## Executive Summary — Stress Testing Best Practices (_082212) ## Source details **Canonical URL:** [Executive Summary — Stress Testing Best Practices (_082212)](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/np/pp/eng/2012/_082212.pdf) ## Other formats - [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/np/pp/eng/2012/_082212.pdf.md) - [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/np/pp/eng/2012/_082212.pdf.json) --- ### Role, purpose, and seven best-practice principles - Recent crisis showed pre-crisis stress tests missed material vulnerabilities; stress tests useful for crisis management (example: SCAP 2009). - Fund experience: FSAPs since 1999, contributions to GFSRs, Expert Forum on Advanced Stress Testing Techniques. - Definition: quantitative “what if” exercise covering coverage, quantitative framework, communication strategy, and follow-up measures. - Purpose of paper: - Propose seven operational “best practice” principles to tailor stress tests and enable critical evaluation. - Focus on macrofinancial surveillance stress tests; also relevant for microprudential and institution-specific assessments. - Provide yardstick for exercises and agenda for toolkit improvements ahead of next FSAP review (tentatively 2014). - Seven principles: - Define appropriately the institutional perimeter for the tests. - Identify all relevant channels of risk propagation. - Include all material risks and buffers. - Make use of the investors’ viewpoint in the design of stress tests. - Focus on tail risks. - When communicating stress test results, speak smarter, not just louder. - Beware of the “black swan.” *Key operational insights* - Success depends on more than parameter choices; critical elements include: - Choice of risk scenarios covering all relevant risk factors and severity. - Design covering transmission channels and realistic buffer assumptions. - Choice of hurdle rates. - Clear ex ante objectives and knowledge of institutions, business models, and transmission channels. - Appropriate perimeter and coverage, complementary tools, communications strategy, and credible commitment to remedial measures. - Stress tests are hypothetical and cannot by themselves predict or prevent crises. *Empirical survey* - Survey (November 2011): 26 central banks/supervisory authorities from 23 countries responded (some joint submissions). - Among the 23 countries: seven emerging market economies, 16 developed economies; 13 European, 6 Asian, 4 Western Hemisphere. - Notable shortcomings concentrated in three areas: - Identifying channels of risk propagation. - Using the investors’ viewpoint. - Focusing on tail risks. *Operational implications summarized in Table 4 (page 48).* --- ### Solvency and Liquidity Tests; Typology and Implementation ### Solvency tests — definition, scope, and measurement - Objective: assess whether firm has sufficient capital to remain solvent in a hypothetically challenging environment by estimating profit, losses and valuation changes. - Main risk factors: - Credit risk. - Market risk (interest rates, exchange rates, equity prices). - Test types and shock design: - Single factor or multiple factor tests. - Combined shock tests (ad hoc) or macro scenario tests (coherent via macroeconomic framework). - Balance sheet coverage/valuation: - Credit risk may cover total loans or segmented loans. - Market risk assessed for trading and AFS accounts; HTM debt securities may be excluded. - Off-balance sheet exposures can be highly non-linear in stress. - Bank solvency tests usually do not adjust liabilities for interest rate changes; insurance and pension fund tests should. - Macro scenario implementation: - Requires macrofinancial models linking risk parameters (NPL ratio, PD, LGD, credit rating) to macro variables (GDP, unemployment, exchange rate, interest rate). - Typical horizon: several years (typically one to three for country supervisors; often longer in FSAPs). - Pre-impairment profits need projection; assumptions on dividend policy and deleveraging required. - Solvency measurement and hurdle rates: - Measures: statutory/core Tier 1/Tier1 to RWA; leverage ratios; losses percent of capital; capital shortfalls. - Institutions “pass” or “fail” against a pre-determined hurdle rate. - Hurdle rates often set at current minimum regulatory requirement but could differ (e.g., Basel III expectations or market funding cost); choice is critical. ### Liquidity tests — objective, mechanics, and interaction with solvency - Objective: examine whether institutions have enough cash inflows and liquid assets to withstand cash outflows in a stress scenario. - Sources of sudden outflows: maturity mismatch, deposit withdrawals, funding market freezes, margin calls, collateral downgrades, SPV support. - Mechanics and hurdle rates: - Liquid assets can be sold or repo’d; non-marketable loans or declining collateral values (haircuts) cause shortages. - Hurdle rates include: days of tolerance before negative cash flow; net cash flow position; stressed liquidity ratios. - Interaction with solvency: - Funding distress can convert liquidity shortages into solvency problems via fire sales and increased funding costs. ### Typology of stress tests (four objectives) - Internal risk management (bank-level). - Microprudential/supervisory (Pillar 1 & Pillar 2 under Basel II; guiding capital requirements and buffers; liquidity ratios in Basel III; Solvency II for insurance). - Macroprudential/surveillance (systemic risk, Financial Stability Reports; IMF FSAPs since 1999). - Crisis management (assess recapitalization needs and public support; examples: SCAP, CEBS/EBA exercises). - Evolution: from single-factor to macro scenario tests and from institution-specific to enterprise-wide; supervisory tests increasingly guide actions. ### Top-down (TD) versus Bottom-up (BU) - TD: national authorities/Fund using bank-by-bank data with consistent methodology (typical in FSAPs). - BU: individual institutions use internal data/models under common assumptions. - Strengths: - BU: granular coverage of exposures, hedges, complex products; uses internal models. - TD: uniform methodology and assumptions; resource-effective; validates BU. - Weaknesses: - BU: resource intensive; cooperation-dependent; cross-institution comparison issues. - TD: precision limits due to data; may not reflect institution-specific decisions. - Common practice: liquidity tests often BU; FSAPs almost always include TD, frequently supplemented by BU. --- ### Model Families, Emerging Markets, and Complementarity ### Two broad model families - Balance sheet-based (“fundamental”): detailed analysis of institutions’ balance sheets. - Strengths: identify source of vulnerabilities; applicable in thin market countries; adjust for supervisory weaknesses. - Weaknesses: backward-looking; data intensive; limited in capturing interdependence and integrated risks. - Market price-based: summary default measures embedded in market prices (stocks, bonds, derivatives). - Strengths: flexible; captures portfolio effects and tail events; high update frequency; less data-intensive. - Weaknesses: “black box” source identification; sensitive to short-term market swings; not applicable without market data. ### Application in emerging and developing economies - Constraints: limited data/weak accounting systems. - Feasible approaches: - Simple balance sheet-based tests with single or multi-factor shocks using basic supervisory data. - Improve supervision and data availability before developing advanced approaches. - Use cross-country data when time series are short (example referenced: Annex 1.6, April 2010 GFSR). ### Complementarity and systemic estimation - Balance sheet and market price approaches provide different information and are complementary. - Common challenge: assessing joint default probabilities and default dependence to produce systemic loss distributions. - Balance sheet extension: superimpose network of claims. - Market price approach: treat system as a portfolio and derive distribution via portfolio analysis. --- ### Principles in Detail: Perimeter, Channels, Risks, Investors’ View, Tail Risks, Communication, Black Swan ### Principle 1 — Define the institutional perimeter - Include systemically important institutions and relevant nonbanks and FMIs. - Criteria for systemic importance: size, substitutability, complexity, interconnectedness. - Network approaches can identify central institutions and clusters; examples include cluster analysis to identify gatekeeper institutions. - Empirical coverage: - FSAPs typically cover at least seventy percent of locally-incorporated commercial banks by assets. - Survey-based coverage ranges 60 to 100 percent (median 85 percent and 16 banks). - Number of banks in samples varies from below 5 to over 1,000. - Inclusion of nonbank sectors: FSAPs always test banking sector; insurance sector occasionally; pension funds and FMIs tested ad hoc. ### Principle 2 — Identify all relevant channels of risk propagation - Key channels: - Liquidity–solvency feedbacks. - Financial instability → real economy feedbacks (second-round effects). - Bank–sovereign linkages. - Policy feedbacks. - Gaps and challenges: - Incorporating real–financial interactions is a major challenge; calibration often relies on historical data or expert judgment. - Most liquidity tests are implemented independently of solvency tests; many do not capture intensified runs when banks sustain substantial losses. - Feedbacks from banking distress to macroeconomy and complex bank–sovereign linkages are rarely incorporated in country authority tests and FSAPs. - Emerging model development: - Work underway: DSGE with banking sector (GFSR 2011b), Bank of Japan econometric model, IMF expanding GIMF to include financial sector. ### Box 3 — Integrating liquidity and solvency, bank reactions - Banks’ reactions: use high-quality collateral, fire sales, deleveraging, debt-to-equity conversions. - Two-round practical approach: BU first round; TD second round calibrated with BU-reported reactions (haircuts, reduced repo roll-over). - Analytical approaches: models endogenizing funding costs and market/funding liquidity feedbacks (examples: Schmieder et al. (2012); Dutch Central Bank model; Hong Kong Monetary Authority framework; RAMSI). ### Principle 3 — Include all material risks and buffers - Post-crisis expansion: include sovereign, funding, systemic liquidity, counterparty risks, cross-border activities. - Buffers: pre-impairment income must be included, especially in multi-year tests. - RWA changes can substantially affect outcomes; careful modeling required. - Gaps: - Sovereign risks often modeled as market haircuts; methodological differences exist. - FSAPs in emerging markets include judgmental adjustments for supervisory weaknesses. - Persistently neglected risks: downgrades and interest rate risks in the banking book. ### Box 4 — Projection of pre-impairment income - Elasticities and averages: - Elasticity of pre-impairment income-to-capital to economic growth: range 1 to 1.5. - Average pre-impairment income-to-capital ratio in advanced and emerging economies: about 12 to 15 percent. - A moderate stress (4 percentage point drop in annual GDP growth in advanced economies; 6 percentage point drop in emerging economies) would reduce pre-impairment income-to-capital by about half. - For banks in advanced economies, elasticity rises to 4 under severe stress. - Components: net interest income, fee/commission income, non-interest expenses. - Funding-cost feedbacks can sharply reduce net interest income near regulatory minimums. ### Principle 4 — Make use of the investors’ viewpoint - Operational implications: - Adopt mark-to-market valuation under baseline and adverse scenarios. - Use economic capital complementing statutory capital. - Use point-in-time (PIT) parameters rather than through-the-cycle (TTC). - Explicitly stress market risk appetite and impose market-based hurdle rates in addition to regulatory minima. - Market-based hurdle rates are generally not used by country authorities; rare in FSAPs (U.S. FSAP 2010, IMF 2011a). ### Principle 5 — Focus on tail risks - Traditional calibration: “extreme but plausible” without systematic determination (examples: “worst-in-a-decade”, “1 percent probability”, “x standard deviation”). - Problems with historical calibration: varied experience, short histories, failure for unprecedented events. - In crisis situations: tests should not shy from severity but disclosure should be managed with credible support measures. - Interdependence and joint tails: - Balance sheet models: use networks to measure spillovers and joint default probabilities. - Market price models: CoVaR, MES, SES and systemic CCA to capture joint tail dependence. - Evolving practice: - Common FSAP rule of thumb: two-standard deviation shock on GDP growth for two years using long (20–30 years) history unless justified otherwise. - Survey respondents adopt tail events probability ranging from 1–5 percent or use crisis-magnitude shocks. ### Box 6 — Importance of risk dependence and distributional approaches - Methods: refined quantile statistics, copulas, EVT (generalized Pareto, GEV), systemic CCA with multivariate extreme value distribution (MGEV). - Implied capital requirement under distributional approaches identifies capital shortfall consistent with market expectations and extreme dependence. - Elements: New Expected Loss (after shock), VaR 95%, VaR 99%, additional regulatory and economic capital. ### Principle 6 — Communicate smarter, not just louder - Disclosure trends: - Increased disclosure since crisis; FSSAs often published with aggregate results. - Survey: almost 85 percent communicate macroprudential solvency stress test results; 50 percent communicate macroprudential liquidity stress tests. - Typical publication: Financial Stability Reports; aggregates and distribution metrics without identifying institutions (exceptions: EBA, U.S. Federal Reserve). - Tradeoffs: - Benefits: remove information asymmetry, restore confidence, better pricing, market discipline. - Risks: gaming, moral hazard, undermining confidence if support measures absent, media misinterpretation. - Preconditions for publication: - Candid assessments; explicit coverage and limitations; accompaniment by credible remedial measures. ### Principle 7 — Beware of the “black swan” - Stress tests cannot predict truly novel or highly improbable events. - Practical approaches: - Use expert judgment and current vulnerabilities to choose scenarios. - Reverse stress tests and institution-specific unique scenarios (Federal Reserve Board practice: each institution chooses one unique and faces one common scenario). - Scenario distribution approach (example: South Africa FSAP using Monte Carlo simulations) to generate distribution of bank capital ratios. - Complement stress tests with qualitative bank risk analysis, early warning indicators, debt sustainability models, and dialogue with supervisors. --- ### Crisis Management Exercises, FMIs, CCPs, Sovereign–Bank Linkages, and Frontiers ### Crisis management stress tests - Increased use post-crisis (SCAP 2009; CEBS/EBA 2010–2011). - Characteristics: - Combine microprudential and macroprudential features. - Bank-by-bank assessment with systemic shocks. - Follow-up: recapitalization, restructuring, public support. - Typically focus on solvency using balance-sheet-based tests and often static balance sheet assumptions. - Comparison of prominent exercises (selected quantitative findings): - 2011 European adverse scenario → -4.1 percentage points cumulative GDP shock. - 2010 European adverse scenario → -3.1 percentage points cumulative GDP shock. - SCAP adverse scenario involved a -2.8 percentage points cumulative GDP shock; SCAP assumed a cumulative two-year contraction in GDP of -2.7 percent. - 2011 EBA exercise assumed a cumulative contraction of -0.2 percent. - 2010 EBA exercise assumed a cumulative contraction of -0.4 percent. - Organizational complexity example: EU-wide tests involved 21 countries, 24 national supervisors, around a dozen European institutions, 91 participating banks, 19 languages, seven currencies. - Outcomes: - SCAP credited with restoring confidence; EBA 2011 had criticisms for inadequate severity and ambiguous backstops. ### FMIs and CCPs - FMIs tested for proper functioning (operational, credit, liquidity), focusing on ability to finish the day and subsequent operation. - CCPs concentrate credit and liquidity risk: - CCPs must satisfy settlement obligations of defaulting participants, cover liquidation losses, and meet cash flows for hedge transactions. - CPSS/IOSCO standard: CCPs should withstand default of participant with the largest exposure; some CCPs test multiple major defaults. - Central clearing of OTC derivatives increases complexity; improved rules expected by end-2012 with guidance on stress testing. - The Fund examines FMI stress testing arrangements in FSAPs but has not conducted stress tests on FMIs itself. ### Sovereign–bank linkages and modeling frontiers - Framework linking financial sector and government balance sheets emphasizes feedbacks: - Distressed financial institutions → government contingent liabilities → higher sovereign default risk → higher bank funding costs and asset valuation effects. - Recommended modeling enhancements: - Build stress test models with explicit feedback channels: financial shock → GDP reduction → increased borrower defaults → bank asset deterioration. - Use dynamic factor models and contagion models (networks, joint default probabilities, systemic CCA, co-risk indicators). - Techniques to capture non-linearities: - Risk-adjusted balance sheet and contingent claims approaches (CCA, Merton 1973; Cox and Ross 1976). - Two methods: portfolio-loss-based asset distribution or market-implied asset and volatility from equity and debt market values. - Funding-cost channel: higher bank funding costs → higher lending rates → credit rationing → lower output → feedback to banking distress. - Improvements: time-varying correlations, market price–based contingent claims models, “quasi-RWA” dynamic adjustment factors. --- ### Table 4 — Practical Implications (summary of operational guidance by principle) - Principle 1: If universal coverage infeasible, prioritize systemically important institutions including nonbanks and FMIs; use network models to assess interconnectedness; understand conglomerate structure. - Principle 2: Identify and model channels (solvency–funding relation, bank–sovereign links, bank reactions, macro feedbacks); if explicit modeling infeasible use heuristic second-round TD tests or expert judgment. - Principle 3: Be comprehensive in including risks (cross-border, nonbank activities, contingent liabilities); model buffers like pre-impairment income conservatively in multi-year tests. - Principle 4: Use mark-to-market valuations, point-in-time parameters, adjust initial capital for supervisory over-reporting, stress market appetite, and use market-based hurdle rates alongside regulatory minima. - Principle 5: Use multiple approaches for “extreme but plausible” shocks; capture tail events and dependencies via market-based and balance-sheet-based models supplemented with network tools. - Principle 6: Tailor communications to circumstances and test goals; ensure candid tests, serious shocks, appropriate hurdle rates, and credible follow-up before disclosure; integrate into broader financial stability communications. - Principle 7: Supplement statistical approaches with expert judgment and reverse stress tests; use a distributional approach to scenarios; combine stress tests with other assessment techniques and maintain ongoing reevaluation. --- *Prepared by IMF staff; chapter text as provided in the source PDF _082212.* ### Executive Summary....................................................................................................... ### Executive Summary ### Role and Context of Stress Testing - The recent financial crisis highlighted both failures of pre-crisis stress tests to capture material vulnerabilities and the usefulness of stress tests as crisis management tools (for example, the Supervisory Capital Assessment Program (SCAP) exercise in 2009). - The Fund has extensive practical experience from more than a decade of Financial Sector Assessment Programs (FSAPs) and contributions to Global Financial Stability Reports (GFSRs), and collaborates with supervisory agencies and central banks, including through the Expert Forum on Advanced Stress Testing Techniques. - Stress testing is defined as a quantitative “what if” exercise that measures the vulnerability of a portfolio, an institution, or an entire financial system under different hypothetical events or scenarios, and the term “stress test” in this paper denotes the whole process (coverage, quantitative framework, communication strategy, and follow-up measures). ### Purpose of the Paper - The paper proposes seven “best practice” principles for stress testing intended as practical operational guidelines derived from experience, especially from the recent crisis, and meant to: - Help practitioners tailor stress tests to specific circumstances (degree of financial sector development, business models, macroeconomic environment). - Empower informed observers to critically evaluate stress tests by country authorities, FSAP teams, or others. - The focus is mainly on stress tests conducted for macrofinancial surveillance purposes, with much of the discussion also relevant for microprudential and institution-specific assessments. - The proposed principles provide a yardstick for evaluating individual stress testing exercises and form an agenda for improvements in the Fund’s stress testing toolkit ahead of the next FSAP review tentatively scheduled for 2014. ### Seven “Best Practice” Principles - Define appropriately the institutional perimeter for the tests. - Identify all relevant channels of risk propagation. - Include all material risks and buffers. - Make use of the investors’ viewpoint in the design of stress tests. - Focus on tail risks. - When communicating stress test results, speak smarter, not just louder. - Beware of the “black swan.” ### Key Design and Implementation Insights - Success of stress tests depends on more than parameter choices; critical elements include: - Choice of risk scenarios covering all relevant risk factors and their severity. - Design that covers important transmission channels and realistic assumptions about buffers. - Choice of hurdle rates. - Clear ex ante understanding of stress test objectives. - Knowledge of key institutions, business models, principal risk sources, and transmission channels. - Appropriate perimeter and coverage decisions. - Use of complementary assessment tools. - A communications strategy tailored to circumstances and purpose. - A credible commitment to take remedial measures if vulnerabilities are uncovered. - Stress tests remain hypothetical statements and cannot by themselves predict or prevent crises. ### Empirical Findings on Current Practices - Practices have improved since the crisis but adherence to the principles is uneven, based on: - A survey conducted in November 2011 covering (i) the broad use and definition of stress tests; (ii) banking sector stress tests (process and organization; framework for solvency tests including risk/scenario selection, macrofinancial linkages, determining capital adequacy; framework for liquidity stress tests; communication strategy); and (iii) issues regarding the use of stress tests and their application to the nonbank financial sector. - A total of 26 central banks and supervisory authorities from 23 different countries responded (in some country cases, responses were jointly submitted by more than one agency). - Among the 23 countries: seven are emerging market economies and 16 are developed economies; 13 are European, 6 are Asian, and 4 are Western Hemisphere countries. - Notable shortcomings are concentrated in three areas: - Identifying the channels of risk propagation. - Using the investors’ viewpoint. - Focusing on tail risks. - These shortcomings reflect gaps in the analytical toolkit and weaknesses in implementation; understanding their implications is crucial for interpreting stress test results and closing these gaps is a key priority. ### Operational Implications - The paper summarizes key operational implications of the principles (referenced as Table 4 on page 48) to guide practitioners in implementing more reliable and policy-relevant stress testing exercises. *Executive Summary, _082212 - Executive Summary* ### 10. Stress tests typically evaluate two aspects of financial institutions’ performance: ### _082212 - 10. Stress tests typically evaluate two aspects of financial institutions’ performance: ### Solvency tests - Definition and objective: - An institution is solvent when the value of its assets is larger than its debt, i.e., there is a certain amount of positive equity capital. - A solvency test assesses whether the firm has sufficient capital to remain solvent in a hypothetically challenging environment by estimating profit, losses and valuation changes. - Main risk factors: - Credit risk: potential losses from borrowers’ default. - Market risk: losses from securities due to changes in market prices such as interest rates, exchange rates and equity prices. - Types of tests and shock design: - Single factor tests (one source of risk) or multiple factor tests (multiple sources). - Combined shock tests (ad hoc combination) or macro scenario tests (coherent generation using a macroeconomic framework). - Balance sheet coverage and valuation considerations: - Credit risk tests may cover total loans (including interbank lending) or loans to segments (corporate, mortgages, credit card). - Market risk assessed for securities held for trading and in the available-for-sales (AFS) accounts; debt securities in held-to-maturity (HTM) accounts may be excluded. - Off-balance sheet exposures (derivatives, contingent claims, securitization exposures) can be affected by market and credit risks and may have highly non-linear responses to stress. - Bank solvency tests usually do not adjust liabilities for interest rate changes; insurance and pension fund tests should adjust liabilities for interest rate sensitivity. - Macro scenario implementation: - Requires estimating macrofinancial models that link risk parameters (non-performing loan (NPL) ratio, probability of default (PD), loss given default (LGD), credit rating, etc.) to macro variables (GDP, unemployment, exchange rate, interest rate). - Macro scenario stress tests typically cover several years (typically one to three in the case of country supervisory authorities or central banks, and often longer in FSAPs). - Pre-impairment profits need projection since retained earnings in the test horizon affect capital; assumptions about bank behavior (dividend payout policies, deleveraging) are required. - Solvency measurement and hurdle rates: - Solvency measured by capital ratios: statutory/core Tier 1/Tier1 capital to risk-weighted assets (RWA); leverage ratios (capital to assets); losses in percent of capital; capital shortfalls. - Institutions or the system “pass” or “fail” if target capital ratio is above a pre-determined threshold or “hurdle rate.” - Hurdle rates are often set at the current minimum regulatory requirement, but could be set differently (e.g., when new regulations such as Basel III are expected) or to maintain market funding cost. - The choice of the hurdle rate is a critical factor, especially when test results are linked to capital planning directed by supervisors. ### Liquidity tests - Objective: - A liquidity stress test examines whether financial institutions have enough cash inflows and liquid assets to withstand cash outflows in a stress scenario. - Sources of sudden cash outflows: - Sudden distress with funding due to maturity mismatch; deposit withdrawals or funding market freezes (repos, commercial paper). - Interlinkages between asset market liquidity and funding liquidity: margin calls, increased collateral needs, collateral downgrades, support of SPVs for reputational reasons. - Mechanics and hurdle rates: - If a bank has enough liquid assets it can generate cash by selling assets or using repos without large losses; if assets are non-marketable loans or collateral values decline (haircut), liquidity shortage results. - Possible liquidity-test hurdle rates: number of days the institution can tolerate a liquidity shock before negative cash flow; net cash flow position; stressed liquidity ratios. - Interaction with solvency: - Liquidity and solvency stress events are closely related; funding distress can turn liquidity shortages into solvency problems via fire sales and increased funding costs translating into solvency stress. ### Typology of stress tests (main objectives and features) - Four types based on ultimate objective: - Stress testing as an internal risk management tool: - Used by financial institutions to measure and manage risks from investments (example: J.P. Morgan’s VaR in the mid-1990s). - Microprudential/supervisory stress testing: - Basel II requires banks to conduct stress tests for market risk and, in some cases, credit risk as part of Pillar 1; Pillar 2 allows supervisors to order additional tests or actions. - Increasing use to set capital requirements, determine explicit capital buffers, or limit capital distributions. - Liquidity ratios in Basel III and insurance regulation in Europe (Solvency II) utilize stress testing. - Macroprudential/surveillance stress testing: - Used to unveil systemic risk and vulnerability in surveillance and system-wide monitoring; results often reported in Financial Stability Reports. - IMF includes stress testing in FSAPs since 1999; some countries indicated FSAP stress tests were their first exercises. - Crisis management stress testing: - Used to assess recapitalization needs of key institutions, possibly with public support (examples: U.S. SCAP; CEBS/EBA exercises in 2010 and 2011). - Follow-up is almost certain; tests designed to estimate capital shortfalls. - Evolution and scope: - Risk coverage/methodologies broadened from single factor to macro scenario tests and from institution-specific to enterprise-wide risk management. - Macroprudential tests typically do not prescribe bank-specific action but can lead to macroprudential policy recommendations. - Supervisory stress tests increasingly guide actions: improve data collection, target examinations, closer monitoring, require bank management actions (raise capital, reduce exposures, cap dividends, update resolution plans). ### Top-down (TD) versus Bottom-up (BU) tests - Definitions: - Top-down (TD): conducted by national authorities or Fund staff using bank-by-bank data with a consistent methodology and assumptions (typical in FSAPs). - Bottom-up (BU): carried out by individual financial institutions using their internal data and models, often under common assumptions. - Strengths of BU: - Reflects granular data and covers exposures and risk-mitigating tools comprehensively, including complex structured products, hedging strategies, and counterparty risks. - Utilizes advanced internal models; may reveal risks missed by TD and provides insight into risk management capacity and culture. - Strengths of TD: - Ensures uniformity in methodology and consistency of assumptions; effective tool to validate BU tests; resource-effective once core framework is in place; feasible where institutions have limited risk management capacities. - Weaknesses of BU: - Resource intensive; depends on cooperation; results may be influenced by institution-specific assumptions that hamper cross-institution comparisons. - Weaknesses of TD: - Estimates may lack precision due to data limitations; standardization may not reflect institution-specific strategic and managerial decisions. - Common practice: - Liquidity tests often BU because they require granular data and depend on bank liquidity strategies. - FSAPs almost always include a TD test, frequently supplemented by a BU test. - Many national authorities use both TD and BU and emphasize running TD tests to validate BU results. ### Communication, follow-up, and publication practices - Follow-up actions: - Depending on objective, follow-up managerial or supervisory actions may be taken. - Macroprudential tests typically do not prescribe bank-specific action but can lead to macroprudential policy recommendations. - Supervisory stress tests guide supervisory action and may require management actions. - Crisis management stress tests almost certainly lead to follow-up (recapitalization, restructuring). - Communication practices: - Close communication between banks and supervisors, and among supervisory agencies or FSAP teams, is required for effective stress tests. - Public communication of stress test results is not common but has increased, especially for crisis management stress tests. - Macroprudential/surveillance stress test results typically reported in Financial Stability Reports or Financial System Stability Assessments (FSSAs) in FSAPs—often in aggregated form without identifying individual institutions. - Publication practices are controversial: authorities worry public dissemination may create unrealistic expectations, media misinterpretation, and distract banks from supervisory purposes. - For FSAPs, publication of FSSAs is voluntary; the majority of countries publish the document. The accompanying stress testing Technical Notes are published much less frequently. *Source: _082212 - 10. Stress tests typically evaluate two aspects of financial institutions’ performance:* ### 16. Stress tests use a wide variety of analytical models that fall broadly into two ### Stress tests use a wide variety of analytical models that fall broadly into two categories ### Overview of model families - Two broad families of stress testing models: - Balance sheet-based (sometimes called “fundamental” approaches): predicated on detailed analysis of balance sheets of individual institutions. - Market price-based approaches: based on summary default measures for individual portfolios, institutions, or entire systems embedded in market prices (stocks, bonds, derivatives). ### Strengths and weaknesses of the two families - Balance sheet-based approaches - Strengths: - Can identify the source of individual vulnerabilities in the balance sheet. - More informative for pinpointing types of risk (e.g., credit losses from housing loans, market valuation losses from exposures to sovereigns, losses from currency mismatches). - Applicable to emerging and low income countries where stock markets are thin or illiquid. - Possible to adjust for supervisory weakness (e.g., under-provisioning, forbearance). - Weaknesses: - Backward-looking, data intensive, hard to update frequently. - Not well-suited for capturing interdependence (portfolio) and contagion effects across institutions. - Face severe limitations capturing all risks in an integrated manner; often focus overwhelming attention on some risks at expense of others (sovereign default risk, downgrading risk, counterparty risks of derivatives, liquidity risks). - Data intensive, especially for network analysis; quality depends on granularity and availability of data. - Market price-based approaches - Strengths: - More flexible, can easily incorporate portfolio effects and risk factors as perceived by the market. - Can be updated as frequently as desired (daily or lower frequency). - Focuses on systemic risks/losses and tail events; incorporates risk factors priced by the market. - Less data-intensive than accounting-based approach. - Weaknesses: - Difficult to disentangle precise source of vulnerabilities (“black box”). - Sensitive to short-term swings in market perceptions that may be unrelated to fundamentals; estimated vulnerability measures may be very volatile during periods of significant market stress. - Cannot be applied to countries or entities with limited or no market price data (mutuals, privately-held, or government-owned companies). ### Application in emerging and developing economies - Model choice constraints: - Limited data or weak accounting and regulatory systems constrain model choices. - Simpler financial and economic systems mean key risks and vulnerabilities are relatively straightforward to identify. - Feasible approaches: - Simple balance sheet-based tests (as illustrated in Čihák, 2007) with single or multi-factor shocks can be implemented in most countries using basic supervisory data. - When supervisory data are patchy or unreliable, priority should be to improve supervision and data availability rather than develop advanced stress testing approaches. - Developing macro-financial linkage models for macro scenario tests can be challenging given limited long time series; one option is to utilize cross-country data. - Footnote example: Annex 1.6 in the April 2010 GFSR shows that an NPL projection model using data from Latin America and emerging Asia performs reasonably well in predicting loan quality in Central and Eastern European countries, where time series are relatively short. ### Box 1 summary: Balance Sheet-Based and Market Price-Based Approaches - Balance sheet-based models: - Spell out all (on- and off-balance sheet) positions and the risks to which these are exposed. - Typically used by financial stability/supervisory authorities since they need disaggregated information to adopt risk-mitigating measures. - Market price-based models: - Based on summary bank default measures embedded in asset prices; measures extracted by solving for default probability implicit in prices using standard pricing formulas. - IMF often uses the Contingent Claims Approach (CCA), a market price-based approach built on standard option pricing model (Gray, 2008). - For stress tests, market-based default measures need to be projected for the period covered by the test. - Complementarity: - While data limitations may dictate use of one approach, the methodologies do not convey the same type of information and should be considered complementary rather than substitutes. - Integrated risks and common challenges: - Balance sheet-based models struggle to value all positions and assess all risks in an integrated manner; methodologies to value all positions are being developed but are data-intensive and yet to be refined. - Common challenge for both types: stress testing individual institutions to system-wide risks requires assessing joint entity default probabilities and measures of default dependence to produce systemic loss distributions. - Balance sheet-based extension: superimpose a network of claims to track default effects across institutions. - Market price-based approach: treat system as a portfolio of banks and derive distribution of systemic losses using portfolio analysis techniques. ### Table 3 comparative operational aspects (selected points) - Primary input data: - Balance Sheet-Based: Accounting data (balance sheet, profit and loss account, matrix of interbank exposures). - Market Price-Based: Financial market data (equity prices, bond yields, CDS spreads or equity option-based probability of distress of banks). - Secondary input data: - Balance Sheet-Based: Probability of distress (and loss-given-default) or NPL ratios/loan classification of borrowers (for credit risk); market data to calibrate shocks. - Market Price-Based: Balance sheet data (combination of equity prices and accounting data obtain key input variables, such as the Expected Default Frequency by Moody’s). - Type of test: - Balance Sheet-Based: Solvency, liquidity and network analyses. - Market Price-Based: Largely focused on solvency and its interdependence among key financial institutions; initial attempts at liquidity stress testing. - Frequency: - Balance Sheet-Based: Varies depending on reporting cycle (quarterly, semi-annual, annual). - Market Price-Based: Daily or lower frequency. - Application: - Balance Sheet-Based: Most banks or financial systems (including emerging markets and low income countries) as long as financial reporting or supervisory data exists and is available. - Market Price-Based: Limited to market data-rich countries and institutions quoted on the market; generally cannot cover mutuals, privately-held, or government-owned companies. - Links to macro scenarios: - Both approaches: Possible by estimating additional macro-financial model(s), linking macro scenario variables and risk factors (PDs, NPL ratios, volatility, leverage, etc.). - Estimation of systemic effects: - Balance Sheet-Based: Consider common macro shocks across banks (e.g., GDP, inflation) and incorporate network effects (interbank exposures). - Market Price-Based: Consider common macro shocks and incorporate interdependence (portfolio) effects among banks, estimated using asset prices. - Output examples: - Balance Sheet-Based: Various capital ratios, liquidity ratios, capital shortfalls, number/share of banks breaching minimum requirements. - Market Price-Based: Expected losses, unexpected or tail losses, contingent liabilities for the government, probabilities of spillover among banks. - Selected examples of methodologies referenced: - Balance Sheet-Based: Stress tester 101 (Cihak, 2007); Next generation stress testing (Schmieder, Puhr and Hasan, 2011); Network analysis (Espinoza-Vega and Sole, 2010). - Market Price-Based: CreditRisk+ (Credit Suisse Financial Products); Systemic CCA (Gray and Jobst, 2010); CoVaR (Adrian and Brunnermeier, 2008); Distress dependence (Segoviano and Goodhart, 2009); SES and MES (Acharya et al. 2010). ### Recent developments and institutional activity - The area is very dynamic, with model refinements and new methodological approaches constantly being developed. - Fund staff are active in this area, in close cooperation with advanced economy central banks, financial stability agencies, and academia. - The Expert Forum on Advanced Stress Testing Techniques, organized by the IMF’s Monetary and Capital Markets Department (MCM) and a cooperating central bank annually, is a pre-eminent forum for practitioners. ### Lessons from the global financial crisis and the European sovereign debt crisis - Major impacts on stress testing design, conduct, and use: - Pre-crisis stress tests by the financial industry, country authorities, and FSAPs failed to detect key vulnerabilities (Haldane 2009a and 2009b). - Post-crisis: increased use of stress tests for crisis management (U.S. and Europe) with mixed results. - Move toward greater transparency in disseminating stress test results, though controversies remain. - Why pre-crisis stress tests failed (selected reasons): - Institutional perimeter too narrow: did not cover “shadow banking” (money market funds, insurance companies writing credit insurance). - Key shock transmission and propagation channels not covered: interconnectedness and second-round feedback effects between financial sector and real economy and/or sovereign risk were not incorporated. - Some key risk factors were missed: counterparty risk, basis risk, contingent risks. - Balance sheet valuations did not fully reflect economic value: regulatory/accounting norms overestimated resilience; regulatory capital may not value every security at market values. - Shocks were not severe enough and did not examine truly tail events: scenarios were sometimes too benign; data series for some instruments (CDOs, structured products) too short; time-varying correlation and extreme market risks not well reflected. - Many shocks considered “unthinkable” were excluded: severe liquidity risks involving seizure of key funding markets and sovereign default risk in advanced economies were often not covered. - Ongoing agenda: - Some lessons have been incorporated (scenario types and severity), but challenges remain with incorporating all relevant risk propagation channels and feedback effects between financial sector and real economy. - Appendix IV (referenced) presents emerging methodological approaches attempting to tackle these challenges. ### “Best practice” principles for stress testing (proposed) - Seven principles proposed (mainly focused on stress tests for macroprudential surveillance): - Define appropriately the institutional perimeter for the tests. - Identify all relevant channels of risk propagation. - Include all material risks and buffers. - Make use of the investors’ viewpoint in the design of stress tests. - Focus on tail risks. - When communicating stress test results, speak smarter, not just louder. - Beware of the “black swan.” - Purpose of principles: - Provide operational guidance across diverse country and sector circumstances while maintaining minimum standards to enhance comparability (important for FSAPs). - Protect against obvious pitfalls highlighted by recent crises. - Allow informed observers to critically evaluate stress testing exercises. - Recognize that stress testing is as much an art as a science. ### Principle No. 1: Define appropriately the institutional perimeter for the tests - Objective: select institutions to include in system-wide tests, assessing which banks are systemically important (capable of triggering or amplifying systemic risk). - Criteria used to assess systemic importance: - Size. - Substitutability. - Complexity. - Interconnectedness. - Rationale: - A bank’s distress or failure is more likely to cause damage to other banks, markets, or the economy if its activities comprise a large share of financial intermediation. - A higher network of contractual obligations increases likelihood that failure will materially raise distress at other institutions. - Lower substitutability (where specific institutions provide critical services or market infrastructure) increases systemic impact. - More complex institutions have higher systemic impact because costs and time needed to resolve them are greater. *Prepared by IMF staff; chapter text as provided in the source PDF.* ### 28. While size, degree of substitutability, and complexity are observable features, an ### While size, degree of substitutability, and complexity are observable features, an assessment of interconnectedness requires the use of sophisticated network approaches ### Network approaches to measuring interconnectedness and informing stress tests - Network approaches model financial institutions’ bilateral claims as a matrix and can identify systemically important institutions or groups of institutions that occupy central positions in the network. - Network models can directly measure the likelihood of multiple defaults by systemically important institutions—addressing systemic risk that would be underestimated by stress testing individual institutions and aggregating results (see Principle 5). - Simple network models add contagion channels to stress testing by modeling direct default on borrowings and the indirect channel of a failed bank ceasing to lend (liquidity shock). - Advanced network models examine four measures of interconnectedness to identify key nodes: - (1) “in-degree” = number of links that point to a particular node; - (2) “closeness” = inverse of the average distance from one node to others; - (3) “betweenness” = focus on the shortest path between nodes; - (4) “prestige” = assigns increasing scores to nodes connected to other high-scoring nodes. - Cluster analysis separates the network into subgroups (clusters) with closer internal connections and can identify “gatekeeper” institutions bridging clusters. - Applications cited: - Fund work used cluster analysis to identify the 25 jurisdictions with the most systemically important financial sectors. - Reserve Bank of India used network techniques to identify core institutions including banks and nonbanks and to show contagion paths where failure of one bank could affect ten other institutions, including three insurance companies. - Network models can help set the appropriate perimeter for stress testing and can be integrated ad hoc (triggering randomly selected failing banks) or linked with macro stress testing (banks failing solvency/liquidity tests become trigger banks). ### Defining the perimeter of stress tests: group activities, cross-border and nonbank entities - Large financial conglomerates can include banking, insurance, pension fund, investment fund, various financial SPVs, and non-financial activities; ultimate ownership and economic links may not be clear. - Some SPVs are legally independent but may imply contingent liabilities for parent companies if parents provide support beyond legal obligations for reputational reasons. - Home-country analyses often focus on domestic activities, but cross-border exposures are relevant for inward and outward spillover assessment. - Host-country analyses may omit parent-company health from stress tests due to limited group-wide information; stress testers must use judgment on consolidation versus segregation of activities. - Inclusion of nonbank institutions: - FMIs (payments systems, central securities depositories, securities settlement systems, central counterparties) are systemically important because of the critical services they provide; stress testing for FMIs emphasizes safe and reliable functioning rather than balance sheet size. - Stress tests of insurance companies are increasingly common; while traditional insurance activities are unlikely to trigger systemic risk, insurers can create systemic risk via complex transactions and links with banks. - A network approach integrating all financial institutions reduces the chance of omitting potential systemic risk sources. ### Empirical practices and coverage in stress testing - FSAP stress tests typically cover at least seventy percent of the locally-incorporated commercial banks (including subsidiaries of foreign banks and state-owned banks) by assets. - Country authorities’ exercises focus on private domestically-owned commercial banks, followed by foreign subsidiaries and state-owned banks. - Survey-based coverage ranges from 60 to 100 percent of the system by assets (with the median being 85 percent and 16 banks). - The number of banks in stress test samples varies from below 5 to over 1,000. - When only a subset is included, perimeter methodology typically prioritizes size and interconnectedness, followed by size of local retail activities and legislative definitions. - Formal network approaches are increasingly used, though sophisticated models are still used in a minority of cases, even in advanced economies. - Inclusion of nonbank sectors: - FSAPs always test the banking sector and occasionally the insurance sector; other sectors are rarely tested. - Among country authorities, about 40 percent indicate they only test the banking sector and another 45 percent also test the insurance sector. - Tests of pension funds and FMIs are undertaken on a much more ad hoc basis, if at all. - Examples: U.S. SCAP and CCAR included a life insurance company (Metlife) and the former auto loan arm of General Motors (Ally), in addition to investment and commercial banks (these companies operate with bank holding company licenses). - Some central banks simulate operational disruptions of FMI to identify recovery times, critical participants, and contingency measures; CCP stress tests typically frame outcomes in terms of the number of participant defaults a CCP can withstand. ### Principle No. 2 — Identify all relevant channels of risk propagation - Key transmission channels to consider include: - Feedback between liquidity and solvency risks (non-linear relation between funding costs and creditor risk perception; asset sales/fire sales and ensuing declines in asset values aggravating solvency). - Feedback from financial instability to the real economy (bank reactions such as deleveraging and capital flight can degrade bank asset values further — second-round effects). - Bank-sovereign linkages (more complex than fiscal contingent liabilities alone; see Appendix IV). - Policy feedbacks (policy actions or inaction materially affect risk transmission and crisis duration). - Proper stress test design requires careful examination and calibration of these transmission channels and understanding potential responses of financial institutions and capital markets to shocks. ### Gaps in practice and implementation challenges - Incorporating interactions between real and financial sectors remains a major operational challenge; calibration of propagation channels often relies on historical information or expert judgment and must contend with incomplete information and tail risk from “unknown unknowns.” - Liquidity and solvency feedbacks: most surveyed country authorities and almost all FSAPs incorporate liquidity shocks (deposit withdrawals, reductions in interbank positions, haircuts for liquid assets); a few account for off-balance liquidity needs and wholesale funding withdrawals. - About half of respondents consider domestic currency and foreign currency liquidity positions separately. - Most liquidity stress tests are implemented independently of solvency tests, excluding the intensified runs when banks sustain substantial losses (examples: Bear Stearns; some banks in the European periphery). - Large European banks compute “liquidity-at-risk” or “liquidity value at risk” using stochastic approaches, which do not adequately incorporate traditional credit risk or links to macroeconomic scenarios. - Feedbacks from financial sector distress to the macroeconomy and complex bank-sovereign linkages are rarely incorporated in FSAPs or country authorities’ stress tests. - In some IMF programs, fiscal-financial linkages are modeled by adding bank recapitalization needs from stress tests to fiscal debt sustainability analysis as contingent liabilities, or by using market-price based approaches. - Linkages such as the effect of a distressed, deleveraging banking sector on economic growth through credit crunch are not captured. - The field requires a new generation of macroeconomic models including the financial sector and intermediation; development is underway but sufficiently operational models are not yet broadly available. - Examples of advances: GFSR (IMF 2011b) presented a DSGE model with a banking sector; Bank of Japan (Ishikawa et al., 2012) developed an econometric model incorporating financial–macro interactions; IMF Research Department is expanding GIMF to explicitly include the financial sector. - In practice, some effects are integrated ad hoc (e.g., adding a second-round top-down stress test to bottom-up bank-by-bank tests), and financial institutions may provide qualitative descriptions of likely responses (portfolio allocation, deleveraging, securing central bank credit lines) to inform design of second-round shocks. *Source: _082212 - 28. While size, degree of substitutability, and complexity are observable features, an* ### Box 3. Integrating Liquidity and Solvency Risks and Bank Reactions in Stress Tests* ### Box 3. Integrating Liquidity and Solvency Risks and Bank Reactions in Stress Tests ### Banks’ reactions to credit and funding shocks - First lines of defense: high-quality capital and retained earnings (pre-impairment income). - Banks can generate liquidity by using high-quality eligible securities as collateral for market or central bank funding; fire sales are an option but costly in sharply declining asset-price environments. - Deleveraging (targeting high risk-weight assets) reduces risk-weighted assets (RWAs) and raises capital adequacy ratios. - Hybrid measures observed in practice include debt-to-equity conversions, issuance of convertible bonds, and optimizing RWAs. - Integrating bank reactions is critical for informative stress tests, especially over longer horizons, because: - The source of shocks (solvency vs. liquidity) may be unclear when banks react. - Reaction measures have both capital and liquidity aspects that are hard to disentangle. ### Two-round stress-testing (practical, ad hoc approach) - Conduct a bottom-up (BU) first round and a top-down (TD) second round. - Use BU-reported behavioral reactions to calibrate TD second-round impositions: - If many banks report asset sales of particular asset classes, impose haircuts on those assets in the TD round. - If banks report discontinuing reverse repos, incorporate reduced repo roll-over rates. - Quantification of haircuts or roll-over rates can be based on historical information, cross-country experience, or expert judgment. ### Analytical approaches that integrate solvency and liquidity - Schmieder et al. (2012): simulate increase in funding costs resulting from a change in solvency indicated by a change in a bank’s (implied) rating. - Dutch Central Bank model: endogenizes market and funding liquidity risk by including feedback mechanisms capturing behavioral and reputational effects; used by several central banks and supervisors. - Hong Kong Monetary Authority framework: links default risk and deposit outflows; simulates mark-to-market losses → deposit outflows → asset fire sales → sharp increase in contingent liquidity risk. - Barnhill & Schumacher (2011): general empirical model incorporating previous approaches to be more comprehensive about solvency shock sources and compute longer-term impact of funding shocks. - Risk Assessment Model for Systemic Institutions (RAMSI), Bank of England: simulates banks’ liquidity positions conditional on capitalization under stress and other dimensions such as decreases in market confidence. - Increasing use of network analysis and market-based systemic risk measures to capture spillovers and interconnectedness (e.g., Luxembourg FSAP network effects, Bank of England RAMSI adding contagion via network models, market-based reduced-form approaches such as the 2010 U.S. FSAP). *Prepared by Heiko Hesse.* ### Principle No. 3 — Include all material risks and buffers - Pre-crisis focus was on credit risk from customer loans and market risk from marketable securities; crisis highlighted need to include sovereign, funding, systemic liquidity, and counterparty risks. - For globally active institutions, incorporate cross-border activities: cross-ownership, credit and market exposures, and funding (including parent-subsidiary funding and liquidity transfers). - Controversies and challenges: - Some risks (e.g., own sovereign) can be so large and hard to hedge that protecting against them via additional capital may be infeasible. - System-wide liquidity shocks raise questions about the usefulness of stress testing those risks. - Interest rate risks (refinancing and reinvestment) are often neglected in low-stable inflation/rate environments. - Buffers such as pre-impairment income must be included, especially in multi-year stress tests where profits materially affect results. - Changes in RWA can substantially affect stress test outcomes; proper modeling of RWA changes is important. - Bottom-up tests may be preferable because banks have better portfolio knowledge, but careful supervisory scrutiny of bank models and assumptions is essential. ### Gaps and developments in risk coverage - Post-crisis expansion of included risks: liquidity and funding, sovereign risk, low profitability, regulation-related risks, and contagion from interconnectedness. - Sovereign risks in tests often modeled as market risks via haircuts on sovereign securities, with methodological differences in coverage and shock sizing. - FSAPs in emerging/developing economies attempt to quantify supervisory weaknesses (under-provisioning, loan misclassification, collateral overvaluation, forbearance, concentration risk) via judgmental adjustments that require cross-checking with supervisory expertise. - Ensuring methodological consistency for pre-impairment income and RWA projections remains challenging across institutions and jurisdictions. - Persistently neglected risks in system-wide tests include downgrades (capital buffers for downgrading risk) and interest rate risks in the banking book. - Cross-border exposures pose data and coordination challenges; some FSAPs rely on bottom-up exercises and the IMF’s global macro framework for global scenarios. ### Box 4 — Projection of pre-impairment income under adverse scenarios - Pre-impairment income (profit before loan/security impairment losses) contributes to capital and can be substantial relative to shock-related losses over multi-year horizons. - No single agreed methodology; common techniques include estimating elasticities and projecting subcomponents. - Elasticity estimates (Hardy and Schmieder, forthcoming): - Elasticity of pre-impairment income-to-capital to economic growth ranges from 1 to 1.5. - Average pre-impairment income-to-capital ratio in both advanced and emerging economies is about 12 to 15 percent. - A moderate stress event (a 4 percentage point drop in annual GDP growth in advanced economies and a 6 percentage point drop in emerging economies) would reduce pre-impairment income-to-capital by about half. - For banks in advanced economies, the elasticity rises to 4 under severe stress conditions, implying a strongly non-linear relation between macro conditions and bank profits. - Component-based projections: - Key subcomponents: net interest income, fee and commission income, non-interest expenses (including salaries). - Net interest income is more closely linked to macro conditions than other sources, except trading income under highly adverse conditions. - Fees/commissions should be linked to their sources (sales/trading → asset prices; loan origination/credit cards → credit growth or employment). - Interactions between solvency and liquidity: - Suspected solvency losses can trigger depositor and wholesale funding withdrawals, forcing banks to pay higher funding rates. - Higher funding costs, if not passed through to borrowers, can reduce net interest income. - Wholesale funding costs are more elastic than deposit costs and harder to pass on in weak conditions. - Structural models for credit spreads (similar to Moody’s KMV) can link bank funding cost to solvency distress (Schmieder et al. (2012) and other IMF exercises). - Funding-cost increases are relatively contained for well-capitalized banks but rise sharply near regulatory minimums; depending on pass-through, net interest income can shrink substantially. *Prepared by Christian Schmieder and Hiroko Oura.* ### Principle No. 4 — Make use of the investors’ viewpoint - Market perceptions of solvency and asset values critically affect bank funding and viability; reliance on uninsured short-term wholesale funding can expose banks to investor-driven confidence shocks. - Market discipline can lead to higher funding costs or complete withdrawal of funding, causing higher losses, deleveraging, and second-round impairments. - Operational implications — complement regulatory/accounting-based stress tests with market views: - Adopt mark-to-market (MTM) valuation for all assets and liabilities under baseline and adverse scenarios to reflect investor valuation (may require ignoring HTM amortized-cost accounting for stress benchmarking). - Use economic capital rather than statutory capital to capture underlying economic value (e.g., unrealized losses on securities and inadequate provisioning when collateral values fall). - Use point-in-time (PIT) parameters for expected and unexpected losses rather than through-the-cycle (TTC) parameters to reflect investors’ assessments. - Explicitly stress market risk appetite—e.g., adverse scenarios of 2008–09 magnitude could incorporate a similarly large increase in market price of risk, leading to higher credit spreads/funding costs and lower equity values. - Impose market-based hurdle rates (targeted funding-cost based) in addition to regulatory minima to reflect market-required capital/rating targets; otherwise banks that pass regulatory stress tests may fail market tests. - For supervisory authorities and central banks with better access to bank-specific and market information, identifying the market’s perceived risk-return trade-off should be simpler. *Prepared by Heiko Hesse.* *Source: Box 3 and related boxes from the supplied IMF content unit.* ### 45. Imposing hurdle rates that may be higher than regulatory minima is particularly ### _082212 - 45. Imposing hurdle rates that may be higher than regulatory minima is particularly ### Macroprudential rationale for higher hurdle rates - Financial institutions must be sufficiently capitalized not only to ensure their own viability in the event of a system-wide shock but also to prevent them from becoming transmission channels. - This may require higher capital than is necessary on a stand-alone basis (OFR, 2012). ### Dissemination of stress test results - Publishing stress test results can: - Remove asymmetric information during periods of uncertainty and restore market confidence. - Create awareness of risks, promote more realistic risk pricing, and enhance market discipline during good times, potentially reducing probability of sudden investor reversals. - Preconditions for publication to yield benefits: - Stress tests must be candid assessments of risk and explicit about coverage and limitations. - Announcement of results must be accompanied by measures that convincingly redress vulnerabilities unveiled by the stress tests, including but not necessary limited to capital injections. ### Current practices and shortcomings - Market-based hurdle rates: - Generally not used by country authorities, and only rarely used in FSAPs (U.S. FSAP 2010, IMF 2011a). - Funding costs are increasingly being modeled and incorporated explicitly in stress tests, but this alone is insufficient because the relationship between funding costs and bank solvency is non-linear. - Some market-related solvency thresholds can and should be used as hurdle rates. - Sovereign risk assessment: - Market considerations are generally not incorporated in stress tests. - Regulatory and accounting standards could understate valuation risks for sovereign securities held in HTM accounts. - Some FSAPs include unrealized losses from securities in HTM accounts by applying mark-to-market valuations in response to yield changes, recognizing economic rather than accounting valuation matters for sustainability. - The EBA chose not to account for unrealized valuation losses from sovereign securities in HTM accounts in their 2010 and 2011 exercises, opting to disclose exposure details instead. - The Bank of Japan regularly tests interest rate shocks on all securities, including those in HTM accounts. - Adjusting initial capital for already realized shocks: - EBA’s capital exercise of end-2011 required banks to mark-to-market all sovereign bond holdings incorporating large 2011 yield changes. - Blackrock exercises for Ireland and Greece included re-assessment of existing loans via a re-underwriting process. - Surveillance stress tests often do not include these adjustments. - On a positive note: - Recent FSAPs and EBA exercises are ensuring the use of point-in-time PDs. ### Box 5 — Market-Based Hurdle Rates (summary) - Market-based hurdle rates complement regulatory hurdle rates: capital ratios compatible with stable funding costs. - Methodology examples: - Schmieder et al. (2012) derive a non-linear relationship between solvency (Basel II IRB-implied capitalization ratios) and funding costs for a sample of German banks using 12 quarters during 2007–2009. - Funding costs measured as excess spread above government bond rate; German banks’ weighted average MKMV’s Expected Default Frequency (EDF) plotted against funding costs; EDFs translated into capitalization ratios using the IRB formula, including an additional capital cushion of 2.5 percentage points. - Key quantitative findings reported: - A regulatory capital ratio of about 15 percent would be a good benchmark. - According to the authors’ calculations, this would be equivalent to a Basel III Core Tier 1 capital ratio of about 11 percent. - Below this level, banks become very sensitive to funding costs as the relationship becomes highly non-linear: - The spread increases by about 50 bps when capital ratios go from 9.9 percent to 8.5 percent. - The spread increases by 184 bps when capital ratios are reduced further down to 7.1 percent. - Savings in funding costs above 15 percent appear small. - Contingent Claims Analysis (CCA): - Recovers probability of default implied in bank stock prices, market value of bank assets, and market value of risky debt. - Can be used to construct a CCA capital ratio (market capitalization to market value of assets) mapped onto creditor-required spreads. - Results for a sample of French banks show a non-linear relationship useful to determine a market-based hurdle rate. ### Principle No. 5 — Focus on Tail Risks - Traditional rule of thumb: apply “extreme but plausible” shocks, but no systematic way to determine them. - Typical calibrations: “worst-in-a-decade”, “one percent probability” tail events, or “x standard deviation” shocks. - Problems with historical calibration: - Historical experience varies across systems and changes over time. - Short history of some financial products (e.g., asset-backed securities) leads to shocks that are too mild. - Historical approaches fail for unprecedented events. - In crisis or near-crisis situations: - Institutions/systems already distressed; authorities may be reluctant to apply excessively negative tail scenarios. - Publishing extreme-scenario outcomes could trigger self-fulfilling crises. - Political economy factors may influence scenario choice when results affect closure, rescue, recapitalization, or rescue packages. - Effective stress test in near-crisis should not compromise on scenario severity but should mitigate adverse market impact by having credible support measures in place; if not feasible, consider withholding stress tests and releasing critical exposure data with qualitative analysis. - Interdependence among shocks and institutions: - Modest but correlated shocks can generate extreme outcomes if dependence is ignored. - Risk factors may be weakly correlated in normal times but highly correlated in distress (excess skewness, excess kurtosis). - Measuring non-linear dependencies is necessary to capture joint tail risks. - Approaches to capture dependencies: - Balance sheet-based models: use network models to measure spillovers and probability of multiple defaults; most conventional models do not formally account for default dependencies; identifying networks requires substantial information. - Market price-based models: possible to model distribution of joint default probabilities; key challenge is modeling dependence structure. - Conditional Value-at-Risk (CoVaR) measures VaR of an institution conditional on distress of others using equity or market price data. - Marginal Expected Shortfall (MES) specifies historical expected losses conditional on breaching a systemic risk threshold. - Adjusting MES by firm-specific leverage and capitalization yields Systemic Expected Shortfall (SES), measuring average linear bivariate dependence when the sector is undercapitalized. - More complex approaches require significant departures from conventional statistical methods (see Box 6). ### Evolving practices on tail risks and interdependence - Recent FSAPs: - Systematically include severe shocks comparable across countries. - A common rule of thumb in many FSAPs: assume a two-standard deviation shock on GDP growth rate for two years based on a long (20–30 years) history, unless a different magnitude is justified. - Among country authorities: - Most survey respondents adopt tail events with small probability (ranging from 1–5 percent); some use qualitative criteria (e.g., in line with, or worse than, historical worst). - Using samples that include 2008/09 crisis observations should generate fairly severe shocks. - Another frequent choice is testing a shock of similar magnitude as the recent crisis. - Historical-data approaches carry complacency risk; manage via qualitative approaches (Principle 6). - Examples of variant practices: - U.S. SCAP: macro scenario moderate compared to actual turnout, but corresponding rise in loan loss rate comparable to the Great Depression. - EBA exercises: assumed only moderate deterioration in sovereign yields, quickly exceeded by market developments; 2011 exercise required revised assumptions mid-exercise. - 2010 U.S. FSAP: examined a moderate shock on top of the baseline, noting it would push output gap to one of the lowest post-war levels. - 2012 Spain FSAP: complemented quantitative stress testing with qualitative assessments. - Incorporating risk interdependence remains technically challenging: - Some authorities capture interdependence heuristically via common macro/financial shocks and interbank contagion network models. - The Fund has increasingly used network models standalone or as part of stress testing. - Recent FSAPs in Germany, Sweden, the United Kingdom, and the United States combined institution-level stress tests with system-wide capital assessments controlling for dependence structure and stochastic input parameters. - Technical limitations mean mainstreaming these approaches in FSAPs will take time. ### Box 6 — The Importance of Risk Interdependence (summary) - Traditional correlation measures detect only linear dependence under normality assumptions; alternative measures are needed for non-linear dependence and joint tail behavior. - Methods and concepts: - Refined quantile-based Chi-square statistic of independence to detect shared extremes. - Copula functions and other non-parametric methods to combine distributions with flexible dependence structures, capturing “joint asymptotic tail dependence.” - Extreme Value Theory (EVT) methods (e.g., generalized Pareto distribution, GEV) define limiting behavior of extreme observations and model exceedances beyond thresholds. - Systemic CCA framework: - Extends contingent claims analysis to measure systemic risk from interlinkages between risk-adjusted balance sheets. - Applies option-pricing theory to market information, explicitly acknowledging non-linearities in default risk measurement. - Employs EVT to generate a multivariate extreme value distribution (MGEV) capturing tail realizations of market-implied joint potential losses. - Estimation process: - Estimate non-parametric dependence function of individual expected losses. - Combine with marginal distributions assumed to conform to a GEV distribution; marginal distributions estimated via the Linear Combination of Ratios of Spacings method. - Dependence function estimated iteratively on a unit simplex to optimize coincidence of cross-classified extremes. - Conditional Value-at-Risk estimate of joint expected losses determined as probability-weighted residual density beyond a pre-specified confidence level. - Differs from classical multivariate approaches by separating marginal distribution analysis from dependence analysis and capturing both linear and non-linear joint asymptotic tail behavior. *Source: IMF staff discussion in PDF chapter _082212 - 45. Imposing hurdle rates that may be higher than regulatory minima is particularly* ### Box 6 (contd.). The Importance of Risk Dependence in System-Wide Stress Testing ### Box 6 (contd.). The Importance of Risk Dependence in System-Wide Stress Testing ### Implied Capital Requirement under “Distributional Approaches” - Incorporating time-varying extreme dependence of risk factors that determine system-wide default risk—ideally together with a market-derived measure of capital adequacy—offers a more realistic capital assessment in stress tests. - This approach identifies the amount of capital shortfall to current capital levels based on market expectations on solvency that far exceed the minimum regulatory requirements. - Assuming higher uncertainty about the realization of expected losses (i.e., greater historical volatility) reflects that, especially during distress periods, firms would need higher capital buffers to absorb losses above existing provisioning levels so that current capital levels remain unaffected over a specific risk horizon. - Elements shown in the aggregate loss distribution illustration: - New Expected Loss (after shock) - VaR 95% - Probability - Extreme Loss - VaR 99% - Time - aggregate loss distribution at time T (after shock) - Mean - additional regulatory capital - additional economic capital ### Principle No. 6: When Communicating Stress Test Results, Speak Smarter, Not Just Louder - Public communication of stress test results has increased since the global crisis; disclosure practices varied in coverage, assumptions, and level of detail. - FSSAs—usually published with country authorities’ consent—include stress test assumptions and results on an aggregate basis or preserving anonymity of institutions. - The U.S. and Europe saw unprecedented disclosure in the aftermath of the crisis; in the U.S. this was enshrined in law. - Comparative outcomes: - The U.S. SCAP restored confidence, allowed investors to differentiate banks by resilience, and arguably facilitated raising additional private capital. - The 2011 EU-wide stress test did not fully achieve its goals despite transparency; weaknesses included perceived inadequate severity, incomplete capture of weaker national banking systems’ risk profiles, and ambiguous follow-up actions and policy backstops. - Tradeoffs in disclosure: - Benefits: raises public awareness of risks; promotes realistic risk pricing; strengthens market discipline; promotes meaningful conversation about financial stability policies. - Risks: may entice institutions to “game” tests; increase moral hazard if investors over-rely on published results; undermine confidence if support measures are not in place; cause confusion when multiple exercises produce differing results. - Guidance on when to disclose: - In the midst of a crisis, the case for public disclosure of details of system-wide stress tests is compelling; in normal times the case is more finely balanced. - Disclosure of system-wide tests (macroprudential or crisis management) is stronger than for microprudential tests because they apply consistent assumptions across institutions and focus on systemic risk factors relevant for market confidence. - Preconditions for beneficial disclosure: - Credible tests that cover relevant risks and transmission channels, assume serious shocks, set appropriate hurdle rates, and produce candid assessments. - Accompaniment by a convincing framework for follow-up action, including government support if needed. - Disclosure should be part of a broader communication strategy on financial stability and complemented with regular outreach and disclosure of a broad set of indicators. ### Survey Findings on Disclosure Practices - Communication rates from survey respondents: - Almost 85 percent communicate the results of macroprudential solvency stress tests. - 50 percent communicate the results of macroprudential liquidity stress tests. - Typical publication practices: - Results usually appear in annual or semi-annual Financial Stability Reports, placed in the context of broader financial stability assessment. - Results are typically communicated using system aggregates, often with some distribution metrics, without disclosing identities of individual institutions—exceptions include recent EBA and U.S. Federal Reserve tests. - FSAP stress test results are reported in FSSAs, typically without revealing individual institution identities. - Publication rate of FSSAs has historically been relatively high at 60–70 percent and has risen in recent years. - FSAP Technical Notes on stress tests, which report more detailed results, are usually not published. - Motivations and concerns: - Objectives cited for publishing: raising public awareness of financial stability issues, achieving greater transparency, and providing information to market participants. - Concerns: risk of exaggerated expectations about stress tests, inconsistent media interpretation, and excessive focus by banks on published results undermining supervisory effectiveness. ### Principle No. 7: Beware of the “Black Swan” - Stress tests cannot predict the future; they measure resilience to assumed shocks but cannot capture truly novel or highly improbable events. - Sources of unforeseeable shocks include new products, unexpected events (e.g., the breakup of currency unions), low-volatility factors, or long-dormant risks (e.g., advanced country sovereign defaults). - Practical approaches to incorporate such factors: - Use expert judgment and current vulnerabilities as guides for scenario choice (e.g., concentrated real estate exposure warrants a large real estate price decline scenario regardless of historical probability). - The Federal Reserve Board practice: use two scenarios—one unique to each institution (chosen by it) and one common—so institutions face scenarios they consider particularly damaging. - Policy trade-offs in shock design: avoid letting weak banks pass tests; avoid designing shocks so large that many sound banks fail. - Reverse stress tests by institutions and surveys of such exercises can extend the frontier of tail risks. - Apply distribution theory to scenarios (scenario distribution approach) rather than choosing a single adverse scenario: - Example: IMF staff used a scenario distribution in the first South Africa FSAP, using statistical properties of historical price-change distributions and Monte Carlo simulations to represent combinations of changes in prices (including credit spreads) and revalue bank assets, producing a distribution of bank capital ratios per bank where each point corresponds to a particular scenario. - Proper context and use: - Stress tests should not be undertaken in isolation; results have margins of error and are subject to model risk, data limitations, and potential underestimation of shock severity. - Complement stress tests with other tools: qualitative and quantitative bank risk analysis, early warning indicators, debt sustainability models, and informed dialogue with supervisors and market participants. - Regularly back-test and re-assess stress test design, models, and implementation; for adverse scenarios consider external expert review; stress the sensitivity of results to key parameters and remain vigilant for new risks. ### V. Conclusions and Operational Implications - The paper proposes operational best practice principles to improve stress test design and implementation, addressing concerns about interpretation and comparability as stress tests take on new, more public roles. - Key messages: - Stress tests are forward-looking tools to assess solvency and liquidity resilience under possible adverse scenarios but do not predict the likelihood of those scenarios. - Critical initial decisions for stress testers include choice of risk scenarios (coverage and severity), test design to cover transmission channels and realistic buffer assumptions, and hurdle rate selection. - Biased or politically influenced choices (modest shocks, optimistic income projections, trivial hurdle rates) can lead to complacency and costly outcomes; public dissemination of such benign results can undermine credibility. - Stress tests’ effectiveness depends on broader context: clear scope and objectives; knowledge of key institutions, business models, and transmission channels; appropriate perimeter and coverage decisions; complementary assessment tools; tailored communications strategies; and credible commitment to remedial measures. - Current implementation gaps: - Adherence to principles is uneven across jurisdictions. - Notable shortcomings exist in identifying channels of risk propagation, using the investors’ viewpoint, and focusing on tail risks—reflecting analytical and implementation gaps that must be closed. *Source: Box 6 (contd.). The Importance of Risk Dependence in System-Wide Stress Testing* ### 69. The table below summarizes key operational implications that flow from each ### Table 4. Practical Implications of “Best Practice” Principles for Stress Testers ### Principle 1. Define appropriately the institutional perimeter for the tests - For system-wide stress tests, if universal coverage is not an option, identify systemically important institutions to cover in the tests, including relevant nonbanks and financial market infrastructures. Criteria for assessing systemic importance should include an assessment of interconnectedness, for example through the use of a network model. - Gain a basic understanding of the structure of financial conglomerates, if present, and cover in the tests any of their banking or non-banking activities that may have a major impact in a stress scenario. ### Principle 2. Identify all relevant channels of risk propagation - Identify and understand the main channels of risk propagation, including but not limited to the relation between solvency and the cost of funding for institutions with large wholesale funding; the risk transfers between banks and the sovereign; potential bank reactions to stress; and feedback effects between financial sector stress and the macroeconomy (e.g., through deleveraging and lower GDP growth). - Incorporate as many of these channels in the design of the stress tests as possible. If modeling them explicitly is not feasible, use heuristic means, such as a second-round top-down test that maps the impact of first-round tests on individual institutions onto key macro variables (asset prices, credit, aggregate liquidity, etc.); or expert judgment, to capture all potential channels of risk transmission. ### Principle 3. Include all material risks and buffers - Understand the tested institution’s—or, for system-wide tests, the systemically important institutions’—business model, including markets where they operate and sectoral or cross-border exposures. - Be as comprehensive as possible in including potential sources of risk in the tests. Do not omit risks because they may be “too big to mitigate.” - Assess whether any banks may be exposed to risks coming from cross-border exposures, non-banking activities (such as employee pension funds, investment funds) or other institutions that may be legally separate but with significant economic links (such as parent companies, subsidiaries, and off-balance sheet vehicles), and include as many of these as possible in the tests. - If stress tests include nonbanking institutions, consider relevant non-economic risks that may have a major impact on them and the rest of the system (such as natural disasters for insurance companies). - In stress tests with longer time horizon, model explicitly likely buffers, such as pre-impairment income. Pre-impairment income should be scenario-dependent and conservatively projected. ### Principle 4. Make use of the investors’ viewpoint in the design of stress tests - Adopt mark-to-market methodologies to the valuation of bank assets and liabilities and point-in-time parameters to measure expected and unexpected losses. - Adjust institutions’ initial capital for supervisory practices or other factors that result in systematic over-reporting of capital, such as underprovisioning, collateral overvaluation, and forbearance. Use economic capital estimates as a complement to statutory capital in the stress tests. - Stress market risk appetite by including explicitly shocks involving increases in the market price of risk. - Use hurdle rates reflecting market views (e.g., target funding costs) in addition to regulatory minima. ### Principle 5. Focus on tail risks - Use a variety of approaches to determine “extreme but plausible” shocks to be used in stress tests, including calibration on historical data, cross-country experience, “worst-ever” events, etc. - Whenever possible, adopt methodologies that capture tail events and risk dependencies. Ideally, use both balance sheet-based models and market price-based models, as the latter can incorporate more easily measures of risk dependencies using market prices. If using balance sheet-based models alone, supplement stress tests of individual financial institutions with tools that measure spillovers and joint default probabilities, such as network models. ### Principle 6. When communicating stress test results, speak smarter, not just louder - Tailor the communication of stress test assumptions, methodologies, and results to circumstances and the goals of the tests. Disclosure can yield substantial benefits for macroprudential stress tests, especially in crisis situations, when market confidence is at a premium. - Before disclosing stress test results, ensure that the tests cover relevant risks and transmission channels, assume serious shocks, set appropriate hurdle rates, produce a candid assessment, and are accompanied by a credible framework of follow-up action, including government support, if needed. - To the extent possible, set the communication of stress test results in the context of a broader communication strategy for financial stability objectives and policies, and complement them by other financial soundness indicators. ### Principle 7. Beware of the “black swan” - Supplement statistical approaches to the identification of potential tail events with expert judgment and new information. Consider analyzing the sensitivity of results to variations in the assumed shock, instead of choosing just one adverse scenario. - Supplement stress tests with other assessment techniques, such as qualitative and quantitative bank risk analysis, early warning indicators, models of debt sustainability, and informed dialogue with supervisors and market participants, among others. Final conclusions on the resilience of an institution or a system should be based on all these sources, not just the results of stress tests. - Evaluate and regularly re-assess the institutional coverage, risks, and channels of risk transmission included in the tests; as well as the models and other approaches. Constantly assess the suitability of new tools. Remain vigilant for the emergence of new risks. --- ### APPENDIX I. Existing supervisory guidelines for stress testing by banks - In the Basel II framework, banks are asked to apply rigorous stress testing programs. Pillar 1 focuses on minimum capital requirements for market risk, credit risk, and operational risk. - The Capital Requirements Directive (CRD) that introduced Basel II in the EU requires banks that apply the Internal Ratings-based approach to conduct stress tests under Pillar 1 in order to determine the regulatory capital for market risk. Internal ratings-based (IRB) banks are further asked to run credit risk stress tests in order to examine the robustness of their internal approach. - The Supervisory Review Process, Pillar 2 of the framework, requires banks to take a forward-looking and more comprehensive view on risk, which in turn should determine both capital and strategic planning. - Supervisors should make regular and comprehensive assessments of the processes, strategies, mechanisms, and systems that banks integrate in their Internal Capital Adequacy Assessment Process (ICAAP). The Supervisory Review and Evaluation Process (SREP) by supervisors is supposed to review and evaluate the banks’ internal stress testing framework according to the principle of proportionality. - Supervisors should consider banks’ results to assess capital adequacy and, if necessary, require additional capital and liquidity buffers. - The Basel Committee and other supervisors have issued guidance: stress testing should be an integral part of corporate governance and the risk management culture of a bank (CEBS, 2010). The board is ultimately responsible; senior management is accountable for implementation. - Senior management must identify and communicate risk appetite, identify relevant and plausible stress scenarios tested firm-wide, and ensure results feed into decision-making, including strategic business planning and capital and liquidity planning. - Scenarios should reflect bank-specific risks, take into account system-wide interactions, and be flexible to adapt to portfolio changes and new risks (BCBS, 2009). Scenarios should range from optimistic forecasts (baseline) to tail events that challenge the bank’s business model (reverse stress tests). - Historical shortcomings: BCBS (2009) found stress testing often technical, isolated, and insufficiently integrated with business units, leading to underestimation of risks. BCBS (2012) argued pre-crisis scenarios were based on pre-crisis data that did not cover heavy downturns or simultaneous realization of several risks, causing systematic underestimation. Banks failed to integrate guidelines on reputational risk, off-balance sheet vehicle risks, risks from highly leveraged counterparties, and deficiencies in risk mitigating techniques. - Supervisors share part of the blame: based on comprehensive SREP assessments, banks should have been required to take corrective action, and constructive open dialogue with other national and international public institutions would have helped identify systemic vulnerabilities (see BCBS, 2009 and 2012). ### APPENDIX II. Stress testing of nonbank financial institutions: insurance companies and financial market infrastructures A. Insurance Companies - Stress testing of the insurance sector is gaining growing acceptance as a risk management tool. Insurers increasingly use it to assess and manage their risks. Insurance supervisors use it to assess the risks facing specific insurers and to identify possible vulnerabilities of the sector as a whole. - From a financial stability perspective, stress testing of insurance companies is also important. If an insurer is significantly involved in activities such as providing protection against credit exposures—which are closely linked to the broader financial sector—then its failure could have systemic implications similar to those of the failure of a large, highly interconnected bank. Similarly, the failure of a large insurance company for which there is no quick substitution in the market, would also make a systemic case. - Insurance stress testing has only been performed in only nine FSAPs between 2003 and 2009, as compared with the banking sector, where stress testing is an integral part of every FSAP. - In FSAP insurance stress tests, efforts have been made to align the economic shocks to those used for the banking stress tests. However, insurers can also be vulnerable to other types of events: natural events (floods, earthquakes, windstorms) for nonlife insurers; a pandemic or long-term improvements in mortality for life insurers. - Insurance stress testing also incorporates insurance-specific shocks relevant to the jurisdiction or insurance sector. - Insurance stress testing tends to rely heavily on bottom-up tests. Reasons: contract-level linkage between assets and liabilities for many life products; effects of insurer-specific reinsurance programs on nonlife insurers; many supervisors lack detailed data and models for top-down tests. - Top-down insurance stress testing at meaningful granularity may be impossible in most jurisdictions until supervisory data and modeling capabilities evolve; efforts should be made to improve bottom-up stress testing quality. - The need to analyze vulnerabilities of global insurers has become clear as a result of the recent crisis. Their interconnectedness and cross-border activities are still not well understood and could pose a financial stability risk. The IMF is in the process of developing appropriate stress test methodologies and tools to gain a deeper understanding of the key risks affecting global insurers. B. Stress Testing of Financial Market Infrastructures (FMIs) - Stress testing measures FMIs’ resilience to extreme but plausible shocks. Unlike banks, FMIs are tested for proper functioning (operational, credit, or liquidity risk), focusing on immediate reaction and ability to finish the day and operate in subsequent days. - The Fund helps specify the stress testing requirements embedded in the CPSS/IOSCO standards and checks whether those requirements are met by the FMIs it assesses in the context of FSAPs. - FMI stress testing by central banks: - Central banks play a crucial role ensuring payment systems are designed and operated safely and efficiently. Central banks often operate the main national large-value payment system and oversee core payment systems and other FMIs (central counterparties, securities settlement systems). - Some central banks use simulation tools to analyze payment flows and participant behavior in different scenarios, in addition to standard-based qualitative approaches. Central banks may propose operational, organizational, or financial changes, such as new risk mitigation facilities. - The Bank of Finland developed a payment system simulator (BoF-PSS2), available to other central banks for customization, and organizes seminars to share practical experience. Eurosystem overseers decided to develop a tailor-made TARGET2-specific simulator based on BoF-PSS2 to run quantitative simulation-based stress tests on TARGET2. - Operational disruptions of an FMI or a major participant are tested, as well as financial default of major participants. Simulations identify recovery times, critical participants, contingency measures, and stop-sending limitations under parameters like concentration level, availability of liquidity, back-up procedures, reactions of non-defaulted banks, and money market structure. Simulations can indicate if the system can complete settlement before end of day per CPSS standards and define contingency capacity needs. They can quantify liquidity impact by simulating suspension of outgoing payments and test handling of financial defaults of major participants. - Stress-testing by FMIs: - Many FMIs have developed financial stress tests to ensure they can manage liquidity and credit risks. CHIPS instituted a program to ensure participants understand consequences of failures to honor closing positions and to encourage liquidity contingency plans. - Continuous Link Settlements (CLS) Bank conducts a range of stress and back-testing scenarios to review adequacy of risk management procedures, including adequacy of haircuts and failure of the settlement member with the single largest funding obligation in a single currency. Since the recent crisis, CLS Bank has worked on more extreme scenarios, such as failure of all settlement members of a given currency to complete funding and simultaneous failure of all liquidity providers in the same currency. - Some securities settlement systems, in particular the International Central Securities Depositories (ICSD), whose activities go beyond mere settlement (they offer securities loans and credit lines), face large credit and liquidity risks that need close monitoring and mitigation and most conduct regular stress tests of their financial resources. *Source: _082212 - 69.* ### 84. Among FMIs, CCPs exhibit the highest concentration of liquidity and credit risks. ### _082212 - 84. Among FMIs, CCPs exhibit the highest concentration of liquidity and credit risks. ### CCP concentration of credit and liquidity risks - CCPs become principal to every transaction they clear, which implies that market participants no longer have credit exposures to their trading counterparties, but only to the CCP. - As a result, CCPs concentrate credit risk and would face large liquidity needs if a participant defaults because they need to: - fulfill the settlement obligations of the defaulting participant, - cover potential losses when the cleared position or related collateral are liquidated, - meet the cash flows relating to possible hedge transactions. ### Role and design of stress testing for CCPs - Stress testing is key in managing CCP credit and liquidity risks. - Stress tests: - take into account extreme but plausible market conditions, - are typically framed in terms of the number of participant defaults a CCP can withstand. - Current CPSS/IOSCO standards prescribe that CCPs should be able to withstand the default of the participant with the largest exposure. - Some CCPs choose more stringent tests, including the default of several major participants. - CCP models that calculate margin requirements, default fund contributions, collateral requirements, and other risk control mechanisms are expected to be subjected to rigorous and frequent stress tests reflecting: - product mix, - other risk management choices. - Key elements of CCP stress tests: - assumed market conditions and default scenarios, - test frequency, - evaluation of extreme market conditions (that is, price changes significantly larger than the prevailing levels of volatility), - assessment of potential losses in individual participants’ positions. - Additional considerations: - distribution of positions between defaulting participants and their customers, - resources of potential defaulters available to a CCP (margins, clearing fund contributions or other assets), - CCP’s own resources, - perspective on potential size of losses and liquidity gaps of the CCP. ### Central clearing of OTC derivatives and regulatory attention - Since the financial crisis, regulators, overseers, and supervisors’ attention has focused on CCP stress testing, particularly when clearing OTC derivatives. - Central clearing of OTC derivatives presents more challenges than clearing listed or cash-market products because of complex risk characteristics. - Following the G-20 commitment, improved rules are to enter into force by end-2012; in some jurisdictions this will include guidance on stress testing (specifying types of tests for different classes of financial instruments and portfolios, involvement of clearing members or other parties, test frequency, and time horizon). - CCP supervision and oversight have generally been strengthened since the crisis, for example by systematically analyzing CCP internal risk management models, including stress testing parameters. ### FMI stress testing and the IMF’s role - The Fund often examines FMI stress testing arrangements in the context of FSAP and ROSC missions. - The Fund has not conducted stress tests on FMIs itself. - The Fund helps specify the stress testing requirements embedded in the CPSS/IOSCO standards and checks whether those requirements are met by the FMIs it assesses in the context of FSAPs. ### Crisis management stress tests: role and features - Since the recent financial crisis, stress tests have increasingly been used as instruments for crisis management; results feed into political decision-making and often determine allocation of public support measures. - Well-known crisis management exercises include: - Federal Reserve’s Supervisory Capital Assessment Program (SCAP), completed in May 2009, - EU-wide stress tests by CEBS in 2010 and EBA in 2011, - EBA’s 2011–2012 Capital Exercise. - Characteristics of crisis management stress tests: - combine microprudential and macroprudential features, - assess institutions’ resilience bank-by-bank while incorporating systemic risk through macroeconomic and market-level shocks, - potential follow-up actions include requiring recapitalization from private market, acceptance of governmental support packages, restructuring, and changes in business models, - concern for system-wide stability; shocks are common macroeconomic scenarios of systemic nature, - aggregate indicators are considered to evaluate resilience and potential need for macroprudential measures, - results are usually published with varying degrees of granularity and transparency. - Crisis management tests are tailored to current risks and are not necessarily regular; they typically involve higher-probability shocks than extreme-but-plausible tail risks in surveillance/supervisory tests. - Methodology combines bottom-up and top-down approaches: - supervisors design detailed methodology for banks to examine impact of consistent, common macroeconomic scenarios, - results are checked for completeness, consistency, and plausibility by supervisors, - centralized components and cross-checking are used for quality control. - Crisis management stress tests primarily focus on solvency, use traditional balance-sheet-based tests, and apply static balance sheet assumptions; liquidity risk was often not tested or assessed separately in major exercises. ### Comparison of prominent crisis management exercises (key quantitative findings) - Purpose differences: - SCAP: designed to "estimate losses, revenues, and reserve needs" and evaluated size of governmental capital injections contingent on bank performance. - EU-wide tests (2010, 2011): aimed to examine resilience, improve transparency, identify vulnerabilities, and inform policy-makers about capacity to absorb shocks and dependence on public support measures. - EBA’s Capital Exercise: designed to "create an exceptional and temporary capital buffer to address current market concerns over sovereign risk and other residual credit risk related to the current difficult market environment". - Severity and design of macroeconomic stress scenarios: - Adverse scenario of the 2011 European-wide stress test translated into a -4.1 percentage points cumulative GDP shock. - Adverse scenario of the 2010 European-wide stress test translated into a -3.1 percentage points cumulative GDP shock. - The SCAP adverse scenario involved a -2.8 percentage points cumulative GDP shock. - SCAP assumed a cumulative two-year contraction in GDP of -2.7 percent. - 2011 EBA exercise assumed a cumulative contraction of -0.2 percent. - 2010 EBA exercise assumed a cumulative contraction of -0.4 percent. - Organizational complexity of EU-wide tests: - involved 21 countries, - 24 national supervisors and authorities, - around a dozen European institutions, - 91 participating banks, - had to deal with 19 different languages and seven currencies. - Outcomes and credibility: - SCAP widely seen as credible and effective: after publication of results, the correlation among major banks’ stock prices fell by more than 10 percent; capital injections for banks not passing tests were mandatory. - 2011 EU-wide stress test was praised for risk coverage and disclosure but criticized for: - how sovereign exposures were stressed and the severity of stress scenarios in general, - use of a common adverse scenario that did not adequately reflect some national banking sectors’ risk profiles, - ambiguous backstops and uncertainty over actions for banks just above the hurdle rate, which undermined investor confidence. - Specific cases (e.g., Dexia in 2011 and Irish banks in 2010) further undermined credibility of EU test results. ### Frontiers and challenges in stress testing: integrating sovereign risk and feedbacks - EBA 2011 stress tests highlighted conceptual and technical difficulties in accounting for sovereign risk exposures in bank stress testing. - Need for models that account for banks’ impact on sovereign risk and feedbacks from sovereigns to banks. - Channels of interaction between sovereigns and banks include: - mark-to-market fall in value of government bonds held by local banks reduces bank asset values, - distressed banks increase government contingent liabilities, - high sovereign spreads raise bank funding costs, - erosion of the value of official support (guarantees) if sovereign is distressed, - contagion effects to foreign banks and sovereigns. - The document presents a schematic of spillovers (domestic and foreign) and multiple channels (A through I) showing how sovereign and bank stresses interact and propagate. *Source: _082212 - 84. Among FMIs, CCPs exhibit the highest concentration of liquidity and credit risks.* ### 102. To include sovereign risk in stress testing, the key interlinked risk exposures between ### 102. To include sovereign risk in stress testing, the key interlinked risk exposures between ### Framework linking sovereign and financial sector balance sheets - A stylized framework starts with the economic, i.e., risk-adjusted, balance sheets of the financial sector (portfolio of financial institutions) and is then linked to, and interacts with, the government’s balance sheet. - Channels and feedbacks: - Distressed financial institutions → large government contingent liabilities → reduced government assets → higher risk of default on sovereign debt. - Dynamic macrofinancial linkage models used for bank stress tests can be linked to sovereign risk models, incorporating feedback of banking risk to sovereigns via contingent liabilities and sovereign spreads affecting bank funding costs. ### Enhancing analysis of macroeconomic–banking feedbacks and contagion - Typical stress testing limitation: - Uses macroeconomic scenarios to assess banking/financial sector risk and capital adequacy without feedbacks to the macroeconomy. - Lesson from recent crises: banking/financial sector often becomes distressed first, credit supply contracts, and lower GDP growth follows—these feedbacks are usually excluded from stress tests. - Recommended modeling enhancements: - Build stress test models with explicit feedback channels: shock to financial sector → estimate reduction in GDP growth and other factors → negative impacts on corporate and household borrowers → increased credit risk on banks’ balance sheets. - Use dynamic factor models to enhance modeling of feedback between banks and the macroeconomy. - Contagion modeling: - Traditional macro factor models embed some bank correlation via common macro factors but do not capture inter-institution correlations, dependencies, and feedbacks. - Contagion effects can be modeled with networks, joint probabilities of default, systemic CCA, co-risk indicators, and other recently developed models. - Enhanced stress testing can incorporate features of these systemic risk models to analyze interdependence and joint risk. ### Techniques to improve modeling of non-linearities in bank assets, capital, credit risk, and funding cost - Risk-adjusted balance sheet and contingent claims approach: - Conceptual foundations: contingent claims approach, CCA (Merton, 1973) and risk-neutral valuation (Cox and Ross, 1976). - Liabilities (equity and (risky) debt) are claims on uncertain underlying bank assets; volatility affects risk premiums and values of equity and debt liabilities. - Two methods to construct a risk-adjusted balance sheet: - Use estimated loan portfolio loss distribution and other bank balance sheet components; using risk-neutral valuation, estimate probability distribution of the bank’s risky loans and distribution of assets over a specific time horizon; combine asset distribution with promised payments on debt and deposits to construct a risk-adjusted (CCA-type) balance sheet. This technique does not rely on market prices. - Estimate bank implied asset level and asset distribution from observed market value and volatility of the bank’s equity and its book value of debt and deposits (market-implied method). - Comparing risk indicators and tail risk shapes from the two methods provides insight into dynamics and differences between the “fundamental” loan portfolio loss approach and the market-implied view (which varies between calm and stress periods). - Funding cost channel in stress tests: - Include impact on banks’ funding costs: higher bank funding costs → higher lending rates for corporates and households, credit rationing, and lower credit growth → negative impact on economic output → feedback causing further banking system distress. - Funding costs depend on the risk-free interest rate, the bank’s credit spread, market risk appetite, and the impact of the government’s (implicit and explicit) guarantees. - Risk-adjusted balance sheet models are useful tools for calculating bank funding costs. - Capturing non-linearities and dynamic correlations: - More work needed to capture non-linearities between changes in bank assets and capital. - Balance sheet-based stress testing with fixed correlations and static risk weights can underestimate capital shortfalls under stress. - Improvements: - Use time-varying correlation between exposures when estimating portfolio loss distributions. - Market price–based contingent claims models already build in dynamic changes between assets and market capital: changes in assets → changes in market equity and expected losses to creditors/guarantors. - The change in bank market capitalization given a change in asset values is analogous to an aggregate RWA adjustment factor in the Basel rules; it can be seen as a “quasi-RWA” dynamic adjustment factor. ### Key references and methodological anchors cited in text - Foundational models and methodologies: - Merton, 1973; Cox and Ross, 1976. - Empirical and methodological literature referenced spans work on systemic risk measurement, funding liquidity modeling, contagion, network perspectives, stress testing principles, and next-generation balance sheet stress testing (multiple authors and institutional reports cited). *Source: Excerpt paragraphs 102–107 from provided IMF chapter/section content.* --- _Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/np/pp/eng/2012/_082212.pdf_