market risk and bond losses: direct effects of interest rate shocks; direct effects of exchange rate shocks; shocks to sovereign bond yields.
  - Market risk, including foreign exchange risk.
  - Credit losses.
  - Losses from bonds and money market instruments (sovereign and other issuers) in the banking and trading books.
  - Funding costs.
- Behavioral adjustments assumptions:
  - Bottom-up by banks: Dynamic balance sheet assumption.
  - Top-down by Central Bank of Ireland: Dynamic balance sheet assumption; No dividends are paid out.
  - Top-down by FSAP Team: Balance sheet grows with nominal GDP; Dividends are paid out by banks that remain adequately capitalized throughout the stress.
- Calibration of risk parameters:
  - Point in time risk parameters for credit risk parameters or proxies (bottom-up, top-down Central Bank of Ireland, top-down FSAP Team).
- Regulatory/accounting frameworks referenced:
  - National regulation.
  - Basel II IRB approach + Basel III.

### Banking Sector: Liquidity Risk (Top-down by Central Bank of Ireland and FSAP team jointly)
- Institutional perimeter:
  - Institutions included: 5 largest banks in the system.
  - Market share: 61 percent of banking sector’s assets.
  - Data and baseline date: Latest data: June 2015. Source: supervisory data. Scope of consolidation: perimeter of individual banks.
- Channels and methodology:
  - NSFR and LCR type proxies.
  - Cash-flow based liquidity stress test using maturity buckets by banks.
- Risks and buffers:
  - Funding liquidity (liquidity outflows).
  - Market liquidity (price shocks).
  - Buffers: Counterbalancing capacity; Central bank facilities.
- Tail shocks and size:
  - Run-off rates calculated following historical events and LCR/NSFR rates.
  - Bank run and dry up of wholesale funding markets, taking into account haircuts to liquid assets.
- Regulatory standards cited:
  - European Commission Delegated Regulation (EU) 2015/61; and Basel Committee on Banking Supervision (2014), “Basel III: The Net Stable funding ratio”.
- Reporting / output presentation:
  - Liquidity gap by bank, and aggregated.
  - Survival period in days by bank, number of banks that can still meet their obligations.

### Banking Sector: Contagion Risk
- Institutional perimeter:
  - Institutions included: 3 banks; Financial sub-index of ISEQ index.
  - Market share: 50 percent of total banking system assets.
  - Data and baseline date: Latest data: June and December 2015. Source: supervisory and market data. Scope of consolidation: perimeter of individual banks.
- Channels and methodology:
  - Network interbank model by Espinosa-Vega and Solé (2010).
  - Diebold-Yilmaz variance decomposition connectedness methodology.
  - Data-driven correlation networks.
- Tail shocks (size of shock):
  - Pure contagion: default of institutions.
  - Spillover index and transmission.
- Reporting / output presentation:
  - Number of undercapitalized and failed institutions, and their shares of assets in the system.
  - Evolution and direction of spillovers within the network.

### Appendix highlights: BIS International Banking Statistics (key characteristics)
- Locational Banking Statistics included:
  - Cross-border transaction within a financial group.
  - On-balance sheet items for both claims and liabilities.
  - On an immediate counterparty basis.
  - By sector (bank vs. non-banks).
  - 44 reporting countries.
- Locational Banking Statistics excluded:
  - Off-balance sheet items.
  - Disregard risk transfers.
  - Bilateral data are confidential, need to request to BIS.
- Consolidated Banking Statistics included:
  - Cross-border claims only.
  - Key off-balance items (guarantees extended, credit commitments and derivatives).
  - Immediate risk basis and ultimate risk basis coverage with 31 and 26 reporting countries respectively.
  - By counterparty sector (bank, official sector, non-bank financial institutions, non-bank private sector).
- Consolidated Banking Statistics excluded:
  - Cross-border transactions within a financial group.
  - Domestic claims in the reporting country.
  - Data by sector on immediate risk basis.
  - Off-balance sheet items excluding guarantees, credit commitments and derivatives.
- Note: Bilateral data are publicly available.

### Appendix IV: Size of Capital Buffers in the Sample Countries (As of end-2014)
- Selected entries (Millions of US dollars; Millions of Local Currency):
  - Ireland: 27,080; 22,379.
  - United States: 1,559,796; 1,559,796.
  - United Kingdom: 861,848; 553,220.
  - France: 459,416; 379,662.
  - Germany: 532,663; 440,193.
  - Spain: 268,273; 221,701.
  - Japan: 366,799; 43,961,000.
  - Netherlands: 156,405; 129,253.
  - Sweden: 82,765; 647,594.
  - Australia: 174,227; 213,167.
  - Austria: 74,864; 61,868.
  - Belgium: 74,449; 61,525.
  - Canada: 192,713; 223,566.
  - Finland: 24,034; 19,862.
  - Greece: 23,448; 19,377.
  - Italy: 224,804; 185,778.
  - Portugal: 37,093; 30,653.
  - Switzerland: 217,808; 216,370.
  - Source: IMF Financial Soundness Indicators; Central Bank of Ireland; Haver Analytics; and IMF staff calculation.

### Appendix V: Simulation results (summary indicators)
- Simulation experiments report indicators such as:
  - Induced Failures.
  - Contagion Rounds.
  - Absolute Hazard 1/ (Number of simulations in which that the banking system in each country fails).
  - Hazard Rate 2/ (Percentage of failures out of 17 rounds of the simulation).
  - Index of Contagion 3/ (Percentage of total capital impairment in other banking systems due to the failure of the banking system in each country).
  - Index of Vulnerability 4/ (Average of capital impairment of the banking system in each country due to the failure of all other banking systems).
- Selected country-level examples (from reported tables):
  - Ireland: entries include "00425541" in one table and "00425842" in another (as tabulated under Induced Failures / related columns).

### Appendix VI: Market data-based interconnectedness analysis — technical details
- Methodology:
  - VAR specification: Y_t = A(L) Y_t + ε_t, with A(L) lag polynomial order chosen by Bayesian Information Criterion (BIC).
  - Generalized forecast-error variance decomposition using Pesaran and Shin (1998) to compute percent contribution of entity A to H-step ahead forecast error variance of entity B.
  - Complemented with Granger-causality tests (Wald test) at the five percent significance level to identify direct connections beyond common exposures.
- Data:
  - Analyses conducted at the sectoral level; two worldwide indexes used as global factors: world real estate MSCI index and the global hedge fund index.
  - Countries covered include: Ireland, China, France, Germany, Greece, Italy, the Netherlands, Portugal, Spain, the U.K and the U.S.
  - For Ireland a stock price index for the financial system as a whole (ISEQ financial index) was used.
  - Equity returns data from Bloomberg spanning January 1, 2000 to February 4, 2016 at the daily frequency.
  - Forecast error variance calculated on a 10 days ahead basis.

*Source: _cr16315 - 3. Tail shocks (INTERNATIONAL MONETARY FUND).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16315.pdf_