Chapter 16: Key Takeaways — Stress Testing and Scenario Analysis
Core Concepts Summary
Stress testing is the structured application of hypothetical but plausible adverse conditions to a financial institution's balance sheet, income statement, and capital position. It serves capital adequacy assessment, liquidity adequacy assessment, and strategic planning purposes. Post-2008 regulatory frameworks elevated stress testing from internal best practice to a mandatory pillar of prudential supervision.
Stress Testing Typology
| Type | Definition | Starting Point | Primary Use |
|---|---|---|---|
| Sensitivity Analysis | Examines impact of changing one variable while holding all others constant | Single risk factor shock (e.g., +100bp interest rate) | Communicating specific risk factor exposures; internal hedging decisions |
| Scenario Analysis | Applies a coherent, internally consistent set of macro-financial conditions simultaneously | Full macro scenario (GDP, unemployment, house prices, rates) | Regulatory submissions (CCAR, ACS, EBA); ICAAP; strategic planning |
| Reverse Stress Testing | Identifies scenarios that cause business model failure or capital/liquidity floor breach | The failure outcome (breach of capital/liquidity floor) | Recovery planning; identification of hidden vulnerabilities; Pillar 2 |
Key limitation of sensitivity analysis: By design it holds all other variables constant, so it cannot capture the systemic interdependencies that characterize actual crises — in which multiple risk factors deteriorate simultaneously in mutually reinforcing ways.
Key requirement of scenario analysis: Coherence. Variable values must be internally consistent with one another and with the narrative logic of the scenario. A scenario in which GDP falls 5% but unemployment is unchanged is internally inconsistent.
Key requirement of reverse stress testing: Genuine severity. Regulatory supervisors (particularly the PRA) distinguish between mechanically severe scenarios (e.g., GDP -20% in one year — historically unprecedented and therefore not a genuine risk) and genuinely institution-specific scenarios that identify real fragility points.
Regulatory Program Comparison
| Dimension | CCAR (US) | DFAST (US) | Bank of England ACS (UK) | EBA EU-Wide (EU) |
|---|---|---|---|---|
| Full Name | Comprehensive Capital Analysis and Review | Dodd-Frank Act Stress Test | Annual Cyclical Scenario | EU-Wide Stress Test |
| Administrator | Federal Reserve Board | Federal Reserve / State Regulators | Bank of England / PRA | European Banking Authority |
| Primary Applicability | BHCs / US IHCs of FBOs with ≥$100bn assets | Banks with ≥$100bn (Fed) or ≥$10bn (state regulators in some cases) | 7 largest UK banks and building societies | ~50 largest EU banks (~70% of EU banking assets) | |
| Scenario Authority | Fed publishes scenarios | Fed publishes scenarios | BoE publishes scenarios; counter-cyclical severity | EBA / ECB / ESRB publish scenarios |
| Scenarios Published | Baseline, Adverse, Severely Adverse | Baseline, Adverse, Severely Adverse | Typically one main adverse scenario (counter-cyclical) | Baseline + Adverse |
| Projection Horizon | 9 quarters | 9 quarters | Typically 5 years | 3 years |
| Balance Sheet Assumption | Dynamic (institutions model own balance sheet evolution) | Dynamic | Dynamic | Static (no balance sheet growth during stress) |
| Pass/Fail Threshold | Yes — minimum CET1 ratios apply; Fed can object to capital plans | Less formal than CCAR; feeds into supervisory assessment | Capital depletion assessed against systemic reference point | No formal pass/fail; feeds into SREP / Pillar 2 |
| Public Disclosure | Yes — institution-level results | Yes — results published | Yes — institution-level results | Yes — granular institution-level results |
| Capital Plan Review | Yes — qualitative process quality evaluated | No formal capital plan review | No formal capital plan review | No formal capital plan review |
| Frequency | Annual | Annual | Annual | Biennial (every 2 years) |
| Key Unique Feature | Qualitative objection to capital plan even if quantitative results pass | Applies to bank entity, not holding company | Counter-cyclical scenario severity calibration | Granular cross-institution comparability |
ICAAP vs. ILAAP: Key Distinctions
| Dimension | ICAAP | ILAAP |
|---|---|---|
| Full Name | Internal Capital Adequacy Assessment Process | Internal Liquidity Adequacy Assessment Process |
| Core Question | Do we hold enough capital to absorb losses under stress? | Can we fund our obligations through a stress period? |
| Regulatory Basis | Basel III Pillar 2; CRD IV Art. 73; PRA SS31/15 | CRD IV Art. 86; Basel III LCR/NSFR; PRA SS24/15 |
| Key Output | Capital Planning Buffer; Pillar 2A/2B calibration | Internal Liquidity Adequacy Standard; Pillar 2 liquidity buffer |
| Stress Scenarios | Macro adverse and severely adverse; reverse stress test | Idiosyncratic funding stress; market-wide funding stress; combined stress |
| Survival Concept | Capital remains above minimum requirements throughout stress horizon | Institution can fund itself for ≥90 days under combined stress |
| Board Approval | Required | Required |
| PRA Assessment | SREP — feeds into Total Capital Requirement | ILAA — feeds into Individual Liquidity Guidance |
| Common Deficiency | Implausible management actions (especially new equity issuance in stress) | Unrealistic deposit runoff assumptions; underestimation of contingent liquidity needs |
The Expected Loss Framework
The fundamental equation connecting macro scenarios to credit losses:
EL = PD × LGD × EAD
Under stress conditions: - Stressed PD: estimated via satellite models linking macro variables (GDP, unemployment, property prices) to segment-level default rates - Stressed LGD: incorporates collateral value changes (house price falls increase mortgage LGD; CRE price falls increase commercial mortgage LGD) - EAD: may increase for revolving facilities and committed lines as drawdown behavior changes
The total stressed Expected Loss across all segments represents the credit loss projection. This is combined with stressed Net Interest Income, operating expense projections, and tax to produce the stressed P&L — which, in turn, determines the stressed capital ratio.
Climate Risk Stress Testing: Emerging Framework
| Program | Jurisdiction | Year | Type | Key Finding |
|---|---|---|---|---|
| CBES (Climate Biennial Exploratory Scenario) | UK / Bank of England | 2021 | Learning exercise (not capital test) | Significant data and model gaps; limited counterparty-level emissions data |
| ECB Climate Stress Test | EU / ECB | 2022 | System-wide stress test | Material exposures to carbon-intensive sectors; methodology still developing |
| NGFS Scenarios | Global (network) | Ongoing | Reference scenario framework | Three pathways: Early Action, Late Action, No Additional Action |
Physical risk vs. Transition risk: - Physical risk: financial impact of climate change itself (extreme weather, sea level rise, chronic temperature shifts) - Transition risk: financial impact of the response to climate change (carbon taxes, stranded assets, technology disruption)
Key challenge: Climate scenarios operate over 10–30 year horizons vs. the 2–3 year horizon of conventional stress tests, requiring fundamentally different modeling approaches and scenario design methodologies.
Reverse Stress Testing: Regulatory Requirements Summary
| Regulator | Requirement | Key Standard |
|---|---|---|
| PRA (UK) | Required under SYSC 20 and SS31/15 | Must identify scenarios causing business model non-viability; results must be material to capital planning |
| EBA | Required under EBA/GL/2018/04 | Must be conducted as part of ICAAP; results reviewed by management body |
| BCBS | Endorsed in 2017 supervisory guidance | Complement to conventional scenario analysis; particularly useful for tail risk identification |
| Federal Reserve | Not formally mandated in same way; but part of CCAR qualitative assessment | Quality of reverse stress testing evidence reviewed in qualitative evaluation |
Common ICAAP/Stress Testing Deficiencies (Supervisory Findings)
- Insufficient scenario severity: Adverse scenarios too mild relative to the institution's specific risk concentrations
- Implausible management actions: New equity issuance in systemic stress; asset sales with unrealistic pricing or timing assumptions
- Box-tick reverse stress testing: Scenarios that are technically severe but mechanically constructed rather than institution-specific
- Variable inconsistency in scenarios: GDP and unemployment shocks not coherently linked; property price shocks not reflected in LGD estimates
- Insufficient board challenge: Board approves ICAAP without genuine engagement with scenario selection rationale or reverse stress test results
- Poor documentation: Audit trail of model choices, overlays, and governance approvals insufficient
- No IRRBB basis risk capture: Interest rate stress uses parallel yield curve shift without capturing basis risk between rate indices
- Static operational risk overlay: Operational risk stress not integrated with macro stress (OpRisk losses assumed independent of macro conditions)
Practitioner Checklist: Stress Testing Submission Readiness
Data and Modeling
- [ ] Portfolio exposure data reconciled between risk systems and finance general ledger
- [ ] Satellite model documentation complete and validation signed off
- [ ] Model overlays documented with written business justification
- [ ] Market risk and operational risk projections completed in parallel
- [ ] All results traceable from input data to final capital ratios (full audit trail)
Scenario Design
- [ ] Scenario narratives are internally consistent (macro variable cross-checks completed)
- [ ] Scenario severity calibrated to institution-specific risk concentrations (not generic)
- [ ] All three scenario types present: sensitivity, scenario, reverse
- [ ] Reverse stress test result linked to specific business model vulnerabilities
- [ ] Comparison to prior-year scenarios documented (explain changes)
Management Actions
- [ ] Each management action has a written plan with realistic timeline
- [ ] New equity issuance not assumed unless anchor investors documented
- [ ] Dividend suspension adequately captures preference shares and AT1 coupons
- [ ] Asset sale assumptions use stressed market pricing, not current valuations
- [ ] Board resolution (or equivalent approval) in place for each significant action
Governance
- [ ] Stress Testing Committee minutes document substantive challenge (not just ratification)
- [ ] Board review session conducted; board questions and management responses documented
- [ ] ICAAP/ILAAP document version-controlled with sign-off history
- [ ] Results shared with Recovery Plan team and cross-referenced
- [ ] Internal Audit independent review completed
Climate Risk (where applicable)
- [ ] Physical risk and transition risk both addressed (even if qualitatively where data gaps exist)
- [ ] Data gaps explicitly acknowledged and remediation plan documented
- [ ] NGFS or equivalent scenario framework used as reference
- [ ] Time horizon appropriate for chronic physical risk assessment (10+ years)
Key Regulatory References
- BCBS: "Supervisory and bank stress testing: range of practices" (December 2017)
- EBA: EBA/GL/2018/04 — Guidelines on institutions' stress testing
- PRA: SS31/15 — "The Internal Capital Adequacy Assessment Process (ICAAP) and the Supervisory Review and Evaluation Process (SREP)"
- Federal Reserve: "Dodd-Frank Act Stress Test 2023: Supervisory Stress Test Methodology"
- Bank of England: "Stress testing the UK banking system: 2022/23 results"
- ECB: "2022 climate risk stress test"
- NGFS: "NGFS Climate Scenarios for central banks and supervisors" (updated 2023)