## _wp13178

## Source details

**Canonical URL:** [_wp13178](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13178.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13178.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13178.pdf.json)

---

### Purpose and scope of the chapter
- Objective:
  - Formulate guidelines for designing a crisis solvency stress test, focusing on solvency risk.
  - Distinguish effective crisis stress tests using financial market impact studies of exercises in the United States, European Union, Ireland and Spain.
  - Use case study analysis and juxtapose findings with relevant “best practice” principles (e.g., BCBS, 2009; Board of Governors of the Federal Reserve/FDIC/OCC, 2012; IMF, 2012a).
- Case studies covered:
  - Seven crisis stress tests conducted in the United States, the European Union, Ireland and Spain between 2009 and 2012 (Table 1).
- Market proxies used (all sourced from Bloomberg):
  - Financials stock price indices and sovereign CDS spreads (Table 2).

### Key introductory observations and principal conclusions
- Stress tests described as the “new normal” in financial crisis management and used to regain public trust in banking systems.
- Crisis stress test definition:
  - A supervisory exercise accompanied by detailed public disclosure, with microprudential execution but macroprudential objectives (Figure 1).
- Principal conclusions:
  - Crisis stress tests must be credible, timely, and have clearly specified objectives.
  - Authorities must be fully committed and ready to take follow-up actions with necessary resources and backstops.
  - Supporting activities such as AQRs and follow-up stress tests are necessary for credibility.
  - Design choices (timing, governance, disclosure, coverage, scenario design, AQR) materially affect effectiveness.

### Historical evidence and market impact (selected findings)
- United States — SCAP/CCAR outcomes:
  - SCAP release halted and reversed a 2-year slide in investor confidence toward U.S. banks (Figure 2).
  - Capital increases for SCAP sample: weighted Tier 1 (T1) Common Equity ratio of 18 BHCs in the SCAP sample increased from an average 5.6 percent at end-2008 to 11.3 percent in 2012 Q4.
  - T1 Common Equity increased from $393 billion to $792 billion over the same period.
  - SCAP disclosed estimated losses and capital needs of USD599.2 (Billions).
- European Union — CEBS/EBA outcomes:
  - CEBS 2009, CEBS 2010 and EBA 2011 exercises were less convincing in stabilizing markets; stock prices weakened and sovereign CDS spreads widened following each exercise in several instances.
  - CEBS 2010 disclosed EUR565.8 (Billions); EBA 2011 disclosed EUR400.0.
- Ireland — PCAR 2011:
  - PCAR 2011 contributed to stabilizing market sentiment; recapitalization/restructuring plans incorporated PCAR outcomes.
- Spain — TD/BU 2012:
  - Third-party BU stress test and a comprehensive restructuring strategy put a floor under market sentiment.
  - Private sector stress tests applied thresholds and assumptions that produced large capital holes (observed examples: CT1 thresholds up to 11 percent, capital hole of up to €120 billion).

### Credibility, governance, and independence
- Credibility requirements for crisis stress tests:
  - Governance must be perceived as independent and technically expert (stress tester and overseer).
  - Tests should be sufficiently stringent yet plausible, simultaneous, consistent, and comparable across firms.
  - Results should be granular to enable differentiation among institutions.
  - Clarify early how results will be addressed, backstopped, and used.
- Governance examples:
  - United States: overseen and executed by Fed, FDIC, OCC; supervisors consulted but not conducting analyses.
  - EU: national supervisors conducted tests; coordinated/overseen by CEBS/EBA; EBA sought more legal powers over input-data reliability.
  - Ireland and Spain: third-party contractors (e.g., BlackRock, Oliver Wyman, Roland Berger) appointed to strengthen perceived independence.

### Timing, scope, and coverage
- Timing trade-offs and empirical observations:
  - Timing is crucial: earlier action can reduce asset-value destruction and fiscal costs; delayed action can raise reputational and fiscal costs and necessitate broader and more expensive remedial actions.
  - Examples of distance from historical peaks to exercise:
    - United States: index peak 509.6 (Feb 20, 2007); end prior 168.8 (Dec 31, 2008); Change from peak (percent): -67.9.
    - Europe: index peak 538.8 (Apr 20, 2007); end prior 151.1 (Dec 31, 2008); Change from peak (percent): -72.0.
    - Ireland: index peak 17,951.5 (Feb 21, 2007); end prior 414.3 (Dec 31, 2010); Change from peak (percent): -97.7.
    - Spain: index peak 158.8 (Feb 14, 2007); end prior 62.1 (Dec 31, 2011); Change from peak (percent): -60.9.
  - Sovereign CDS spread changes (basis points):
    - United States: narrowest 5.8 (Apr 29, 2008); end prior 67.4 (Dec 31, 2008); Change from narrowest +61.6 (+1,062.1 percent).
    - Ireland: narrowest 16.1 (Mar 24, 2008); end prior 608.7 (Dec 31, 2010); Change from narrowest +592.6 (+3,780.7 percent).
    - Spain: narrowest 1.5 (Jun 20, 2005); end prior 380.4 (Dec 31, 2011); Change from narrowest +378.9 (+25,260.0 percent).
- Coverage principles:
  - Capture systemically important banks and majority of banking system assets; sample may expand with crisis severity.
  - Empirical coverage:
    - United States SCAP: 19 largest BHCs (assets > $100 billion) representing two-thirds of banking system assets.
    - CEBS 2009: 22 large cross-border banks with 60 percent of EU banking assets; later rounds increased to 91 banks covering at least 50 percent of each national sector.
    - Ireland PCAR 2011: four institutions accounting for 80 percent of banking system assets.
    - Spain TD/BU 2012: banks accounting for around 90 percent of total system assets.

### Scenario design, risk-factor coverage, and modeling
- Scenario calibration approaches:
  - CEBS method: calibrate growth shocks in number of standard deviations of 2-year growth rates over preceding 30 years (steps 1–4 specified).
  - Rule of thumb: IMF FSAP scenario stress tests apply two standard deviation shocks to growth.
- Observations on extremity and parameterization:
  - Magnitude of macro shocks not sole determinant of credibility:
    - CEBS 2009/2010 and EBA 2011 applied cumulative growth shocks averaging 1.9, 1.3 and 1.5 standard deviations, respectively.
    - SCAP adverse growth scenario equivalent to cumulative one standard deviation from baseline over two years.
    - Spain FSAP severe adverse: adverse shock exceeded 3 standard deviations of mean annual growth rate of past 30 years.
  - Number of macroeconomic parameters not decisive for credibility:
    - SCAP used 3 parameters (real GDP growth, unemployment, house prices); Ireland and Spain used more than a dozen.
- Risk-factor coverage:
  - Credit risk: U.S. tested 12 categories of loans; other exercises ranged from 5–8 categories plus foreclosed assets and segments of foreclosed assets (Spain: 6 categories of loans plus 5 segments of foreclosed assets).
  - Market risk: trading book and banking-book treatment varied; EU exercises often limited sovereign stress to trading book AfS adjustments; Spain and IMF exercises treated sovereign exposures differently, with LTRO presence affecting treatment.
- Revenue and loss modeling:
  - Revenue projections (pre-provision profits) are crucial and seen as a “black box”—credibility of these estimates key to perception of the exercise.

### Capital definitions, hurdle rates, and trade-offs
- Jurisdictional practices and hurdle rates (verbatim):
  - United States:
    - Applied existing capital framework.
    - Required T1 capital hurdle of 6 percent post-stress and T1 Common Equity ratio of 4 percent post-stress.
    - Basel I risk weights used to calculate RWA.
  - EU, Ireland, Spain:
    - Applied CRD II/Basel II risk weights and capital definitions prevailing at the time.
    - CEBS 2009 and 2010: T1 hurdle rate of 6 percent.
    - EBA 2011: EBA Core Tier 1 (CT1) post-stress hurdle rate of 5 percent.
    - Ireland: baseline hurdle 10.5 percent and 6 percent EBA CT1 under stress.
    - Spain TD/BU: EBA CT1 hurdle rate 9 percent under baseline and 6 percent under adverse scenario.
- Box 2 illustrative implications of hurdle rates (verbatim example):
  - Bank pre-shock CT1 ratio: 9 percent; constant RWA.
  - Required CT1 post-stress hurdle rate: 11 percent (includes a buffer).
  - Baseline scenario:
    - (a) Under baseline, CT1 reduced by 2 percentage points to 7 percent.
    - (b) Bank expected to raise CT1 to 11 percent: increase of 4 percentage points.
    - (c) Bank would need to top up existing 9 percent CT1 by another 4 percentage points to 13 percent in anticipation.
    - (d) Implies total capital adequacy ratio of more than 16 percent once additional requirements to make up T1 and total capital are included.
    - (e) If central case materializes, bank would have CT1 of targeted 11 percent.
  - Adverse scenario:
    - (i) Under severe adverse, losses reduce CT1 by 6 percentage points to 3 percent.
    - (ii) Bank expected to raise CT1 back to 11 percent: increase of 8 percentage points.
    - (iii) Bank would need CT1 of 17 percent (existing 9 percent plus another 8 percentage points).
    - (iv) Implies total capital adequacy ratio of more than 20 percent once additional requirements are included.
    - (v) If baseline materializes, bank would carry CT1 of 15 percent (17 percent less the 2 percentage points impact).
- Policy trade-off:
  - Very high post-stress capital ratio requirements can exacerbate deleveraging and credit contraction (microprudential vs macroprudential tension).
  - Recommended practice: build buffers in good times so they can be drawn down in stress, and align capital amount versus ratio objectives to avoid excessive contraction.

### Transparency, disclosure, AQRs, and follow-up
- Disclosure best practices observed in effective exercises:
  - Publish stress test design, methodology, implementation, macroeconomic and loan loss assumptions, and bank-level results showing projected losses by portfolio category, capital components, and projected capital shortfalls (Table 8 summary).
  - SCAP’s public disclosure of estimated losses and capital needs, combined with a financial backstop, reassured markets.
- AQR role and cost trade-offs:
  - An AQR should precede stress tests when data quality is suspect.
  - AQRs are more costly but reduce the risk of misleading stress-test outcomes; they introduce a “double hit” (provisioning first, then capital shortfall), increasing eventual recapitalization needs.
  - Numerical illustration (verbatim):
    - Assumption: stress test shows total capital needs of €50 billion (constant RWA).
    - Example 1 — No AQR:
      - Revenues fall €10 billion; expected loan losses €20 billion; net profit decreases from €40 billion to €10 billion.
      - Projected capital shortfall: €10 billion addressed by a direct cash injection.
    - Example 2 — AQR first:
      - AQR uncovers additional provisioning of €30 billion; stress test projects further loan losses €20 billion and revenue drop €10 billion.
      - Result: net loss of €20 billion and projected capital shortfall of €40 billion.
  - Figure 8 starting balance-sheet values (verbatim): Various revenues: 150; Various expenses: 110; Loans: 240; Various liabilities: 170; Provision: 0; Less provisions: 20; Profit: 40; Loans net of provisions: 220; Shares: 20; Cash: 20; Retained earnings: 10.
- Follow-up stress tests and evolution:
  - U.S. path: SCAP followed by CCAR supervisory tests and DFA-related publication; CCAR 2011 initially confidential; CCAR 2012/2013 and DFA 2013 published in detail.
  - Scenario severity examples:
    - CCAR 2011 adverse projected growth shock: 1.4 standard deviations from projected baseline.
    - CCAR 2012 and CCAR 2013 adverse: 2.5 standard deviations.
  - EBA planned EU stress tests for 2014; Ireland and Spain adjusted sequencing with EBA/SSM AQRs and balance-sheet assessments.

### Design scorecard and operational best practice (selected elements)
- Essential design elements (scorecard themes):
  - Timing: early enough to arrest loss of confidence but linked to capacity to act.
  - Oversight and stress tester independence: perceived independence critical; third-party contractors used where supervisory credibility damaged.
  - Approach: bottom-up cross-validated by top-down preferred; if TD only, perform bank-by-bank TD.
  - Coverage: aim to include systemically important banks and a large share of system assets (examples: 2/3, 60–90 percent depending on jurisdiction).
  - Scenario severity: capture extreme but plausible shocks (avoid “tail of the tail”); standardize calibration (e.g., CEBS method) for comparability.
  - Standardized assumptions: macro scenarios standardized across banks; behavioral assumptions can vary with supervisory cross-checks.
  - Capital standards and hurdle rates: clear definition and communication required (Basel/CRD frames applied variably across jurisdictions).
  - Disclosure: publish design, methodology, assumptions, bank-level results where feasible.
  - AQR and DIV: recommend preceding stress tests when data quality suspect.
  - Liquidity stress testing: increasingly important; complement solvency exercises where feasible.
  - Follow-up actions and financial backstops: predefined action plans and credible backstops (e.g., SCAP/TARP CAP) enhance market acceptance.
- Appendix Table 1 (verbatim effectiveness entries):
  - Effectiveness as defined: financials stock index stabilizes/improves and returns volatility falls.
  - Outcomes summarized verbatim:
    - United States: Yes.
    - European Union (CEBS 2009): No.
    - European Union (CEBS 2010): No.
    - European Union (EBA 2011): No.
    - Republic of Ireland (PCAR 2011): Yes.
    - IMF FSAP 2012 TD 2012 Exercise: Not applicable.
    - IMF FSAP 2012 BU 2012 Exercise: No.
    - Spain: Yes.

### Policy recommendations and operational implications (verbatim guidance and synthesis)
- Authorities undertaking crisis stress tests should:
  - Be fully committed and prepared to take action once valuations have fallen to certain levels.
  - Clarify objectives, design, timelines and follow-up actions at the outset.
  - Ensure credible financial backstops or contingency financing arrangements are available and communicated.
  - Precede stress tests with AQR/DIV when data reliability is suspect, recognizing the “double hit” provision-versus-capital trade-off.
  - Standardize macro scenarios for comparability; allow reasonable variation in behavioral assumptions subject to supervisory challenge.
  - Publish sufficient technical detail (design, methodology, assumptions, bank-level summaries) to enable market replication and restore confidence, while balancing operational risks of disclosure.
  - Complement solvency tests with liquidity stress testing where appropriate.
  - Use independent overseers or third-party stress testers where supervisory credibility is weak.
- Caveat noted:
  - Political economy considerations and luck (timing, actual bank health, market conditions) materially affect outcomes; without commitment and resources, stress tests can backfire.

*Source: _wp13178 - References and selected chapter excerpts (provided content).*

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

### References

### Document components and listings
- Tables included (by table number and title):  
  - 1. Case Studies: Crisis Stress Tests  
  - 2. Market Data: Financials Stock Price Index and CDS Spreads  
  - 3. Crisis (and Follow-up) Stress Tests: Effectiveness Scorecard  
  - 4. Crisis Stress Tests: Design Scorecard  
  - 5. Crisis Stress Test Jurisdictions: Financial Markets Statistics  
  - 6. Crisis Stress Tests: Macro-financial Parameters Scorecard  
  - 7. Crisis Stress Tests: Risk Factors Scorecard  
  - 8. Crisis Stress Tests: Disclosure Scorecard  
  - 9. European Union and the United States: Evolution of Publicized Stress Testing Exercises

- Figures included (by figure number and title):  
  - 1. Solvency Stress Testing Applications  
  - 2. United States: The Sentiment after the SCAP  
  - 3. European Union: The Ebb from the EBA  
  - 4. Ireland: The Pain before the PCAR  
  - 5. Spain: The Floor under the FSAP  
  - 6. United States: S&P 500 Stock Market and Financial Sector Indices  
  - 7. United States: Baseline and Adverse Growth Scenarios for Crisis and Supervisory Stress Tests  
  - 8. Accounting Entries: Loss Recognition, Provisioning and Recapitalization

- Boxes included:  
  - 1. Designing Crisis Stress Test Growth Scenarios  
  - 2. The Potential Impact of Capital Hurdle Rates for Crisis Stress Tests

- Appendices included:  
  - I. Case Studies of Crisis Solvency Stress Tests: United States, European Union, Ireland and Spain  
  - II. The Cost of Direct Recapitalization versus Loss Recognition and Provisioning

- Appendix table:  
  - 1. Crisis Stress Tests: Features of Design

### Key introductory observations
- Stress tests are characterized as the “new normal” in financial crisis management and are used to regain public trust in banking systems during the global financial crisis.
- A “crisis stress test” is described as a supervisory exercise accompanied by detailed public disclosure, serving as a microprudential exercise with macroprudential objectives (Figure 1).
- Transparency and the quality of disclosure are emphasized as critical to credibility.

### Historical context and examples
- The concept was introduced by U.S. authorities in early-2009 as the Supervisory Capital Assessment Program (SCAP).
- Market reaction and subsequent actions after SCAP:  
  - The announcement initially produced trepidation and skepticism; official clarifications, availability of a financial backstop, and publication of methodology and results reassured markets (Peristian and others, 2010).  
  - Assessed banks were able to add more than $200 billion in common equity in the following 12 months.  
  - U.S. supervisors followed SCAP with the Comprehensive Capital Assessment Program (CCAR) and under the Dodd-Frank Wall Street Reform and Consumer Protection Act (DFA) framework.

- Turkish authorities’ post-2001 banking sector assessment is cited as a public diagnostic exercise with no forward-looking stress test but with published objectives, methods, implementation details, and progress (BRSA, 2002a; BRSA, 2002b).

### Purpose, design features, and governance of crisis stress tests
- Purposes of crisis stress tests (as distinct from supervisory stress tests in crisis):  
  - (i) determine a needed capital buffer over current solvency levels;  
  - (ii) differentiate the soundness of banks in the system as part of triage analysis;  
  - (iii) quantify potential fiscal costs depending on projected shortfalls and urgency of recapitalization.

- Credibility requirements:  
  - Governance must be perceived as independent and technically expert (stress tester and overseer).  
  - Tests must be sufficiently stringent yet plausible, simultaneous, consistent, and comparable cross-firm assessments.  
  - Results should be granular to enable differentiation among institutions.  
  - Clarify early how results will be addressed, backstopped, and used.

- Recommended diagnostic strategy components:  
  - (i) containment; (ii) diagnostics (asset quality review (AQR), data integrity and verification (DIV), stress test); and (iii) restructuring or exit.  
  - AQR and DIV should precede stress tests so that data used are “clean.”

- Political economy considerations may affect design given implications for public confidence and fiscal costs.

### Data, case studies, and methodology
- Case studies: four case studies covering seven crisis stress tests conducted in the United States, the European Union, Ireland and Spain between 2009 and 2012 (Table 1).
- Source authorities for exercise details: Board of Governors of the Federal Reserve (Fed), European Banking Association (EBA) and CEBS, Central Bank of Ireland (CBI), Banco de España (BdE).
- Market indicators used in the analysis (all sourced from Bloomberg, Table 2):  
  - Financials stock price indices as proxies for market assessment of banking-system soundness.  
  - Sovereign CDS spreads to indicate perceived country creditworthiness and potential fiscal burden.

- Caveat: market indicators may be influenced by concurrent events not isolated in this study.

### Identification of successful crisis stress tests (impact study methodology)
- Impact study criteria (a stress test is classified as successful if it stabilizes or improves investor sentiment for at least six months after results announcement):  
  - The financials stock index does not fall below the threshold set at the time the test results are announced.  
  - The volatility of daily returns (standard deviation over 60 days) continues to decline.  
  - The sovereign CDS spread narrows.

- Empirical evidence for the U.S. SCAP (Figure 2):  
  - SCAP release halted and reversed a 2-year slide in investor confidence toward U.S. banks.  
  - Market volatility, which had peaked just prior to the exercise, declined sharply.  
  - Since SCAP, the S&P 500 Financial Sector Index largely remained above the level established by SCAP results, though it “flirted with that floor” during 2012 Q3 volatility.  
  - Capital increases for U.S. firms in the SCAP sample: the weighted Tier 1 (T1) Common Equity ratio of the 18 bank holding companies in the SCAP sample increased from an average 5.6 percent at the end of 2008 to 11.3 percent in 2012 Q4, reflecting an increase in T1 Common Equity from $393 billion to $792 billion over the same period.  
  - U.S. CDS spreads narrowed in tandem with financials index improvement during SCAP but subsequently dissociated in September 2011 as markets focused on U.S. fiscal deficit developments after the CBO announcement.

### Study objectives and analytical approach
- Specific objective: formulate guidelines for designing a crisis solvency stress test, focusing on solvency risk.  
- Methodologies employed:  
  - Distinguish effective crisis stress tests using financial market impact studies of exercises in the United States, European Union, Ireland and Spain.  
  - Case study analysis to identify key elements and design choices, drawing on qualitative information.  
  - Juxtaposition against relevant “best practice” principles where relevant (e.g., BCBS, 2009; Board of Governors of the Federal Reserve/FDIC/OCC, 2012; IMF, 2012a).

### Principal conclusion highlights
- Crisis stress tests must be credible, timely, and have clearly specified objectives.  
- Authorities must be fully committed, prepared to conduct thorough, honest, and transparent examinations, and ready to take follow-up actions with necessary resources and backstops.  
- Supporting activities such as AQRs and possibly follow-up stress tests are necessary for credibility.  
- The design of crisis stress tests matters significantly and can explain differing effectiveness between U.S. and European exercises.

*Source: _wp13178 - References (selected content extracted from the provided PDF content).*

### 2.   Fund staff had previously conducted a rudimentary stress test of the Hungarian banking system during the crisis pro

### 2.   Fund staff had previously conducted a rudimentary stress test of the Hungarian banking system during the crisis pro

### Case studies and market-data context
- Table 1 and Table 2 summarize crisis stress-test case studies and market-data proxies (Stock Price Index and CDS Spreads).  
- Jurisdictions and notable exercises listed include:
  - United States: Supervisory Capital Assessment and Program 2009 (Fed, FDIC, OCC).
  - European Union: CEBS 2009, CEBS 2010, EBA 2011 (national supervisors, CEBS/EBA, EC, ECB, ESRB).
  - Ireland: Prudential Capital Assessment and Review (PCAR) 2011 (Central Bank of Ireland; loan-loss inputs from BlackRock Solutions).
  - Spain: Top-down (TD) 2012 and Bottom-up (BU) 2012 (Oliver Wyman; Roland Berger; BdE; MEC; Troika; EBA involvement).
- Market proxies used:
  - United States: S&P 500 Financials Sector Index (S5FINL [Index]); United States EUR senior 5-year CDS (ZCTO CDS EUR SR 5Y [Corp]).
  - European Union: STOXX Europe 600 Banks Price EUR (SX7P [Index]); iTraxx SovX Western Europe USD 5-year (SOVWE CDSI GENERIC 5Y [Corp]).
  - Ireland: Irish Stock Exchange Financial Index (ISEF [Index]); Ireland USD senior 5-year CDS (IRELND CDS USD SR 5Y [Corp]).
  - Spain: MSCI Spain Financials Index (MSES0FN [Index]); Spain USD senior 5-year CDS (SPAIN CDS USD SR 5Y [Corp]).

### Market reactions and effectiveness of crisis stress tests
- United States (SCAP/SCAP-related actions):
  - The turnaround in confidence in U.S. banks from the SCAP buoyed sentiment globally; U.S. financials stock index rebounded and volatility fell.
  - The SCAP results helped halt and turn around the sharp slide in market confidence despite the SCAP adverse growth scenario being relatively mild (equivalent to a cumulative one standard deviation from baseline over a two-year horizon).
- European Union:
  - EU stress tests were less convincing: stock prices weakened and sovereign CDS spreads widened following each exercise; the stock index has not risen above the level recorded at CEBS 2009.
  - CEBS 2010: Ireland requested a bailout from the EC, ECB and IMF after stress tests indicated banks would remain sufficiently capitalized.
  - EBA 2011: Dexia and Bankia passed the EBA 2011 stress test only to require significant restructuring within months; stock market volatility jumped.
  - EU Capital Exercise (announced Oct 2011; results disclosed Dec 2011) plus ECB LTRO (Dec 2011) halted deterioration in confidence and narrowed sovereign CDS spreads; OMT announcement (Aug 2012) further improved market confidence.
- Ireland:
  - PCAR 2011 contributed to stabilizing market sentiment. IMF’s Third Review in Sep 2011 indicated program benchmarks largely met and PCAR outcomes incorporated into recapitalization/restructuring plans.
  - Financials stock price index troughed; volatility declined; sovereign CDS spreads tightened.
- Spain:
  - Third-party BU stress test and a comprehensive strategy to identify and deal with problem banks put a floor on market sentiment and turned it around.
  - Announcement of IMF FSAP and third-party TD stress test results coincided with increased volatility but signaled progress toward restructuring.
  - Publication of Memorandum of Understanding (MoU) with Eurogroup in July 2012 reassured investors; stock price volatility declined sharply and sovereign CDS spreads narrowed.
- Table 3: an effectiveness scorecard compares indicators (financials stock index, return volatility, CDS spreads) across crisis/surveillance/supervisory exercises (SCAP 2009, CCAR series, CEBS/EBA exercises, PCAR 2011, FSAP 2012, Spain TD/BU 2012).

### Designing an effective crisis stress test — key elements and lessons
- General principle:
  - Crisis stress tests require different design elements from normal supervisory tests (timing, governance, transparency, objective, action plan, financial backstop).
  - Additional supporting activities include AQRs, separate liquidity stress tests, and follow-up solvency stress tests.
  - A design “scorecard” (Table 4) compares features across crisis stress tests; Appendix I provides details.
- Timing:
  - Timing is crucial: information provision steps should be taken as soon as possible during a crisis to reduce uncertainty (Borio and others, 2010; Pritsker, 2010).
  - Early recognition and intervention can avoid hidden deterioration and lower costs of resolution.
  - Empirical observations:
    - U.S. SCAP was introduced after a decline from the Feb 20, 2007 peak; CEBS 2009 was introduced almost 2 years after the STOXX Europe 600 Banks Price Index peak.
    - Ireland and Spain conducted stress tests 4 and 5½ years after their financials stock-price peaks.
    - If decisions to stress test were made around the end of the year prior to each crisis stress test, U.S. and European indices would have dropped by between 65–75 percent by that stage.
  - Market indicators as possible triggers: stock prices, price-to-book (PB) ratios, sovereign CDS spreads — evidence inconclusive.
    - Examples from Table 5:
      - Financials stock index historical peaks and end-of-year-prior-to-first-crisis-stress-test levels:
        - United States: 509.6 (Feb 20, 2007); 168.8 (Dec 31, 2008); Change from peak (percent): -67.9.
        - Europe: 538.8 (Apr 20, 2007); 151.1 (Dec 31, 2008); Change from peak (percent): -72.0.
        - Ireland: 17,951.5 (Feb 21, 2007); 414.3 (Dec 31, 2010); Change from peak (percent): -97.7.
        - Spain: 158.8 (Feb 14, 2007); 62.1 (Dec 31, 2011); Change from peak (percent): -60.9.
      - Price-to-book ratios (historical peaks and end-of-year-prior):
        - United States: peak 3.50 (Sep 12, 2000); end prior 0.99 (Dec 31, 2008); Change (percent): -71.7.
        - Europe: peak 2.21 (May 15, 2002); end prior 0.73 (Dec 31, 2008); Change (percent): -67.0.
        - Ireland: peak 4.16 (Jan 1, 1999); end prior 0.26 (Dec 31, 2010); Change (percent): -93.8.
        - Spain: peak 4.74 (Jul 17, 1998); end prior 0.72 (Dec 31, 2011); Change (percent): -84.8.
      - Sovereign CDS spreads (basis points):
        - United States narrowest 5.8 (Apr 29, 2008); end prior 67.4 (Dec 31, 2008); Change from narrowest +61.6 (+1,062.1 percent).
        - Ireland narrowest 16.1 (Mar 24, 2008); end prior 608.7 (Dec 31, 2010); Change from narrowest +592.6 (+3,780.7 percent).
        - Spain narrowest 1.5 (Jun 20, 2005); end prior 380.4 (Dec 31, 2011); Change from narrowest +378.9 (+25,260.0 percent).
  - Timing trade-offs:
    - Delay increases asset-value destruction, reputational damage to supervisors, and fiscal burdens from loss of market/depositor/creditor confidence.
    - Late action can require third-party stress testers, independent overseers, more comprehensive AQRs/DIVs, broader bank coverage, and higher market-imposed standards.
  - Key judgment: optimal timing is not clear-cut; crisis stress tests ideally before confidence becomes entrenched but must be tied to authorities’ capacity to backstop and address findings.
- Governance:
  - Credibility of protagonists is paramount; expertise, resources, and political-economy considerations matter.
  - Examples:
    - United States: oversight/execution via Fed, FDIC, OCC; supervisors of individual banks consulted but not doing the analyses.
    - EU: national supervisors conducted tests, coordinated/overseen by EBA (which sought more legal powers over input-data reliability).
    - Ireland and Spain: third-party contractors appointed to strengthen perceived independence and objectivity given damaged supervisory reputations.
- Scope and coverage:
  - Preferred approach: bottom-up (BU) test cross-validated by top-down (TD); if only TD is possible, perform bank-by-bank TD rather than aggregated TD.
  - Support from AQRs and DIVs recommended.
  - Coverage principle: capture at least the systemically important banks and the majority of banking system assets; sample may expand depending on crisis severity.
  - Empirical coverage:
    - U.S. SCAP: 19 largest BHCs (assets > $100 billion), representing two-thirds of banking system assets.
    - EU CEBS 2009: 22 large cross-border banks with 60 percent of EU banking assets; later increased to 91 banks covering 21 countries and at least 50 percent of each banking sector (additional 5 percentage points coverage).
    - Ireland PCAR 2011: four financial institutions accounting for 80 percent of banking system assets.
    - Spain TD and BU 2012: banks accounting for around 90 percent of total system assets.
- Scenario design and risk-factor coverage:
  - Crisis stress scenarios should capture extreme but plausible shocks (tail risks) and reflect uncertainty around baselines; avoid “tail of the tail” shocks that prevent constructive follow-up actions.
  - Magnitude of macroeconomic shocks not the overriding success factor:
    - CEBS 2009/2010 and EBA 2011 applied cumulative growth shocks averaging 1.9, 1.3 and 1.5 standard deviations, respectively.
    - SCAP adverse growth scenario was equivalent to a cumulative one standard deviation from baseline over two years — less stressful than CEBS/EBA shocks — yet SCAP was effective in regaining market confidence.
    - Spain TD and BU growth shocks were equivalent to one standard deviation from projected baseline; Ireland used EBA scenarios in PCAR 2011.
  - Number of macroeconomic parameters does not clearly determine credibility:
    - SCAP used three parameters (real GDP growth, unemployment, house prices); Ireland and Spain used more than a dozen.
  - Comprehensive coverage of material risk factors is crucial:
    - U.S. and EU stress tests focused on domestic loan books.
    - EU stress tests criticized for inadequate capture of sovereign exposures and for lenient treatment of sovereign holdings (trading book market-value adjustments only; omission of banking-book sovereign stress).
    - Spain addressed lender forbearance and misclassification concerns in BU exercise via auditors and real-estate appraisers; sovereign risk omission considered less problematic given LTRO availability.
- Transparency, disclosure, and follow-up:
  - Design scorecard (Table 4) assesses features: timing, oversight, stress tester independence, approach (BU vs TD), coverage, scenario severity, standardization of assumptions, hurdle rates (CET > 6 percent noted as “very high”), objective/action plan clarity, follow-up actions, explicit financial backstops, disclosure of design/methodology/models/assumptions, and bank-by-bank result disclosure.
  - AQRs and follow-up stress tests:
    - AQRs are important inputs; lower-intensity quantitative substitutes used in some exercises.
    - Follow-up stress tests and standardized assumptions across banks help ensure robustness.
  - Liquidity stress testing:
    - Liquidity stress tests accompanying solvency exercises were variably applied; EBA conducted a confidential thematic review of liquidity funding risks.
- Practical implications:
  - Crisis stress tests should be timed early but not at the expense of other credibility elements (independence, transparency, coverage, comprehensive risk-factor treatment, AQR support, follow-up actions and financial backstops).
  - Where supervisory credibility is damaged, third-party consultants and explicit independent oversight can restore confidence but may reflect a higher cost of delayed action.

*Source: IMF staff chapter on crisis stress-testing experience and design (excerpt).*

### Box 1. Designing Crisis Stress Test Growth Scenarios

### Box 1. Designing Crisis Stress Test Growth Scenarios

### Standardized calibration approach (CEBS method)
- Rationale:
  - Calibrate growth shocks in terms of the number of standard deviations from the baseline to allow standardized comparison across stress tests at a point in time and over time.
  - A large standardized shock may be less significant for an economy with historically large and volatile growth than for an economy with moderate and stable growth.
- CEBS method steps:
  - 1. Calculate the 2 year growth rates over the preceding 30 years.
  - 2. Calculate the standard deviation of the 2-year growth rates over the 30 year period.
  - 3. Calculate the desired number of standard deviations of the 2-year growth rate.
  - 4. Apportion the standard deviation(s) growth over the 2-year horizon and subtract from each year of the baseline forecast to derive the stressed scenario.

### Typical calibrations and crisis-specific adjustments
- Rule of thumb:
  - IMF’s FSAP scenario stress tests: apply two standard deviation shocks to growth (IMF, 2012a; Jobst and others, 2013).
- Crisis adjustments and examples:
  - Spain FSAP stress test: imposed a “severe adverse” scenario of one standard deviation from the baseline GDP growth trend over a two-year risk horizon (IMF, 2012b).
    - The shock was applied on top of a downward adjustment to the World Economic Outlook baseline forecast to incorporate downside risks plus a projected fiscal adjustment.
    - Most of the shock to baseline growth (about two-thirds) was assumed to occur in the first year, attributable to a sharp decline in output, further declines in house prices and rising unemployment.
    - The 2-year cumulative GDP shock for Spain under the severe adverse scenario was considered extreme by historical standards; the adverse shock scenario extended beyond 3 standard deviations of the mean annual growth rate of the past 30 years and exceeded the sharp contraction experienced in 2009.
    - The third-party crisis stress tests subsequently increased the second-year stress and extended both scenarios to a third year.
    - Actual outcome: growth in 2012—the first year of the risk horizon—approximated the projected baseline.
  - United States examples (chart referenced): SCAP 2009, CCAR 2011, CCAR 2012, CCAR & DFA 2013 scenarios and actual quarterly annualized growth profiles were analyzed (quarter-on-quarter annualized).

### Gauging extremity of scenarios
- Corroborating metric:
  - Determine deviation from the long-term historical average in standard deviation terms to assess extremity.
  - Spain example: adverse shock > 3 standard deviations of the mean annual growth rate of the past 30 years; designed to be more protracted than 2009 contraction.

### Standardization of assumptions and governance
- Standardization objectives:
  - Crisis stress tests standardize assumptions to facilitate comparisons across banks within a jurisdiction; absolute standardization of behavioral assumptions is not necessary for credibility.
  - Behavioral assumptions (management-controlled) are typically allowed to vary but subject to cross-checks to ensure reasonableness.
  - Publication of assumptions is important so market participants can replicate results.
- Institutional practices (selected examples):
  - SCAP (U.S.):
    - U.S. supervisors provided macroeconomic scenario assumptions.
    - Banks adapted assumptions to reflect specific business activities; supervisors reviewed firms’ submissions and quantitative methods (Fed, 2009a and 2009b).
    - Supervisors applied independent quantitative methods to firm-specific data to facilitate horizontal comparisons.
  - EU CEBS/EBA:
    - Applied macroeconomic and sovereign shock scenarios and parameters developed by the ECB.
    - EBA 2011 provided very detailed and prescriptive guidance on assumptions and methodologies; national supervisors reviewed banks’ calculations, then EBA conducted in-depth consistency checks.
  - Ireland PCAR 2011:
    - Incorporated many EBA 2011 parameters.
    - A private consultancy was contracted by the CBI to provide oversight and to challenge third-party stress tester work to ensure consistency across institutions and portfolios (CBI, 2011).
  - Spain:
    - Growth scenarios and guidelines provided by a Steering Committee comprising the authorities, the Troika and counterparts from two European central banks.
    - BU exercise process and methodology were monitored and agreed with an Expert Coordination Committee from the Troika, the EBA and the authorities (Oliver Wyman, 2012b).

### Modeling losses and revenues
- Loss modeling:
  - Crisis highlighted importance of adequately modeling losses for different categories of credit risk (various types of real estate, corporate sector, credit cards), geographic heterogeneity and a rapidly evolving macro-financial environment with no precedent.
- Revenue modeling:
  - Stress testing revenues—especially under stressed conditions—is seen as a “black box” (Schuermann, 2012).
  - Projected pre-provision profits are crucial in determining banks’ loss absorption capacity; credibility of these estimates is key to perception of the stress test.

### Capital standards: role, variability, and trade-offs
- General principle:
  - Capital standards applied to crisis stress tests are crucial for legitimacy; some variability across countries is acceptable if capital definitions are made clear.
  - Composition of capital is important, consistent with Basel III.
- Jurisdictional practices and hurdle rates:
  - United States:
    - Applied existing capital framework.
    - Banks required to meet the T1 capital hurdle of 6 percent post-stress and the higher quality T1 Common Equity ratio of 4 percent post-stress.
    - Basel I risk weights used to calculate RWA.
  - EU, Ireland, Spain:
    - Applied existing Capital Requirement Directive at the time (CRD II) with Basel II risk weights to calculate RWA; capital definition deviated from regulatory directive.
    - CEBS 2009 and 2010: applied a T1 hurdle rate of 6 percent.
    - EBA 2011: implemented commonly-agreed definition of common equity capital (“EBA Core Tier 1 (CT1)”) and applied a post-stress hurdle rate of 5 percent.
    - Ireland: hurdle of 10.5 percent for baseline and 6 percent EBA CT1 under stress (up from the 4 percent required minimum), plus an additional protective buffer.
    - Spain TD and BU stress tests: EBA CT1 hurdle rate of 9 percent under baseline and 6 percent for adverse scenario.
- Microprudential vs macroprudential tensions:
  - High post-stress capital ratio requirements (microprudential) to meet market expectations could lead to excessive deleveraging, reduced credit issuance and exacerbate downturns (macroprudential concern).
  - Banks should build strong prudential buffers during good times so they can reduce them during bad times while respecting microprudential objectives.
  - Ireland and Spain stress tests applied this philosophy: high CT1 adequacy under baseline (incorporating a buffer), with expectation banks could draw on the buffer in extreme stress.
  - During bad times, encouraging increases in capital levels rather than ratios could align micro- and macroprudential objectives.

### Transparency, objectives, action plans, and financial backstops
- Design requirements for crisis stress tests:
  - Should meet a well-specified policy goal: ensure the financial system returns to health and that recovery is durable.
  - The exercise should determine underlying conditions of banks before prejudging outcomes.
  - A comprehensive strategy to address findings must be in place at commencement to avoid uncertainty.
    - Must include a clear action plan and credible financial backstops against possible adverse findings (see Schuermann, 2012).
    - Example: revelation of a large capitalization gap with no market access requires ready sources of funding to fill capital needs.

### Box 2 (illustrative implications of capital hurdle rates)
- Example assumptions:
  - Bank pre-shock CT1 ratio: 9 percent.
  - Constant RWA.
  - Required CT1 post-stress hurdle rate: 11 percent (includes a buffer).
- Baseline scenario example:
  - (a) Under baseline, CT1 reduced by 2 percentage points to 7 percent.
  - (b) Bank expected to raise CT1 to 11 percent: increase of 4 percentage points.
  - (c) Bank would need to top up existing 9 percent CT1 by another 4 percentage points to 13 percent in anticipation.
  - (d) Implies total capital adequacy ratio of more than 16 percent once additional requirements to make up T1 and total capital are included.
  - (e) If central case materializes, bank would have CT1 of targeted 11 percent.
- Adverse scenario example:
  - (i) Under severe adverse, losses reduce CT1 by 6 percentage points to 3 percent.
  - (ii) Bank expected to raise CT1 back to 11 percent: increase of 8 percentage points.
  - (iii) Bank would need CT1 of 17 percent (existing 9 percent plus another 8 percentage points).
  - (iv) Implies total capital adequacy ratio of more than 20 percent once additional requirements are included.
  - (v) If baseline materializes, bank would carry CT1 of 15 percent (17 percent less the 2 percentage points impact).
- Observed application:
  - Private sector stress tests of Spanish banking sector in 2011–12 applied stringent assumptions: projected losses of up to half, CT1 thresholds up to 11 percent, and a capital hole of up to €120 billion.
  - Actual outcome: baseline growth scenario for 2012 eventually became the actual outcome (see Spain example in Box 1).
- Note on replenishment:
  - Absolute amount needed to “replenish” capital may be lower if RWA decreases in line with loan losses.

*Source: Box 1. Designing Crisis Stress Test Growth Scenarios (excerpt).*

### Box 2. The Potential Impact of Capital Hurdle Rates for Crisis Stress Tests

### Box 2. The Potential Impact of Capital Hurdle Rates for Crisis Stress Tests

### Design, objectives, and credible backing of crisis stress tests
- Clear policy objectives and credible financial backstops strengthen market confidence and support recapitalization:
  - SCAP (U.S.): objective was to ensure largest banks had sufficient capital buffers so they would remain well-capitalized and continue providing credit and intermediation services even in a more challenging environment than anticipated (Fed, 2009c; Tarullo, 2010).
  - SCAP process: banks needing to augment capital post-stress test were given one month to design a detailed plan and six months to raise the requisite extra capital; the Treasury provided a contingent common equity bridge under the CAP of TARP.
  - SCAP forward guidance clarified that the exercise would not be used as a pretext for government takeovers if nationalization was not necessary; stock prices of SCAP banks outperformed non-SCAP banks during the period.
- Lack of explicit objectives, action plans, or a collective backstop undermined confidence in other jurisdictions:
  - CEBS 2009 objectives were vague; CEBS 2010 and EBA 2011 explicitly aimed to provide policy information on resilience and inform policymakers about banks’ ability to absorb adverse shocks.
  - CEBS 2010 follow-up measures were left to national authorities; European governments could not provide a collective financial backstop for CEBS or EBA stress tests.
  - Spain’s TD stress test (June 2012) produced an “overall figure” for recapitalization needs but lacked details on financial backstop or follow-up action; markets only firmed after the MoU with the Eurogroup in July and ESM support under the BU stress test.

### Disclosure, transparency, and market acceptance
- Transparency is essential to reduce investor uncertainty about loss recognition, capital adequacy, applicable capital standards, and government actions toward insolvent banks (Peristian and others, 2010; Goldstein and Sapra, 2012).
- Effective disclosure elements observed in successful exercises:
  - Stress test design and methodology, implementation, macroeconomic and loan loss assumptions, and bank-level results showing projected losses for key portfolio categories, capital components, and projected capital shortfalls (Table 8 summary).
  - SCAP’s public disclosures of estimated losses and capital needs provided information that private sector analysis did not have and, together with the financial backstop, reassured markets (Peristian and others, 2012).
- Disclosure pitfalls:
  - Publishing results without credible follow-up powers or backstops can fail to allay market concerns (EU cases).
  - Poorly designed disclosure can introduce noise and possibly create panic (Goldstein and Sapra, 2012).

### Asset Quality Review (AQR) and data reliability
- An AQR of banks’ portfolios should precede stress tests to ensure reliable inputs; scope of AQR depends on market perceptions and reporting reliability.
- AQR includes a DIV exercise to verify data completeness, IT and risk monitoring systems, and may involve:
  - (i) the direct cost of running the exercise (third-party contractors, auditors, asset valuers);
  - (ii) the cost of cleaning up inaccurate reporting via additional provisioning that flows through profits to capital, taken ahead of the stress test.
- Jurisdictional approaches:
  - U.S.: lower-intensity substitute to a full AQR—banks estimated cash flow losses using supervisor-provided indicative loss rate ranges, adjusted by bank-specific factors (past performance, origination vintage, borrower characteristics, geographic distribution, international operations, business mix); submissions were reviewed by accounting and asset-pricing experts and on-site supervisors (Fed, 2009a). Markets were reassured and banks recapitalized.
  - Ireland and Spain: variants of more comprehensive AQRs were conducted.
  - EBA recommended national supervisors conduct AQRs ahead of the 2014 EU stress test; coordinated with the SSM balance sheet assessment to be conducted by the ECB.

### Follow-up stress tests and evolution of public exercises
- Follow-up stress tests help consolidate gains; U.S. supervisory tests have evolved with greater disclosure:
  - Post-SCAP: annual supervisory stress tests; CCAR implemented in 2011 (results initially confidential), with CCAR 2012 and CCAR 2013 and DFA 2013 published in detail (Fed, 2011b, 2012a, 2012b, 2013a, 2013b).
  - From 2013, DFA stress tests run alongside CCAR, requiring annual and mid-cycle supervisory tests for systemically important financial institutions and publication of results.
- Stress scenario severity:
  - CCAR 2011 projected an adverse growth shock of 1.4 standard deviations from the projected baseline.
  - CCAR 2012 and CCAR 2013 assumed adverse growth scenarios of 2.5 standard deviations.
  - DFA mid-cycle: each bank develops its own baseline, adverse, and severely adverse scenarios; banks must publish results of their severely adverse scenarios to promote market discipline (Fed, 2013c).
- European plans:
  - EBA scheduled another round of EU stress tests for 2014 (EBA, 2013).
  - Ireland postponed the next PCAR to 2014 H1, to take place before the EBA 2014 stress test.
  - Spain’s BdE intends to include stress tests internally and regularly; next steps consider the EBA’s AQR recommendation and the SSM balance sheet assessment.

### Liquidity stress testing
- Liquidity stress testing became important after large liquidity shocks in the crisis; earlier crisis stress tests largely focused on solvency.
- Observations:
  - EBA exercises: liquidity risk was not specifically assessed in the EBA stress tests; a confidential thematic review of liquidity funding risks was initiated in 2011 Q1; EBA 2011 solvency test analyzed evolution of funding costs including sovereign stress impacts.
  - Spain: no standalone liquidity stress tests; solvency tests assumed increases in funding costs per scenarios.
  - Ireland: PLAR 2011 conducted the only crisis liquidity stress test to date, covering the four PCAR banks and setting bank-specific funding targets consistent with Basel III and other measures (CBI, 2011); PCAR 2011 specified funding cost and access constraints aligned with the PLAR.
- Uncertainty remains whether supporting liquidity tests would have enhanced solvency exercises in Europe; funding conditions improved only after LTRO and OMT facilities.

### Implementation considerations and risks of transparency
- Benefits of transparency: reduces opacity, improves market discipline, and pre-empts uncertainty in future crises.
- Potential costs/trade-offs during normal times:
  - May encourage banks to reverse-engineer supervisory models rather than map true risk profiles (Schuermann, 2013).
  - May incentivize banks to hold portfolios that perform well in tests but are suboptimal long-term (Sapra and Goldstein, 2012).
  - May lead to market over-reaction ex post or reduce speculative trading that provides useful price information.

### Operational lessons and requirements for effective crisis stress tests
- Credible backstops and clear action plans are critical to translate stress test findings into restored market confidence and recapitalization:
  - Timelines and contingency financing arrangements (e.g., one month to plan, six months to raise capital in SCAP) support orderly recapitalization.
  - Where restoration of solvency requires balance sheet and cost restructuring, capital injections alone may be ineffective; restructuring should be swift and avoid worsening sovereign debt burdens (Borio and others, 2012; Claessens and others, 2011).
  - Resolution frameworks must be adequate to deal with non-viable banks.
- Disclosure best practice:
  - Publish stress test design, methodology, scenarios, loan loss and other key assumptions, and individual bank-level results for key portfolios and capital components.
  - Ensure credible review procedures (internal or third-party) and prior validation of asset quality where necessary to bolster market confidence.
- AQR and data-validation are foundational to credible stress tests and may entail significant upfront costs but reduce the risk of misleading outcomes and the need for larger post-test capital actions.

*Source: Box 2. The Potential Impact of Capital Hurdle Rates for Crisis Stress Tests (continued), IMF working paper content.*

### 2.    No specific numbers are provided for the CCAR 2011 and 2012 ba

### _wp13178 - 2.    No specific numbers are provided for the CCAR 2011 and 2012 ba

### CCAR / DFA assumptions and scope
- No specific numbers are provided for the CCAR 2011 and 2012 baseline growth forecasts; identical industry sources as SCAP are assumed.
- All U.S. domiciled banking organizations are required to compute risk-based capital requirements using the regulatory capital definition (general-risk based capital rules, Basel I); none had entered a transitional floor period for RWAs as at 2011.
- The U.S. T1 leverage minimum:
  - 3 percent for banks with a composite Bank Holding Company Rating System (BOPEC) rating of "1" and for those that have implemented the Board's risk-based capital measure for market risk;
  - 4 percent for all other banks.
- CCAR 2012 applies Basel III framework calculations, fully phased.
- Relationship between DFA and CCAR:
  - Projections of pre-tax net income from the DFA exercise are used as direct inputs to the CCAR.
  - Primary difference: capital action assumptions. The Fed uses a standardized set of capital action assumptions for the DFA; in contrast, BHCs' planned capital actions are incorporated in the CCAR to project post-stress capital ratios.

### Capital metrics, coverage, timing and disclosed loss estimates (selected entries from the comparative table)
- SCAP 2009 (United States):
  - Exercise: SCAP 2009; Objective: capital adequacy guidelines; Number: 19; Percent of System: 67; Period: 2009-10; Risk Horizon baseline period: 1979-2008; Growth Rate Std Devn (SD) over Period: 2.9; SD of 2-Year Losses: 1.0; Metric(s): Fed's risk-based Tier 1 common capital, Tier 1 capital; Hurdle: 4; Publication: Yes; Disclosed Losses: USD599.2 (Billions).
- CEBS / EBA and other EU exercises (selected):
  - CEBS Stress Test 2009: Number: 22; Percent of System: 60; Period: 2009-10; Baseline: 1979-2008; Growth Rate SD: 2.0; SD of 2-Year Losses: 1.9; Metric(s): CRD Tier 1 capital; Hurdle: 6; Publication: No; Disclosed: n.a.
  - CEBS Stress Test 2010: Number: 91; Percent of System: 65; Period: 2010-11; Baseline: 1980-2009; Growth Rate SD: 2.5; SD of 2-Year Losses: 1.3; Metric(s): CRD Tier 1 capital; Hurdle: 6; Publication: Yes; Disclosed: EUR565.8.
  - EBA Stress Test 2011: Number: 90; Percent of System: 65; Period: 2011-12; Baseline: 1981-2010; Growth Rate SD: 2.7; SD of 2-Year Losses: 1.5; Metric(s): CRD Core Tier 1; Hurdle: 5; Publication: Yes; Disclosed: EUR400.0.
- United States — supervisory exercises (selected rows):
  - CCAR 2011:
    - Exercise: Supervisory; Number: 19; Percent of System: 65; Period: 2010 Q4 - 2011-12; Baseline: 1980-2009; Growth Rate SD: 3.4; SD of 2-Year Losses: 1.4; Metric(s): Fed's risk-based Tier 1 common capital, Tier 1 capital, Total risk-based capital, Tier 1 leverage; Hurdle(s): 5, 6, 8, Tier 1 leverage 3 or 4; Publication: No; Disclosed: n.a.; Publication date listed: 2013 Q4.
  - CCAR 2012:
    - Exercise: Supervisory; Number: 19; Percent of System: 67; Period: 2011 Q4 - 2011 Q4 - 1981-2010; Growth Rate SD: 3.5; SD of 2-Year Losses: 2.5; Metric(s): Fed's risk-based Tier 1 common capital, Tier 1 capital, Total risk-based capital, Tier 1 leverage; Hurdle(s): 5, 6, 8, Tier 1 leverage 3 or 4; Publication: Yes; Disclosed Losses: USD534.0 (Billions); Publication date listed: 2014 Q4; Estimated period: 2013 Q3.
  - DFA 2013:
    - Exercise: Supervisory; Number: 18; Percent of System: > 70; Period: 2012 Q4 - 2012 Q4 - 1982-2011; Growth Rate SD: 3.5; SD of 2-Year Losses: 2.5; Metric(s): Fed's risk-based Tier 1 common capital, Tier 1 capital, Total risk-based capital, Tier 1 leverage; Hurdle(s): 5, 6, 8, Tier 1 leverage 3 or 4; Publication: Yes; Disclosed Losses: USD462.0 (Billions); Publication date listed: 2015 Q4; Estimated period: 2014 Q3.
  - CCAR 2013:
    - Exercise: Supervisory; Number: 18; Percent of System: > 70; Period: 2012 Q4 - 2012 Q4 - 1982-2011; Growth Rate SD: 3.5; SD of 2-Year Losses: 2.5; Metric(s): Fed's risk-based Tier 1 risk-based capital, Total risk-based capital, Tier 1 leverage; Hurdle(s): 6, 8, Tier 1 leverage 3 or 4; Publication: Yes; Disclosed: n.a.; Publication date listed: 2015 Q4; Estimated period: 2014 Q3.

### Comparing crisis stress test shortfalls with restructuring costs (Section V findings)
- Crisis stress test capital shortfalls and eventual restructuring recapitalization needs are not expected to match because they operate under vastly different assumptions.
- Reasons restructuring costs are typically higher:
  - Stress tests assume the Going Concern Principle: banks operate as going concerns indefinitely and valuations focus on projected cash flow credit losses rather than liquidation values.
  - Restructuring is a bespoke exercise requiring thorough assessment of banks’ books, recognition of valuation losses (e.g., foreclosed real estate holdings or tax credits), projections of future losses under stress, and potential realization of non-core assets at significant haircuts.
- Spain case example (FROB observations):
  - Three items not captured in crisis stress tests for savings banks: (i) compensation due for breach of insurance contracts; (ii) the fall in dividends from equity holdings; (iii) differences between deferred tax credits and realized tax credits.
  - Explanation: nationalized institutions may face more strenuous stress or marking-to-market rules and are required to sell industry equity stakes, while others may retain such assets and not recognize similar losses.

### Concluding remarks (Section VI key points and policy implications)
- Stress tests have become an important instrument in supervisors’ crisis management toolkit; they use microprudential exercises to address macroprudential risks through improved transparency and disclosure.
- Design elements critical for systemic crisis management stress tests:
  - Public nature requires robustness to intense scrutiny: timing, governance, objective, proposed action plans, and disclosure may need to differ from confidential supervisory tests.
  - Scope of coverage and scenario design must be sufficiently rigorous; scenario complexity need not be high.
  - Supporting activities to enhance credibility:
    - AQRs (Asset Quality Reviews) are vital for input reliability.
    - Follow-up stress tests are important to update markets and consolidate gains.
    - Separate liquidity stress tests to complement solvency tests are increasingly used, though not all are published.
  - Political economy considerations can influence design and effectiveness.
  - Luck can materially affect outcomes (e.g., actual health of banks, timing, market conditions, public receptiveness).
- Policy prescription for authorities undertaking crisis stress tests:
  - Authorities must be fully committed, have a clear objective, and be prepared to take action once valuations have fallen to certain levels.
  - Before a crisis of confidence becomes entrenched, authorities must transparently examine the banking system, take necessary follow-up actions, and have requisite resources to back them.
  - Without commitment and resources, stress tests can backfire and exacerbate market confidence loss, with potential real economy consequences.
- Transparency trends:
  - Crisis stress tests may herald a new era in transparency; public U.S. crisis stress tests were followed by supervisory stress tests providing similar disclosure levels.
  - Some jurisdictions continue to maintain or improve transparency for follow-up crisis stress tests; it remains uncertain if similar transparency will extend to normal-time supervisory stress tests in the future.
  - The bar for disclosure has been set high and markets may demand similar standards when crises recede.

*Excerpted from the provided _wp13178 content.*

### Appendix Table 1. Crisis Stress Tests: Features of Design

### Appendix Table 1. Crisis Stress Tests: Features of Design

### Effectiveness — Financials stock index stabilizes/improves and returns volatility falls
- United States: Yes—The financials stock index never again fell below the level recorded at the time the stress test results were announced; volatility dropped sharply following the announcement of results.
- European Union (CEBS 2009): No.
- European Union (CEBS 2010): No.
- European Union (EBA 2011): No.
- Republic of Ireland (PCAR 2011): Yes.
- IMF FSAP 2012 Top-down (TD) 2012 Exercise: Not applicable—The FSAP stress test was not intended as a crisis management exercise.
- IMF FSAP 2012 Bottom-up (BU) 2012 Exercise: No.
- Spain: Yes—The financials stock index never again fell below the level recorded at the time the stress test results were announced; volatility dropped sharply following the announcement of results.

### Timing of exercise — Stress test is conducted sufficiently early to arrest the decline in confidence
- United States: The exercise was conducted 24 months after the peak of the financials stock index, while the index was still falling.
- European Union (CEBS 2009): The exercise was conducted 23 months after the peak of the financials stock index, while the index was still falling.
- European Union (CEBS 2010): The results were announced 9 months following the CEBS 2009 results—after the Ireland and Greece bailouts provided some support for market sentiment—despite some loss in credibility.
- European Union (EBA 2011): The results were announced 12 months after the CEBS 2010 results, but the exercise had lost significant credibility with markets by that stage.
- Republic of Ireland (PCAR 2011): The exercise was conducted 49 months after the peak of financials stock index, after almost all market value had been lost.
- IMF FSAP 2012 Top-down (TD) 2012 Exercise: Not applicable—FSAPs to each S-25 country are conducted mandatorily once every five years; the Spain FSAP took place 60 months after the peak of the financials stock index.
- IMF FSAP 2012 Bottom-up (BU) 2012 Exercise: The exercise was conducted 64 months from the peak of the financials stock index, after the supervisors had lost significant credibility with markets.
- Spain: The exercise was conducted 3 months after the TD stress test results, of which markets were skeptical given the lack of information on individual banks.

Additional market-condition notes:
- United States: The index had fallen by 68 percent from the peak and the price-to-book ratio was 0.99 as at end-2008.
- EU (CEBS 2009): The index had fallen by 72 percent from the peak and the price-to-book ratio was 0.73 as at end-2008.
- EU (EBA 2011): The index had fallen by 98 percent from the peak and the price-to-book ratio was 0.26 as at end-2010.
- Republic of Ireland (PCAR 2011): The index had fallen by 70 percent from the peak at the time the FSAP report was published in June.
- IMF FSAP 2012 BU 2012 Exercise: The index had fallen by 61 percent from the peak and the price-to-book ratio was 0.26 as at end-2011.

### Governance — Oversight
- United States: No—The stress test was overseen by the authorities.
- European Union (CEBS 2009): Yes—The stress test was overseen and coordinated by the CEBS in cooperation with the EC and the ECB.
- European Union (CEBS 2010): Yes—The stress test was overseen and coordinated by the CEBS in cooperation with the EC and the ECB.
- European Union (EBA 2011): Yes—The stress test was overseen and coordinated by the EBA in cooperation with the EC, the ECB and the ESRB.
- Republic of Ireland (PCAR 2011): No—The stress test was overseen by the CBI.
- IMF FSAP 2012 TD 2012 Exercise: Not applicable.
- IMF FSAP 2012 BU 2012 Exercise: Yes—The stress test was overseen by a Steering Committee comprising the authorities, supported by an Advisory Panel consisting of the Troika, Banque de France and the Dutch National Bank.
- Spain: Yes—The stress test was overseen by a Strategic Coordination Committee consisting of the authorities, the Troika and the EBA and in consultation with an Expert Coordination Committee comprising the same membership.

### Governance — Stress tester(s)
- United States: No—The stress test was conducted by the authorities: The Fed, the FDIC and the OCC.
- European Union (CEBS 2009): No—The stress test was conducted by national supervisory authorities.
- European Union (CEBS 2010): No—The stress test was conducted by national supervisory authorities.
- European Union (EBA 2011): No—The stress test was conducted by national supervisory authorities.
- Republic of Ireland (PCAR 2011): No—The stress test was conducted by the CBI with BlackRock Solutions providing forecast losses; Boston Consulting Group provided oversight and challenge to BlackRock’s work.
- IMF FSAP 2012 TD 2012 Exercise: No—The stress test was conducted by the BdE, in collaboration with the IMF.
- IMF FSAP 2012 BU 2012 Exercise: Yes—The stress test was conducted by Oliver Wyman and Roland Berger.
- Spain: Yes—The stress test was conducted by Oliver Wyman, with inputs from auditors and real estate valuation companies, all coordinated by the Boston Consulting Group.

### Scope — Approach
- United States: Yes—BU and TD mix, with authorities providing macroeconomic scenarios and guidance on the estimation of loan loss parameters.
- European Union (CEBS 2009): Yes—Constrained BU.
- European Union (CEBS 2010): Yes—Constrained BU (of cross-border groups) and simplified stress tests, based on national supervisors and reference parameters provided by the ECB for less complex institutions, with macroeconomic and sovereign shock scenarios and parameters developed by the ECB.
- European Union (EBA 2011): Yes—Constrained BU (of cross-border groups) and simplified stress tests, based on national supervisors and reference parameters provided by the ECB for less complex institutions, with macroeconomic and sovereign shock scenarios and parameters developed by the ECB.
- Republic of Ireland (PCAR 2011): No—Constrained TD, relying on BlackRock’s assessment of forecast losses and incorporating much of the methodology and parameters used for the EBA stress test.
- IMF FSAP 2012 TD 2012 Exercise: No—Constrained TD, with the authorities running both IMF and BdE models, applying growth scenarios and assumptions agreed with the IMF.
- IMF FSAP 2012 BU 2012 Exercise: No—TD of individual banks with growth scenarios and some guidelines provided by Steering Committee comprising the Troika, internal and external agencies.
- Spain: No—“BU” exercise comprising TD of individual banks, with growth scenarios and some guidelines provided by Steering Committee comprising the Troika, internal agencies and other EU counterpart agencies.

Processes described:
- United States: Process includes supervisors applying independent quantitative methods using firm-specific data to support their assessments of banks’ submissions.
- EU (CEBS 2009): Process includes a peer review of results and challenging process, and extensive cross-checks by the CEBS.
- EU (EBA 2011): Process includes consistency checks by the EBA, a multilateral review and TD analysis by the EBA and the ESRB with ECB assistance.
- Spain: Process includes close consultation with an Expert Coordination Committee comprising the Troika, the EBA and the authorities.

### Scope — Coverage
- United States: Yes—19 BHCs each with assets > $100 billion.
- European Union (CEBS 2009): Yes—22 major EU cross-border banking groups.
- European Union (CEBS 2010): Yes—91 banks, comprising major EU cross-border banking groups and a group of additional, mostly larger credit institutions.
- European Union (EBA 2011): Yes—90 banks, comprising major EU cross-border banking groups and a group of additional, mostly larger credit institutions.
- Republic of Ireland (PCAR 2011): Yes—All Irish domestically-owned commercial banks (4).
- IMF FSAP 2012 TD 2012 Exercise: Yes--Commercial banks and intervened savings banks (13), including the largest and problem banks.
- IMF FSAP 2012 BU 2012 Exercise: Yes--Commercial banks and intervened savings banks (14), including the largest and problem banks.
- Spain: Yes—14 largest merged banking groups, disaggregated into 17 following stress test.

Coverage by system assets (selected entries preserved):
- United States: Yes--2/3 of banking system assets.
- European Union (CEBS 2009): Yes—60 percent of the EU banking system assets.
- European Union (CEBS 2010): Yes—65 percent of the EU banking system assets and at least 50 percent of assets of each national banking sector.
- European Union (EBA 2011): Yes—65 percent of the EU banking system assets and at least 50 percent of assets of each national banking sector.
- Republic of Ireland (PCAR 2011): Yes—Approximately 80 percent of banking system assets including foreign subsidiaries.
- IMF FSAP 2012 TD 2012 Exercise: Yes—88 percent of banking system assets excluding foreign branches.
- IMF FSAP 2012 BU 2012 Exercise: Yes—Approximately 90 percent of banking system assets excluding foreign branches.
- Spain: Yes—around 90 percent of banking system assets excluding foreign branches.

### Scenario design — Scenarios (size of shocks)
- United States: No—1 baseline and 1 adverse (equivalent to 1 std. devn. of cumulative shock from baseline over 2 years).
- European Union (CEBS 2009): No—1 baseline and 1 adverse (1.9 std. devn. of cumulative shock from baseline over 2 years).
- European Union (CEBS 2010): No—1 baseline and 1 adverse (1.3 std. devn. of cumulative shock from baseline over 2 years).
- European Union (EBA 2011): No—1 baseline and 1 adverse (1.5 std. devn. of cumulative shock from baseline over 2 years).
- Republic of Ireland (PCAR 2011): No—1 baseline and 1 adverse (consistent with the EBA 2011 stress testing exercise).
- IMF FSAP 2012 TD 2012 Exercise: No—1 baseline and 2 adverse (more adverse scenario is 1 std. devn. of cumulative shock from baseline over 2 years).
- IMF FSAP 2012 BU 2012 Exercise: No—1 baseline and 2 adverse (first 2 years identical to FSAP adverse scenario, plus a third year of negative growth).
- Spain: No—1 baseline and 1 adverse (first 2 years identical to FSAP adverse scenario, plus a third year of negative growth).

### Risk factors — Stress test applies shocks to key risk factors
Credit risk:
- United States: Yes—Credit risk: 12 categories of loans stress tested.
- European Union (CEBS 2009): Credit risk: Stress tested but little information available.
- European Union (CEBS 2010): Yes—Credit risk: 5 main portfolios stress tested.
- European Union (EBA 2011): Yes—Credit risk: 8 categories of loans stress tested; sovereign exposure in AfS banking book treated as credit risk.
- Republic of Ireland (PCAR 2011): Yes—Credit risk: 5 categories of loans stress tested.
- IMF FSAP 2012 TD 2012 Exercise: Yes—Credit risk: 5 categories of loans plus foreclosed assets stress tested.
- IMF FSAP 2012 BU 2012 Exercise: Yes—Credit risk: 6 categories of loans plus foreclosed assets stress tested.
- Spain: Yes—Credit risk: 6 categories of loans plus 5 segments of foreclosed assets stress tested.

Market risk:
- United States: Yes—Market risk: Investment (and trading securities portfolios for firms with trading assets > $100 billion) in the trading book, AfS and HtM in the banking book stress tested.
- European Union (CEBS 2009): No—Market risk: Sensitivity analysis of trading book only.
- European Union (CEBS 2010): No—Market risk: Sovereign and financial institution exposures stress tested, but in the trading book only.
- European Union (EBA 2011): No—Market risk: Sovereign and financial institution exposures stress tested, but in the trading book only.
- Republic of Ireland (PCAR 2011): Market risk: Securities portfolio stress tested but insignificant; no sovereign haircut.
- IMF FSAP 2012 TD 2012 Exercise: Market risk: Sovereign portfolio stress tested, but in the trading book and the AfS banking book only.
- IMF FSAP 2012 BU 2012 Exercise: Market risk: Sovereign portfolio not stress tested, but ECB LTRO facility in place.
- Spain: Market risk: Sovereign portfolio not stress tested, but ECB LTRO facility in place.

### Assumptions — Macroeconomic scenarios standardized across banks
- United States: Yes—Baseline based on industry consensus forecasts; adverse provided by authorities.
- European Union (CEBS 2009): Yes—Based on collaboration among the CEBS, the EC and the ECB.
- European Union (CEBS 2010): Yes—Based on collaboration among the CEBS, the EC and the ECB.
- European Union (EBA 2011): Yes—Based on collaboration among the EBA, the EC and the ECB/ESRB.
- Republic of Ireland (PCAR 2011): Yes—Based on the EBA scenarios.
- IMF FSAP 2012 TD 2012 Exercise: Yes—Agreed upon between IMF staff and authorities.
- IMF FSAP 2012 BU 2012 Exercise: Yes—Based on FSAP scenarios.
- Spain: Yes—Based on FSAP scenarios, extended to a 3-year horizon.

Behavioral assumptions standardized across banks:
- United States: No—Banks’ own.
- European Union (CEBS 2009): Yes—Benchmark risk parameters provided by the ECB; process and guidelines provided by the CEBS.
- European Union (CEBS 2010): Yes—Range of assumptions provided by the CEBS, in cooperation with the ECB, the EC and national supervisory authorities.
- European Union (EBA 2011): Yes—Range of assumptions provided by the EBA, in cooperation with the EC, the ECB, the ESRB and national supervisory authorities.
- Republic of Ireland (PCAR 2011): Yes—Assumptions largely based on the EBA stress test.
- IMF FSAP 2012 TD 2012 Exercise: Yes—Range of assumptions provided by IMF staff.
- IMF FSAP 2012 BU 2012 Exercise: No—Assumptions provided by stress testers except for macro scenarios.
- Spain: No—Assumptions except for macro scenarios provided by stress tester with some guidance from Steering Committee.

### Capital standards — Stress test applies very high hurdle rate(s) (CT1 > 6 percent)
- United States: No—Hurdle rates of T1 capital of 6 percent and T1 common capital of 4 percent for both baseline and adverse scenarios.
- European Union (CEBS 2009): No—Hurdle rate of T1 capital of 6 percent for both baseline and adverse scenarios.
- European Union (CEBS 2010): No—Hurdle rate of T1 capital of 6 percent for both baseline and adverse scenarios.
- European Union (EBA 2011): No—Hurdle rate of CT1 capital of 5 percent for both baseline and adverse scenarios.
- Republic of Ireland (PCAR 2011): Yes—Hurdle rates of CT1 capital of 10.5 percent under the baseline scenario and 6 percent under the adverse scenario.
- IMF FSAP 2012 TD 2012 Exercise: Yes—Hurdle rates of Total capital of 8 percent, T1 capital of 6 percent and CT1 capital of 4 and 7 percent for baseline and both adverse scenarios.
- IMF FSAP 2012 BU 2012 Exercise: Yes—Hurdle rates of CT1 capital of between 6–9 percent.
- Spain: Yes—Hurdle rates of CT1 capital of 9 percent under the baseline scenario and 6 percent under the adverse scenario.

### Objective and action plan — Objective
- United States: Yes—Stated objective was specifically to assess the capital needs of banks that would provide a buffer against higher than generally expected losses and still be able to led to creditworthy borrowers should such losses materialize.
- European Union (CEBS 2009): No—Stated objective was to carry out an EU-wide forward looking stress testing exercise on the aggregate banking system.
- European Union (CEBS 2010): Yes—Stated objective was to provide policy information for assessing the resilience of the system to possible adverse economic developments and to assess the ability of to absorb possible shocks on credit and market risks, including sovereign risks.
- European Union (EBA 2011): Yes—Stated objective was to assess the resilience of individual institutions and the system to hypothetical stress events under certain restrictive conditions.
- Republic of Ireland (PCAR 2011): Yes—Stated objective was to determine the capital resources of domestic banks under a given stress scenario, in order to calculate the cost of recapitalization required to meet central Bank-imposed requirements.
- IMF FSAP 2012 TD 2012 Exercise: No—FSAP stress tests are conducted as part of the overall stability analysis of a financial system and to facilitate policy discussions on crisis preparedness. This objective was not stated explicitly.
- IMF FSAP 2012 BU 2012 Exercise: No—Stated objective was broadly to increase transparency and dispel doubts over the valuation of bank assets.
- Spain: Yes—Stated objective was to adhere to the MoU approved by the Eurogroup on July 20, 2012, which required the estimation of capital needs as an essential element of the roadmap for the recapitalization and restructuring of the banking system.

### Transparency — Follow-up action(s)
- United States: Yes—Banks needing to augment capital buffers were required to develop a detailed capital plan to be approved and implemented within 6 months.
- European Union (CEBS 2009): No—National authorities were responsible for any follow-up to the exercise.
- European Union (CEBS 2010): No—National authorities were responsible for any follow-up to the exercise.
- European Union (EBA 2011): Yes—Banks showing capital shortfalls were required to present their plans to restore capital position and to implement remedial measures by year-end.
- Republic of Ireland (PCAR 2011): Yes—Recapitalization were required based on loan loss projections, along with further calculations of prospective income, expenditure and deleveraging plans.
- IMF FSAP 2012 TD 2012 Exercise: No—FSAP stress tests are surveillance purposes and to facilitate policy discussions; they typically do not require management action.
- IMF FSAP 2012 BU 2012 Exercise: No.
- Spain: Yes—Recapitalization/ restructuring based on stress test findings required by the MoU with the Eurogroup.

Financing backstop:
- United States: Yes—Capital Assistance Program (CAP) under the Troubled Asset Relief Program (TARP).
- European Union (CEBS 2009): No—Not for the region as a whole.
- European Union (CEBS 2010): No—Not for the region as a whole.
- European Union (EBA 2011): No—Not for the region as a whole.
- Republic of Ireland (PCAR 2011): Yes—Already in a crisis program with the Troika.
- IMF FSAP 2012 TD 2012 Exercise: No.
- IMF FSAP 2012 BU 2012 Exercise: No.
- Spain: Yes—The ESM facility per the MoU with the Eurogroup.

### Disclosure of technical details — Design, methodology and implementation
- United States: Yes—Detailed information provided on stress test design, methodology and implementation.
- European Union (CEBS 2009): No—Minimal information on exercise was provided.
- European Union (CEBS 2010): Yes—Detailed information provided on stress test design, methodology and implementation.
- European Union (EBA 2011): Yes—Detailed information provided on stress test design, methodology and implementation.
- Republic of Ireland (PCAR 2011): Yes—Detailed information provided on stress test design, methodology and implementation.
- IMF FSAP 2012 TD 2012 Exercise: Yes—Detailed information provided on stress test design and methodology.
- IMF FSAP 2012 BU 2012 Exercise: Yes—Some information provided on stress test design, methodology and implementation.
- Spain: Yes—Detailed information provided on stress test design, methodology and implementation.

Model(s) disclosure:
- United States: No—Information on banks’ stress test models not disclosed, but projections subjected to detailed review and assessment by supervisors.
- European Union (CEBS 2009): No—Information on banks’ stress test models not disclosed, but results subjected to peer review and challenging process.
- European Union (CEBS 2010): No—Information on banks’/ national supervisory authorities’ models not disclosed, but results subjected to peer review and challenging process.
- European Union (EBA 2011): No—Information on banks’/ national supervisory authorities’ models not disclosed, but results subjected to peer review and challenging process.
- Republic of Ireland (PCAR 2011): No—Information on banks’ and third-party models not disclosed, but supervisory challenges and independent assessment undertaken.
- IMF FSAP 2012 TD 2012 Exercise: Yes—Information provided on IMF and Bank of Spain stress test models; results from two models cross-validated.
- IMF FSAP 2012 BU 2012 Exercise: Some information provided on third-party stress test models.
- Spain: No—Information on third party stress test models not disclosed but modeling process shared and discussed with representatives of the Steering Committee.

Details of assumptions disclosure:
- United States: Yes—High level information on macroeconomic assumptions; detailed information on loan loss assumptions.
- European Union (CEBS 2009): No—Very limited information provided on macroeconomic assumptions.
- European Union (CEBS 2010): Yes—Detailed information on macroeconomic and market assumptions.
- European Union (EBA 2011): Yes—Detailed information on macroeconomic and market assumptions.
- Republic of Ireland (PCAR 2011): Yes—Detailed information on macroeconomic and P&L assumptions and loan losses.
- IMF FSAP 2012 TD 2012 Exercise: Yes—Detailed information on macroeconomic and other behavioral assumptions.
- IMF FSAP 2012 BU 2012 Exercise: Yes—Detailed information on macroeconomic, P&L and loan loss assumptions/ estimates.
- Spain: Yes—Detailed information on macroeconomic and P&L assumptions and loan loss estimates.

Bank-by-bank results disclosure:
- United States: Yes—Summary results disclosed at individual BHC level, including projected losses, capital components and capital needs.
- European Union (CEBS 2009): No—Summary results disclosed at system aggregate level.
- European Union (CEBS 2010): Yes—Summary results disclosed at system aggregate and individual bank levels.
- European Union (EBA 2011): Yes—Summary results disclosed at system aggregate and individual bank level, including projected losses, capital components and capital needs.
- Republic of Ireland (PCAR 2011): Yes—Detailed results disclosed at individual bank level, including projected losses, capital components and capital needs.
- IMF FSAP 2012 TD 2012 Exercise: No—Summary results provided at aggregated groupwise (according to specific characteristics) and system aggregate levels.
- IMF FSAP 2012 BU 2012 Exercise: No—Summary results provided at system aggregate levels.
- Spain: Yes—Summary results disclosed at individual bank level, including projected loan losses, capital components and capital needs.

### Asset quality review (AQR)
- United States: Yes—Lower-intensity, quantitative substitute.
- European Union (CEBS 2009): No.
- European Union (CEBS 2010): No.
- European Union (EBA 2011): No.
- Republic of Ireland (PCAR 2011): Yes—Sample loan review and independent audit conducted.
- IMF FSAP 2012 TD 2012 Exercise: No.
- IMF FSAP 2012 BU 2012 Exercise: No.
- Spain: Yes—Deep dive conducted, supported by real estate appraisers and independent audit.

### Liquidity stress test — Liquidity stress test accompanies solvency stress test
- United States: No.
- European Union (CEBS 2009): No.
- European Union (CEBS 2010): No.
- European Union (EBA 2011): No—But liquidity profile of banks assessed through a confidential specific thematic review for supervisory purposes.
- Republic of Ireland (PCAR 2011): Yes—Prudential Liquidity Assessment Review (PLAR).
- IMF FSAP 2012 TD 2012 Exercise: Yes.
- IMF FSAP 2012 BU 2012 Exercise: No.
- Spain: No.

### Follow-up stress tests — Subsequent stress tests are conducted at regular intervals to maintain transparency
- United States: Yes—CCAR 2011, 2012, 2013 and DFA 2013 by the authorities, designed in a similar manner to the SCAP.
- European Union (CEBS 2009): Yes—CEBS 2010 by the authorities.
- European Union (CEBS 2010): Yes—EBA 2011 by the authorities.
- European Union (EBA 2011): Not conducted in 2012; another stress testing exercise planned for 2013.
- Republic of Ireland (PCAR 2011): Planned for 2014 H1, to take place before the EBA 2014 stress test.
- IMF FSAP 2012 TD 2012 Exercise: Not applicable—As noted above, FSAPs to each S-25 country are conducted mandatorily once every five years.
- IMF FSAP 2012 BU 2012 Exercise: Yes—“BU” exercise as part of MoU with the Troika.
- Spain: Timing to be decided, taking into account the EBA’s AQR and the balance sheet assessment in the context of the SSM, and the EBA 2014 stress test.

Other note:
- Conducted EU Recapitalization Exercise requiring banks to strengthen their capital positions against sovereign debt exposures.

*Sources: BdE; CBI; EBA; Fed; IMF; and authors. 1/ Included for completeness only—not intended as a crisis stress test; surveillance stress testing exercise was conducted in a crisis environment.*

### APPENDIX II. THE COST OF DIRECT RECAPITALIZATION VERSUS LOSS RECOGNITION AND

### APPENDIX II. THE COST OF DIRECT RECAPITALIZATION VERSUS LOSS RECOGNITION AND PROVISIONING

### When to perform an Asset Quality Review (AQR) versus running a stress test on reported data
- An AQR should be done if the quality of data is highly suspect such that any stress test result would be considered meaningless without it, which is typically the case in a crisis.
- AQRs tend to be more costly than running a stress test on existing data:
  - direct costs of running the exercise; and
  - additional costs because they usually require cleaning up the books before any stress test is conducted.
- The higher cost of an AQR reflects a “double hit” to the books: first recognizing losses and increasing provisions in the profit and loss account (via the AQR), and then potentially needing additional recapitalization from the crisis stress test.

### Numerical illustration: total capital needs of €50 billion (constant RWA)
- Assumption: stress test shows total capital needs of €50 billion.

- Example 1 — No AQR (use existing reported data):
  - Stress test projects a fall in revenue of €10 billion.
  - Expected loan losses are €20 billion.
  - Net profit decreases from €40 billion to €10 billion.
  - Projected capital shortfall is €10 billion and is addressed through a direct cash injection.

- Example 2 — AQR conducted first, then stress test:
  - Lender forbearance, loan misclassifications and incorrect valuations discovered, leading to additional required provisioning of €30 billion.
  - Stress test subsequently projects a drop in revenue of €10 billion.
  - Loan losses are estimated at another €20 billion.
  - The result is that the bank makes a net loss of €20 billion.
  - Projected capital shortfall is now €40 billion.

### Figure 8: accounting entries and starting balance-sheet values (as shown)
- Starting position (values in billions of euro, verbatim from figure):
  - Various revenues: 150
  - Various expenses: 110
  - Loans: 240
  - Various liabilities: 170
  - Provision: 0
  - Less provisions: 20
  - Profit: 40
  - Loans net of provisions: 220
  - Shares: 20
  - Cash: 20
  - Retained earnings: 10

- Example 1 (stress test then recapitalize to required €50 billion):
  - Revenues reduced by stress test: 140
  - Provision from stress test: 20
  - Less existing provisions: 20
  - Profit after stress test: 10
  - Cash reduced by drop in revenue: 10
  - Cash injection for capital: 10
  - Resulting loans net of provisions: 200
  - Resulting balance-sheet totals indicated at 220 (assets) and 220 (liabilities + capital) in figure.

- Example 2 (AQR realizes losses of €30 billion first, then stress test and recapitalize to required €50 billion):
  - Provision from AQR: 30
  - Less existing provisions: 20
  - Provision from stress test: 20
  - Loss: 20
  - Less provision from AQR: 30
  - Less provision from stress test: 20
  - Loss transferred from P&L: -20
  - Cash reduced by drop in revenue: 10
  - Cash injection for capital: 40
  - Resulting loans net of provisions: 170
  - Resulting balance-sheet totals indicated at 160 (assets) and 160 (liabilities + capital) in figure.

### Key analytical takeaway
- Conducting an AQR uncovers previously unrecognized losses and increases provisioning, which in turn raises the subsequent capital shortfall revealed by a stress test (the “double hit”).
- Choosing between immediate recapitalization based on reported data and performing an AQR first involves a trade-off: lower upfront recapitalization costs if using unadjusted data versus higher transparency and potentially larger recapitalization needs after loss recognition.

*Source: APPENDIX II. THE COST OF DIRECT RECAPITALIZATION VERSUS LOSS RECOGNITION AND PROVISIONING*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2013/_wp13178.pdf_