## _cr16162 — EXECUTIVE SUMMARY ## Source details **Canonical URL:** [_cr16162 — EXECUTIVE SUMMARY](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16162.pdf) ## Other formats - [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16162.pdf.md) - [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16162.pdf.json) --- ### EXECUTIVE SUMMARY — Purpose and Institutional Mandates - The Bank of England (BoE) has two mandates: a macroprudential mandate through the Financial Policy Committee (FPC) and a microprudential mandate through the Prudential Regulation Authority (PRA). - The stress testing program provides insights into both mandates and is guided by BoE (2015a) and PRA expectations in SS31/15. - Capital buffers are the principal tools; the stress testing program informs their setting: - The U.K. countercyclical capital buffer is informed by results from the concurrent stress test. - The FPC can use results from banks’ annual cyclical scenario in ICAAPs to inform the U.K. countercyclical capital buffer. - PRA prudential buffers capture material bank-specific risks and are sized at authorities’ discretion, informed by concurrent stress tests. - The note considers both the concurrent stress test and ICAAP stress testing, with focus on the concurrent exercise: - ICAAP is core to the supervisory review and evaluation process in the U.K. - ICAAP covers a Pillar 2A assessment and a Pillar 2B assessment (forward-looking severe-but-plausible stress). - The PRA’s PRA buffer assessment considers the maximum change in capital resources and requirements projected in concurrent stress testing and other relevant stress tests, and other factors influencing vulnerability (e.g., leverage ratio, projected use of capital buffers under stress). - BoE (2015b, p. 11) statement preserved: “[t]he main purpose of the [concurrent] stress-testing framework is to provide a forward-looking, quantitative assessment of the capital adequacy of the U.K. banking system as a whole, and individual institutions within it. In doing so, it aims to support both the FPC and the PRA in meeting their statutory objectives.” - Because capital buffers are central to implementing stress test conclusions, the concurrent stress test cannot be evaluated in isolation from ICAAPs. ### Evaluation Dimensions - The BoE’s stress testing program is evaluated along five dimensions: - Scope of coverage - Scenario design - Analytical infrastructure, models, and processes - Disclosure - Governance ### Key Findings and Prioritization of Improvements - Overall: The design of the concurrent stress test largely meets the objectives of the FPC and PRA, but the program is young and not all implementation dimensions are equally well-developed. - Priority improvements (in order of priority): 1) Analytical infrastructure—data, models, processes—requires substantial development. Medium-term recommendation (implement no later than 2018, following next review): - Settle the core data model. The data submission format (e.g., firm data submission framework (FDSF)) and organization of other stress test relevant data (e.g., ICAAP submissions) should be settled promptly. - Build out the supervisory model and analysis infrastructure. BoE uses three model types: - (1) Granular models using firm-submitted FDSF data; - (2) Product- or sector-specific models using industry or other data sources; - (3) “System” models examining interconnectedness and spillovers. - Investment should be accelerated especially in Type 1 and Type 2 models. 2) Coverage of the concurrent stress test may not be wide enough for large banking organizations active in the U.K. - Coverage extends to banks and building societies representing over three-quarters of banking system assets. - Eight of the 15 biggest U.K. banks (as measured by 2014 total assets) are foreign, mainly investment banks; omitting large foreign investment banks may hamper the FPC’s assessment. - Medium-term recommendation: include the largest subsidiaries in the concurrent stress test. --- ### Selected Quantitative and Coverage Facts (preserved exactly) - Banks/building societies included in 2015 concurrent stress test: Barclays plc, HSBC Holdings plc, the Lloyds Banking Group plc, Nationwide Building Society, the Royal Bank of Scotland Group plc, Santander U.K. plc, and Standard Chartered plc. - Coverage rule: total retail deposits greater than GBP 50 billion (in 2015, seven institutions covered). - Banks covered in the 2015 stress test accounted for about 80 percent of bank lending to the U.K. real economy (BoE, 2015b, p. 5). - Scenario variables in 2015: 57 macro variable projections across all geographies plus 11 yield curve paths. - Scenario publication and results timing: publication of scenario March 30, 2015; publication of bank results December 1, 2015; turnaround time eight months; time to publication from bank data reference date (year-end 2014) was 11 months. - Lending growth assumption for concurrent stress testing in 2015: 9 percent over the five year horizon. - Quote preserved: “[t]he main purpose of the [concurrent] stress-testing framework is to provide a forward-looking, quantitative assessment of the capital adequacy of the U.K. banking system as a whole, and individual institutions within it. In doing so, it aims to support both the FPC and the PRA in meeting their statutory objectives.” (BoE, 2015b, p. 11) --- ### OVERVIEW OF THE BOE'S STRESS TESTING PROGRAM — Scope, Scenarios, Methods - Objective: assess capital adequacy of the U.K. banking system (macroprudential) and constituent institutions (microprudential); generate information on potential system vulnerabilities; provide microprudential insight into banks’ stress-testing practices; inform PRA buffer decisions; disclosure intended to enable market discipline and enhance public confidence. - 2015 coverage: - The concurrent stress test covers all banks and building societies with total retail deposits greater than GBP 50 billion, which in 2015 covered seven institutions. - Insurers, asset managers and funds—as well as nondomestic banking entities—are not covered. - No explicit target level of system-wide asset or RWA coverage. - Consolidation perimeter: the banking group as defined by CRD IV, which includes investment banks but excludes insurance activities inside covered banking organizations. - The stress test methodology does not explicitly exclude any balance-sheet or profit-and-loss (P&L) items or risk types. - Scenario types: - Three scenario types: baseline; two adverse scenarios—annual cyclical scenario (ACS) and exploratory scenario. - ACS is performed annually; exploratory scenario added every other year (first time in 2017). - The annual scenario is explicitly countercyclical: severity and associated regulatory capital buffers vary with the financial cycle. - The exploratory scenario addresses a wide range of salient risks (not just cyclical); may not be required of all banks covered under the annual scenario—excluding banks reduces informational value. - 2015 scenario specifics: - Covered 57 macro variable projections across all geographies plus 11 yield curve paths. - A financial market shock was included. - Variable projections extend to a five-year horizon in quarterly time steps. - Turnaround time from publication of the scenario (March 30, 2015) to publication of bank results (December 1, 2015) was eight months; considering the bank data reference date (year-end 2014), the time to publication was 11 months. - Bank ICAAPs: - Banks required to conduct annual stress testing through ICAAP in addition to concurrent stress testing. - ICAAP requires banks to design bank-specific stress scenarios probing unique vulnerabilities. - Balance-sheet assumptions and management actions: - BoE generally allows banks to develop and use their own internal approaches to balance sheet and P&L projections; banks are not restricted to static balance sheets. - For concurrent stress testing the BoE imposed an overall lending growth assumption (in 2015, it was 9 percent over the five year horizon). - Banks may assume mitigating actions in three categories: - (a) Expected actions triggered by falls in capital ratios (e.g., dividend restrictions); - (b) Business-as-usual actions (e.g., reducing staff bonuses); - (c) “Strategic” management actions likely to involve significant board decision-making (e.g., asset disposals). - The BoE assesses realism of proposed management actions and requires banks to estimate results both with and without accepted “strategic” management actions. - BoE modelling and qualitative review: - BoE employs a suite of models to challenge banks’ results and to assess uncertainties/sensitivities; use of BoE models for quality assurance has been limited to informing judgment on individual bank projections; disclosure is based on banks’ results adjusted for BoE judgments. - BoE intends to build modelling capabilities to generate supervisory estimates more comprehensively, increasing coverage and incorporating system-wide dynamics; strict quality standards and model risk-management practices are a priority. - BoE undertakes qualitative review of banks’ stress testing frameworks, including models, data quality, governance and controls; banks must submit documentation to support this assessment. - Policy and disclosure links: - Stress tests support FPC’s statutory responsibility and the PRA’s responsibility to promote safety and soundness. - Concurrent stress test results and qualitative analysis, together with ICAAP information, support policy responses: adjusting regulatory capital buffers (including countercyclical buffer), sectoral capital requirements, and the PRA buffer. - Results are not “mechanically linked” to policy response. - The BoE’s concurrent stress testing framework is meant to be transparent; in 2014 and 2015 BoE disclosed scenario information, aggregate impacts, and post-stress minimum capital ratios (with and without “strategic” management actions) at individual bank level. --- ### A FRAMEWORK FOR EVALUATING CONCURRENT STRESS TESTING PROGRAMS — Dimensions - Five evaluation dimensions: - A. Scope of Coverage - B. Scenario Design - C. Analytical Infrastructure, Models, and Processes - D. Disclosure - E. Governance - Components of a well-designed crisis (wartime) exercise: - Scope of coverage - Severity and post-stress hurdles - Translation conservatism - Disclosure - Government capital backstop - Mapping to peacetime stress testing: all except government capital backstop map readily; BoE’s framework supports risk identification and monitoring and mitigates risks via capital buffers. --- ### Macroprudential and Microprudential Objectives - Microprudential: - Quantitative assessment of capital adequacy for individual banks under a pre-specified stress. - Qualitative insights into risk and capital management practices; relative comparisons across banks and judgments against supervisory expectations. - Helps supervisors assess whether banks have enough capital and adequate risk- and capital-management practices. - Macroprudential: - Insights into resilience of the financial system as a whole. - “Stress tests therefore contribute to the FPC’s statutory objective to protect and enhance the stability of the U.K. financial system.” (BoE, 2015a, p. 5). - Concurrent supervisory stress tests are complemented with banks’ ICAAPs and other information to inform sizing of capital buffers. --- ### Scope of Coverage — Current Design and Possible Expansion - Current scope: - Limited to PRA-regulated banks and building societies with total retail deposits greater than GBP 50 billion (captures seven banking organizations at time of writing). - Possible expansions (three dimensions): - More banks and building societies by reducing the size cut-off. - Other banking organizations, such as U.K. subsidiaries of foreign owned investment banks. - Other financial institutions such as insurers, asset managers, and pension funds. - Trade-off: balance comprehensive coverage against execution effort required by regulators and participating institutions. - Considerations on size threshold and omitted entities: - The smallest bank included in the 2015 stress test, Nationwide Building Society, had a balance sheet of GBP 189 billion at year-end 2014. - Jurisdictional differences: U.K. uses retail deposits; U.S. Dodd-Frank Act thresholds equivalent to about GBP 35 billion and GBP 7 billion; EBA 2016 sample inclusion: minimum of EUR 30 billion of total assets. - Exclusions risk: omitting banks with non-retail-deposit business models (e.g., investment banks) that can be systemically significant; eight of the 15 biggest U.K. bank legal entities by 2014 total assets are foreign. - Omitted entities are captured in home-country stress testing programs; PRA coordinates with home supervisors and uses domestically focused resolution regimes and solvent wind down plans as partial substitutes. - Risk remains that U.K.-only stress tested entities could be brought down by parent failures given global interconnectedness. --- ### Cross-border and Group-wide Considerations - Limitations of U.K.-only stress testing for foreign-owned groups: - Global firms have complex booking models and are highly interconnected; U.K.-only stress test may not include activity undertaken in branches or home-country entities. - Complete perspective on resilience requires group-wide stress tests by or for the home supervisor. - Importance of information exchange between home and host supervisors. - Policy trade-off: - Ring-fencing U.K. investment banking entities may conflict with BoE’s aim to strengthen international cooperation. - Omitting large foreign investment banks could hamper inference on financial stability where home-state information is not shared or global operations fragment regionally. --- ### Scenario Design — Importance and Choices - Scenario choice is central because concurrent stress test results inform capital buffer settings. - From 2016, BoE will add an annual “Exploratory Scenario” to probe vulnerabilities not captured by the countercyclical “Annual Cyclical Scenario.” - Design choices: - Relevant risk factors and number required. - Severity and target of severity. - Shape of scenario (e.g., sharp decline then recovery vs. gradual decline). - Horizon length and time-step frequency. - Example complexity: - U.S. program grew from 3 domestic risk factors in 2009 to 16 domestic variables and 12 variables covering regions outside the United States. - The 2015 BoE scenario comprised 57 macro variable projections across multiple geographies plus 11 yield curve paths. - Trade-offs: - Precision vs. coverage: high-precision few scenarios (microprudential) vs. many heterogeneous scenarios (macroprudential). - BoE approach: one countercyclical scenario annually; one exploratory scenario every other year; ICAAP and reverse stress testing act as mitigants. - Horizon and frequency: - BoE uses a five-year horizon to allow for slow realization of losses. - Comparative horizons: EBA three years; Federal Reserve nine quarters; APRA five years. - BoE uses annual projections to balance capturing long-range losses and managing execution burden. - Proposed hybrid: annual high-precision scenario for microprudential objectives complemented by many alternative scenarios run privately by the supervisor at lower granularity. - Market shock treatment and parameterization: - Banks with significant capital markets activities require a separate financial market shock consistent with the macro scenario. - The BoE 2015 trading shock specified approximately 450 parameters. --- ### BoE 2015 Stress Scenario — Foreign Risk Focus and Key Shocks - The 2015 stress scenario focused on foreign risk factors, recognizing emerging risks in East Asia. - House prices projected paths in the BoE 2015 stress scenario: - Mainland China: drop by 35 percent. - Hong Kong SAR: drop by 40 percent. - Comparative parameter counts: - 2014 EU exercise: approximately 950 parameters plus about 580 sovereign haircuts (by country and maturity). - U.S. Federal Reserve CCAR 2015: about 24,000 parameters across about 20 categories. - U.K. domestic component and bank-level impacts in 2015 scenario: - Unemployment rise in 2015 scenario: rose 3.5 percentage points (from 5.7 percent to 9.2 percent). - Unemployment rise in 2014 scenario: rose 4.6 percentage points (from 7.2 percent to 11.8 percent). - Home price drop in 2015 scenario: dropped 20 percent. - Home price drop in 2014 scenario: dropped 35 percent. - Nationwide’s CET1 capital ratio declined by less than 1 percentage point (from 19.8 percent to 19.1 percent). - Average CET1 decline (ex-Nationwide): 3.9 percent before ‘strategic’ management actions and 3.6 percent after. --- ### Analytical Infrastructure, Models, and Supervisory Processes - BoE strategy over next three years: substantial development of internal analytical infrastructure to: - Ensure consistency in projections across banks. - Produce BoE’s own estimates of institution level capital impact. - Produce BoE’s own estimates of system level capital impact including feedbacks and amplification mechanisms. - Three broad model types: 1. Granular models using FDSF data (primary quality-assurance tools for bank submissions). 2. Product or sector-specific models using industry or other data sources (useful for sector analysis beyond banks). 3. “System” models examining interconnectedness and spillover effects (estimate amplifier effects; example: RAMSI). - Roles across prudential dimensions: - Microprudential: granular models quality-assure bank submissions and uncover qualitative shortcomings. - Macroprudential: models used in scenario design to sift candidate scenarios and for off-line stress tests without burdening banks. - Data, submissions, and investments: - Data sources: FDSF and ICAAP submissions. - Ongoing work: BoE is setting up a more consistent database for flexible analysis. - Medium-term investment recommendations: - Settle the core data model (FDSF and organization of other stress test relevant data elements). - Build out supervisory model and analysis infrastructure; accelerate investment especially in type 1 and type 2 models. --- ### Disclosure Regime and Timing - Disclosure balance: sufficient detail to allow verification vs. protecting proprietary or market-sensitive information. - Observations: - Generous disclosure on process and results is important in a crisis to restore confidence. - Risks of granular disclosure: may drown out bank signals, hamper market discipline, induce strategic bank disclosure behavior, and reveal proprietary information. - Judgment on severity can be assessed at scenario (front end) and financial impact (back end); bank-level and aggregate results disclosure support assessment of translation from scenario to outcomes. - Timeline for BoE concurrent stress test process: - From release of the scenario (late March) to publication of results (beginning of December) takes eight months. - Given balance sheet cut-off date is prior year-end, disclosed results represent bank resilience from 11 months ago. - Potential timing efficiencies: if scenario released by end of January, submission by end-April may cut two months off timeline; as data model stabilizes, further efficiencies may be achievable. --- ### Bank-specific Results, Aggregated Disclosures, and Misconduct Costs - Bank-specific results: - Largely limited to summary measures like post-stress test ratios and capital depletion. - Banks allowed mitigating actions; disclosures show impacts with and without “strategic” management actions. - Aggregated disclosures (asset quality, profitability, trading): - Asset-class and geographic impairment disclosures enable inference of projected loss profiles. - Profitability projections under baseline and stress provide insight into resilience; profitability modeling under stress is less developed but materially affects final results. - Trading shock disclosures accommodated trading business models; six of the seven firms were required to execute a trading shock. - Four components to traded losses: direct marked to market impact (“market risk”); counterparty credit risk and CVA; prudential value adjustment (PVA) for valuation uncertainty. - Box highlights (preserved observations): - Absolute and relative impairments were higher from overseas exposures than from the U.K. - For U.K. retail exposures, nonmortgage impairments swamped mortgage losses in absolute terms. - Geographic impairment rates: baseline to stress was just under two times for the U.K., just over three times for China and Hong Kong SAR. - BoE published an aggregate stressed misconduct cost projection of GBP 40 billion. - At year end-2014, banks had already paid GBP 30 billion in fines and related costs since 2009, and had provided for an additional GBP 13 billion. - The stressed misconduct cost projection was not presented as a central forecast. --- ### Governance of the BoE Concurrent Stress Testing Program - Governance process: - Thorough but time-intensive; bank-specific results are approved by the PRA Board. - Stress test results inform buffer sizes, particularly UK countercyclical buffer and PRA buffers. - Sequencing: FPC considers system-wide buffer adjustments first; PRA then considers bank-by-bank PRA buffer sizing. - Governance challenges: - (a) balancing micro and macroprudential mandates; - (b) avoiding group think; - (c) arriving at objective and independent assessment of bank submissions and overall results. - Mitigations: - Major decisions taken after debate across multiple parts of the BoE. - Broad participation in scenario development and cross-bank challenge sessions; inclusion of external members on FPC and PRA Board. - Use of BoE-built models and analytical tools to provide independent assessments and reduce incentive for banks to understate stress impacts. - Timeline, resources, process logistics: - Expected timeline for 2016: data cut-off December 31, 2015; scenario and guidance release 29 March, 2016. - Banks have three months (April to June) to complete stress tests; submissions often include documentation exceeding 1,000 pages. - Quality assurance begins after submission; results phase (typically October/November) involves FPC and PRA Board discussion; final decisions made day before publication; publication typically around end-November/beginning of December. - Timing considerations: disclosed results reflect resilience from 11 months prior; potential to shorten timeline by earlier scenario release and stabilized data model. --- ### Conclusion and Linkages to Resolution - Stress testing is an important regulatory tool via the concurrent stress test and ICAAP; concurrent test adds simultaneous, published results; ICAAP is staggered and not public. - Room for improvement to make approach additive and proportional: - Improve value to market participants of disclosed information. - Increase speed/timeliness to make results more market-relevant. - Potentially increase extent of disclosure. - Resolution linkage: - If failure of a large institution can occur without major systemic disruption, need for “thick” capital buffers is reduced. - Linkage between stress testing and resolution could benefit from more clarity: plausibility of wind-downs in stressed environments; potential contagion effects of subsidiary or nonbank failures on the U.K. financial system. - BoE’s stress testing framework, supported by analytical infrastructure, could provide insights into these questions. *Prepared by Til Schuermann, external expert for the IMF; assessment is based on the BoE’s stress testing regime as of February 2016.* ### EXECUTIVE SUMMARY __________________________________________________________________________________ 4 ### _cr16162 - EXECUTIVE SUMMARY ### EXECUTIVE SUMMARY — Purpose and Institutional Mandates - The Bank of England (BoE) has two mandates: a macroprudential mandate through the Financial Policy Committee (FPC) and a microprudential mandate through the Prudential Regulation Authority (PRA). - The stress testing program provides insights into both mandates. In October 2015, the BoE released a policy paper describing their current approach (BoE, 2015a), which guides the discussion in this Technical Note. The PRA’s expectations on stress testing for capital planning are additionally set out in SS31/15. - Capital buffers are the principal tools at the BoE’s disposal to effectuate policy objectives; the stress testing program plays a significant role in informing their setting. - The U.K. countercyclical capital buffer, which affects all banks, is informed by results from the concurrent stress test. - The FPC can also use results from banks’ annual cyclical scenario in ICAAPs to inform the U.K. countercyclical capital buffer. - PRA prudential buffers capture material bank-specific risks (e.g., higher sensitivity to cyclical risks or material shortcomings in bank risk management). Both buffer types are sized at authorities’ discretion and are informed by concurrent stress tests. - The note considers both the concurrent stress test and ICAAP stress testing, with a focus on the concurrent exercise. The ICAAP: - Forms a core part of the supervisory review and evaluation process in the U.K. - Covers a Pillar 2A assessment (firms quantify risks not addressed or only partially addressed by Pillar 1) and a Pillar 2B assessment (firms assess forward-looking capital needs under severe but plausible stress). - The PRA carries out a PRA buffer assessment for all firms, considering: - The maximum change in capital resources and requirements projected in concurrent stress testing and other relevant stress tests (including those conducted by other supervisory authorities and the firm’s own stress scenarios); and - Other factors influencing a firm’s vulnerability to stress (e.g., the leverage ratio, projected use of capital buffers under stress). - BoE (2015b, p. 11) statement: “[t]he main purpose of the [concurrent] stress-testing framework is to provide a forward-looking, quantitative assessment of the capital adequacy of the U.K. banking system as a whole, and individual institutions within it. In doing so, it aims to support both the FPC and the PRA in meeting their statutory objectives.” - Because capital buffers are central to implementing stress test conclusions, the concurrent stress test cannot be evaluated in isolation from ICAAPs. ### Evaluation Dimensions The BoE’s stress testing program is evaluated along five dimensions: - Scope of coverage: are the covered institutions, portfolios, assets and geographies sufficient? - Scenario design: is the design of the scenarios appropriate? - Analytical infrastructure: are the processes, models, and other tools appropriate and sufficient? - Disclosure: does the disclosure regime help meet the stated objectives? - Governance: is the governance framework at the BoE appropriate and sufficient? ### Key Findings and Prioritization of Improvements - Overall: The design of the concurrent stress test largely meets the objectives of the FPC and PRA, but the program is young and not all implementation dimensions are equally well-developed. - Priority improvements (in order of priority): 1) Analytical infrastructure—data, models, processes—requires substantial development. Stress testing is analytically intensive across scenario design; evaluation, review, challenge and adjustment of bank submissions (microprudential); and aggregation/policymaking. - Medium-term recommendation (implement no later than 2018, following next review): a) Settle the core data model. The data submission format (e.g., firm data submission framework (FDSF)) and organization of other stress test relevant data (e.g., ICAAP submissions) should be settled promptly to allow infrastructure investment at the BoE and supervised banks. b) Build out the supervisory model and analysis infrastructure. The BoE uses three model types in stress testing quality assurance: - (1) Granular models using firm-submitted FDSF data; - (2) Product- or sector-specific models using industry or other data sources; - (3) “System” models examining interconnectedness and spillovers. - Investment, especially in Type 1 and 2 models, should be accelerated. 2) Coverage of the concurrent stress test may not be wide enough for large banking organizations active in the U.K. - Coverage extends to banks and building societies representing over three-quarters of banking system assets. - Eight of the 15 biggest U.K. banks (as measured by 2014 total assets) are foreign, mainly investment banks. Given London’s role as a global financial center, omitting large foreign investment banks may hamper the FPC’s assessment of financial stability. - Recommendation for the medium term: include the largest subsidiaries in the concurrent stress test. ### OVERVIEW OF THE BOE'S STRESS TESTING PROGRAM — Scope, Scenarios, Methods - Objective: assess capital adequacy of the U.K. banking system (macroprudential) and constituent institutions (microprudential); generate information on potential system vulnerabilities; provide microprudential insight into banks’ stress-testing practices; inform PRA buffer decisions; disclosure intended to enable market discipline and enhance public confidence. - 2015 coverage: - The concurrent stress test covers all banks and building societies with total retail deposits greater than GBP 50 billion, which in 2015 covered seven institutions. - Insurers, asset managers and funds—as well as nondomestic banking entities—are not covered. - No explicit target level of system-wide asset or RWA coverage. - Banks covered in the 2015 stress test accounted for about 80 percent of bank lending to the U.K. real economy (BoE, 2015b, p. 5). - Consolidation perimeter: the banking group as defined by CRD IV, which includes investment banks but excludes insurance activities inside covered banking organizations. - The stress test methodology does not explicitly exclude any balance-sheet or profit-and-loss (P&L) items or risk types. - Scenario types: - Three scenario types: baseline; two adverse scenarios—annual cyclical scenario (ACS) and exploratory scenario. - ACS is performed annually; exploratory scenario added every other year (first time in 2017). - The annual scenario is explicitly countercyclical: severity and associated regulatory capital buffers vary with the financial cycle; shock extent compared to current conditions is greater in good times. - The exploratory scenario addresses a wide range of salient risks (not just cyclical); may not be required of all banks covered under the annual scenario—excluding banks reduces informational value. - 2015 scenario specifics: - Covered 57 macro variable projections across all geographies plus 11 yield curve paths. - A financial market shock was included. - Variable projections extend to a five-year horizon in quarterly time steps. - Turnaround time from publication of the scenario (March 30, 2015) to publication of bank results (December 1, 2015) was eight months; considering the bank data reference date (year-end 2014), the time to publication was 11 months. - Bank ICAAPs: - Banks required to conduct annual stress testing through ICAAP in addition to concurrent stress testing. - ICAAP requires banks to design bank-specific stress scenarios probing unique vulnerabilities. - Balance-sheet assumptions and management actions: - BoE generally allows banks to develop and use their own internal approaches to balance sheet and P&L projections. Unlike the EBA, banks are not restricted to static balance sheets. - For concurrent stress testing the BoE imposed an overall lending growth assumption (in 2015, it was 9 percent over the five year horizon) to ensure adequacy of capitalization to support the real economy in stress. - Banks may assume mitigating actions in three categories: - (a) Expected actions triggered by falls in capital ratios (e.g., dividend restrictions); - (b) Business-as-usual actions (e.g., reducing staff bonuses); - (c) “Strategic” management actions likely to involve significant board decision-making (e.g., asset disposals). - The BoE assesses realism of proposed management actions and requires banks to estimate results both with and without accepted “strategic” management actions. - BoE modelling and qualitative review: - BoE employs a suite of models to challenge banks’ results and to assess uncertainties/sensitivities. Models use supervisory review data, portfolio-level models, sectoral asset class risk reviews, etc. - Use of BoE models for quality assurance has been limited to informing judgment on individual bank projections; stress test disclosure is based on banks’ results adjusted for BoE judgments. - The BoE intends to build modelling capabilities to generate supervisory estimates more comprehensively, increasing coverage and incorporating system-wide dynamics. - Strict quality standards and model risk-management practices are a priority. - BoE undertakes qualitative review of banks’ stress testing frameworks, including models, data quality, governance and controls; banks must submit documentation to support this assessment. - Policy and disclosure links: - Stress tests support FPC’s statutory responsibility to identify, monitor, and take action to remove or reduce systemic risks and the PRA’s responsibility to promote safety and soundness. - Concurrent stress test results and qualitative analysis, together with ICAAP information, support policy responses: adjusting regulatory capital buffers (including countercyclical buffer), sectoral capital requirements, and the PRA buffer. - Results are not “mechanically linked” to policy response. - The BoE’s concurrent stress testing framework is meant to be transparent. In 2014 and 2015, BoE disclosed scenario information, aggregate impacts, and post-stress minimum capital ratios (with and without “strategic” management actions) at individual bank level. - The BoE notes disclosed information and results may change over time as the framework evolves and to reflect that year’s risks. ### A FRAMEWORK FOR EVALUATING CONCURRENT STRESS TESTING PROGRAMS — Dimensions Introduced - The note presents five dimensions to evaluate concurrent stress testing programs: - A. Scope of Coverage - B. Scenario Design - C. Analytical Infrastructure, Models, and Processes - D. Disclosure - E. Governance ### Selected Quantitative and Coverage Facts (preserved exactly) - Banks/building societies included in 2015 concurrent stress test: Barclays plc, HSBC Holdings plc, the Lloyds Banking Group plc, Nationwide Building Society, the Royal Bank of Scotland Group plc, Santander U.K. plc, and Standard Chartered plc. - Coverage rule: total retail deposits greater than GBP 50 billion (in 2015, seven institutions covered). - Banks covered in the 2015 stress test accounted for about 80 percent of bank lending to the U.K. real economy (BoE, 2015b, p. 5). - Scenario variables in 2015: 57 macro variable projections across all geographies plus 11 yield curve paths. - Scenario publication and results timing: publication of scenario March 30, 2015; publication of bank results December 1, 2015; turnaround time eight months; time to publication from bank data reference date (year-end 2014) was 11 months. - Lending growth assumption for concurrent stress testing in 2015: 9 percent over the five year horizon. - Quote preserved: “[t]he main purpose of the [concurrent] stress-testing framework is to provide a forward-looking, quantitative assessment of the capital adequacy of the U.K. banking system as a whole, and individual institutions within it. In doing so, it aims to support both the FPC and the PRA in meeting their statutory objectives.” (BoE, 2015b, p. 11) *Prepared by Til Schuermann, external expert for the IMF; assessment is based on the BoE’s stress testing regime as of February 2016.* ### 11. Given the stated objectives to be achieved by the concurrent stress test, this section ### _cr16162 - 11. Given the stated objectives to be achieved by the concurrent stress test, this section ### Framework for assessing the concurrent stress test - The section outlines a framework to assess the BoE’s concurrent stress test along five dimensions: - Scope of coverage: Are the covered institutions, portfolios, assets, and geographies sufficient? - Scenario design: Since the actual scenario is at the core of a stress testing exercise, is the design of the scenarios appropriate? - Process and models: Are the processes, models, and other tools that exist or are being built appropriate and sufficient? - Disclosure: Does the disclosure regime help to meet the stated objectives? - Governance: Is the governance framework appropriate and sufficient? ### Components of a well-designed crisis (wartime) exercise - A well-designed crisis exercise has five components: - Scope of coverage: The included banks should cover enough of the banking system to ensure proper assessment of financial stability. - Severity and post-stress hurdles: The scenario needs to be severe enough, and the post-stress capital hurdles need to be sufficiently high, to be a credible worst case. - Translation conservatism: The translation of the scenario to the outcomes (e.g., losses, profitability, and capital impact) needs to be appropriately conservative. - Disclosure: The disclosure regime has to be sufficiently detailed to allow verification of process and results. - Government capital backstop: A credible capital backstop by the government in case banks, judged to be viable, who need to raise capital, cannot do so forth their own in a timely manner. - Mapping to peacetime stress testing: All except the last component map readily to “peacetime” stress testing. The BoE’s framework supports risk identification and monitoring through countercyclical and exploratory scenarios and mitigates risks via the capital buffer framework. ### Macroprudential and microprudential objectives - Stress testing serves both macro- and microprudential objectives: - Microprudential: - Provides quantitative assessment of capital adequacy under a pre-specified stress scenario for individual banks. - Yields qualitative insights into risk and capital management practices, enabling relative comparisons across banks and absolute judgments against supervisory expectations. - Helps supervisors answer: do banks have enough capital to support their risk taking activities, and are their risk- and capital-management practices “good enough”? - Macroprudential: - Provides insights into the resilience of the financial system as a whole to shocks. - “Stress tests therefore contribute to the FPC’s statutory objective to protect and enhance the stability of the U.K. financial system.” (BoE, 2015a, p. 5). - The macroprudential objective covers the broader financial system, not just banking. - Concurrent supervisory stress tests are complemented with banks’ ICAAPs and other information to inform the PRA and the FPC on sizing of different capital buffers—the principal mechanism for impacting banks. ### Scope of coverage (current design and possible expansion) - Current scope: - Limited to PRA-regulated banks and building societies with total retail deposits greater than GBP 50 billion. - As of this writing, this criterion captures seven banking organizations. - Possible expansions (three dimensions): - More banks and building societies by reducing the size cut-off. - Other banking organizations, such as U.K. subsidiaries of foreign owned investment banks. - Other financial institutions such as insurers, asset managers, and pension funds. - Trade-off: Regulators must balance comprehensive coverage of the financial system against the effort required to execute the stress test by both regulators and participating institutions. ### BoE capital framework and buffers (Box 1 summary) - The BoE’s capital framework operationalizes macro- and microprudential aspects of stress testing. Requirements and buffers fall into three categories: - The minimum requirements: - Combination of the Pillar 1 minimum (4.5 percent core equity tier 1 (CET1)) and Pillar 2A requirements addressing risks not otherwise captured in Pillar 1 (e.g., pension risk, concentration risk, and interest rate risk in the banking book). - The systemically important financial institutions (SIFI) buffer: - Extra capital buffer for banks deemed systemically important; bank-specific and not impacted by stress test results. - Buffers examined by the stress test (three sets, all available for loss absorption during the stress test): - Capital conservation buffer: 2.5 percent CET1, built up in good times for loss absorption. - Countercyclical capital buffer and sectoral capital requirements: can be used or released in a stress to offer additional protection against systemic risk build-up; U.K. CCyB rate is informed by stress test results. - PRA buffer: prudential, bank-specific buffer to capture material bank-specific risks (e.g., higher sensitivity to cyclical risks or shortcomings in risk management); size is at supervisory discretion by the PRA. - Use of stress test results: - Inform setting of regulatory buffers. - Can lead to bank-level supervisory actions (example: following the 2014 stress test, the PRA Board required The Co-Operative Bank to submit a revised capital plan). - Inform supervisory decisions on remuneration and dividends. ### Relationship between concurrent stress test and ICAAP (Box 2 summary) - Multiple stress testing regimes in the U.K.: - Concurrent stress test: simultaneous stress test with same scenario and approach across sample; results published together. - EBA stress test: European-level exercise for included U.K. banks. - ICAAP stress testing: firms conduct firm-specific stress testing annually as part of ICAAP. - ICAAP: - Core to supervisory review and evaluation process; covers (a) firm-specific Pillar 2A capital assessment and (b) Pillar 2B capital assessment (stress testing capital plans). - Firms develop firm-wide scenarios relevant to their business model; PRA publishes a macroeconomic scenario as guide and severity benchmark. - From 2016 onwards, the PRA macroeconomic scenario will be the annual cyclical scenario used in that year’s concurrent stress test. - Firms must also complete “reverse stress test” identifying scenarios that would test business plans to failure. - Value added by the concurrent stress test (threefold): - Run at the same time for all institutions in the sample, enabling like-for-like comparison. - Results are published, adding transparency for market participants. - Scenarios specified by the BoE may be more focused and topical than those in ICAAP. - Concurrence leverages supervisory advantage: ability to compare exposures, vulnerabilities, models, practices, and resilience across firms; aids quality assurance given the BoE’s evaluation begins with banks’ own calculations. ### Considerations on size threshold and omitted entities - Size threshold rationale: - FPC’s statutory objective: “to protect and enhance the stability of the U.K. financial system.” - If focus on banking system is sufficient, reducing size threshold below GBP 50 billion retail deposits may not yield significant gains. - The smallest bank included in the 2015 stress test, Nationwide Building Society, had a balance sheet of GBP 189 billion at year-end 2014. - Criterion differences across jurisdictions: - U.K. uses retail deposits as size criterion; other jurisdictions use assets. - U.S. Dodd-Frank Act: semi-annual (annual) stress tests for banks with assets of at least USD 50 billion (USD 10 billion) — equivalent to about GBP 35 billion and GBP 7 billion, respectively. - EBA 2016 sample inclusion: minimum of EUR 30 billion of total assets at highest level of consolidation. - Exclusions and risks: - Focusing on retail deposits excludes banks with business models not dependent on retail deposits (e.g., investment banks, trust/custody banks) that can become systemically significant. - Eight of the 15 biggest U.K. bank legal entities by 2014 total assets are foreign (mainly investment banks). - U.K. subsidiaries of foreign investment banks are a sizable omission; largest of these should be included in the medium-term. - Omitted entities are captured in home-country stress testing programs; PRA coordinates with home supervisors and uses domestically focused resolution regimes and solvent wind down plans as partial stress-testing substitutes. - Risk: U.K. entities passing a U.K.-only stress test could be brought down by failure of the parent, due to global interconnectedness, flexible booking, centralized risk management, and significant intra-group flows. ### Cross-border and group-wide considerations - Limitations of U.K.-only stress testing for foreign-owned groups: - Global firms have complex booking models and are highly interconnected; legal-entity level balance sheet coherence can be weak. - A U.K.-only stress test may not include activity undertaken in branches of global firms based in the home country or the EU, which can be larger than subsidiaries. - A complete perspective on resilience requires group-wide stress tests conducted by or for the home supervisor. - Importance of information exchange: - Rich information exchange between home and host supervisors is essential. - Close engagement and strong relationships with home-state regulators are necessary to implement agreed single point of entry (SPE) resolution approaches. - Policy trade-off: - Further ring-fencing (e.g., stand-alone stress testing of U.K. investment banking entities) may conflict with BoE’s aim to strengthen international cooperation. - Given London’s global role, omitting large foreign investment banks could hamper inference on current financial stability, especially if: - (1) The PRA is not provided with information on global group stress tests from home-state regulators; and - (2) Global investment banking fragments regionally, increasing chance that U.K. offices could operate as viable stand-alone entities. ### Scenario design (importance and choices) - The scenario choice is central, especially because concurrent stress test results inform the setting of capital buffers. - From 2016, the BoE will add an annual “Exploratory Scenario” to probe vulnerabilities not captured by the countercyclical “Annual Cyclical Scenario.” - Design choices supervisors face: - Relevant risk factors to focus on and number needed to cover exposures/vulnerabilities. - Severity of the scenario, and for whom it should be severe. - Shape of the scenario (e.g., sharp decline followed by gradual recovery, or gradual decline with no recovery). - Horizon length. - Time-step frequency (quarterly or annual). - Example of variable complexity: - U.S. program grew from 3 domestic risk factors in 2009 (GDP growth, unemployment rate, residential house price index) to 16 domestic variables and 12 variables covering regions outside the United States. - The 2015 BoE concurrent stress testing scenario comprised 57 macro variable projections across multiple geographies plus 11 yield curve paths. *Source: IMF staff compilation from the referenced chapter content.* ### 23. The 2015 stress scenario focussed on exploring foreign risk factors. Recognizing ### _cr16162 - 23. The 2015 stress scenario focussed on exploring foreign risk factors. Recognizing ### BoE 2015 stress scenario: foreign risk focus and key shocks - The 2015 stress scenario focussed on exploring foreign risk factors, recognizing emerging risks in East Asia in 2015, a region where several large U.K. banks have significant exposures. - House prices projected paths in the BoE 2015 stress scenario: - Mainland China: drop by 35 percent. - Hong Kong SAR: drop by 40 percent. - The BoE 2015 concurrent stress test included a trading shock designed to be consistent with the macro scenario, with approximately 450 parameters specified. - Comparative parameter counts in other exercises: - 2014 EU exercise: approximately 950 parameters plus about 580 sovereign haircuts (by country and maturity). - U.S. Federal Reserve CCAR 2015: about 24,000 parameters across about 20 categories. ### U.K. domestic component and bank-level impacts in 2015 scenario - The U.K. domestic component saw less severe stresses than in the 2014 concurrent stress test: - Unemployment rise in 2015 scenario: rose 3.5 percentage points (from 5.7 percent to 9.2 percent). - Unemployment rise in 2014 scenario: rose 4.6 percentage points (from 7.2 percent to 11.8 percent). - Home price drop in 2015 scenario: dropped 20 percent. - Home price drop in 2014 scenario: dropped 35 percent. - Capital impacts: - Nationwide’s CET1 capital ratio declined by less than 1 percentage point (from 19.8 percent to 19.1 percent). - Average CET1 decline (ex-Nationwide): 3.9 percent before ‘strategic’ management actions and 3.6 percent after. ### Scenario design trade-offs and frequency/horizon choices - Key design trade-offs: - Precision vs. coverage: higher precision running few scenarios vs. lower precision running many heterogeneous scenarios. - Supervisory objective alignment: microprudential benefits from high precision; macroprudential coverage benefits from many heterogeneous scenarios. - Heterogeneity dimensions recommended: choice of risk factors, severity, shape, length. - BoE concurrent stress testing regime contemplation: - One countercyclical scenario run annually. - One alternative, more exploratory scenario run every other year. - Mitigants: banks run multiple scenarios as part of their ICAAP and perform reverse stress testing. - Horizon and frequency choices: - BoE has chosen a five-year horizon to allow for possible slow realization of losses. - Comparative horizons: EBA is three years; Federal Reserve asks for a nine quarter projection; APRA also imposes a five-year horizon. - Frequency: BoE uses lower frequency (annual) projections to capture long-range or slow loss emergence profiles while managing execution burden. - Hybrid solution proposed: annual high-precision scenario for microprudential objectives complemented by many alternative scenarios run privately by the supervisor at lower granularity. ### Market shock treatment and parameterization - Banks with significant capital markets activities face market and counterparty credit risks at higher frequency (daily) and therefore require a separate financial market shock consistent with the macro scenario. - The BoE 2015 trading shock specified approximately 450 parameters to probe trading vulnerabilities. ### Analytical infrastructure, models, and supervisory processes - BoE strategy over next three years: more substantial development of internal analytical infrastructure. - Purposes for dedicated stress testing models: - Ensuring consistency in projections across banks. - Producing BoE’s own estimates of institution level capital impact of the stress. - Producing BoE’s own estimates of system level capital impact of the stress, including feedbacks and amplification mechanisms. - Three broad model types used by the BoE: 1. Granular models using data submitted by firms via the FDSF (firm data submission framework). 2. Product or sector-specific models using industry or other data sources. 3. “System” models that examine interconnectedness, spillover effects, etc. - Roles and interactions: - Type 1 (granular) models: project losses (e.g., residential mortgage impairment) and serve as primary quality-assurance tools for bank submissions. - Type 2 (product/sector) models: useful for sector analysis beyond banks and as alternative perspectives to granular-model results. - Type 3 (system) models: estimate amplifier effects on bank losses and provide challenger results at aggregate level (example: RAMSI—risk assessment model of systemic institutions). - Uses across prudential dimensions: - Microprudential: granular models quality-assure bank submissions and uncover qualitative shortcomings in risk- and capital-management practices. - Macroprudential: models can be used in scenario design to quickly sift candidate scenarios and for off-line stress tests without burdening banks. ### Data, submissions, and investments - Data sources: - FDSF: firm data submission framework used for granular model inputs. - ICAAP submissions: firms submit data with supporting documentation used, among other things, to inform the Pillar 2A capital charge. - Ongoing work: BoE is setting up a more consistent database to allow for more flexible analysis; these data are used in the concurrent stress test process to inform views on firm stress test results. - Medium-term investment recommendations (accelerate investments along two dimensions): - Settling the core data model: settle data submission format (e.g., FDSF) and organization of other stress test relevant data elements from firm ICAAP submissions as soon as practicable to allow infrastructure investment at the BoE and supervised banks. - Building out the supervisory model and analysis infrastructure: accelerate investment especially in type 1 and type 2 models to effectuate robust quality assurance of bank submissions. ### Disclosure regime and timing - Disclosure regime balance: information dissemination (sufficient detail to allow verification of process and results) vs. information protection (avoid granular disclosures that invite gaming or reveal proprietary, market-sensitive information). - Observations: - Generous disclosure on process and results is important in a crisis to restore confidence in the banking system. - Risks of granular disclosure: may drown out bank signals, hamper market discipline, induce strategic bank disclosure behavior, and reveal proprietary information. - Judgment on severity: can be assessed at the scenario (front end) and at the financial impact on banks (back end); bank-level and aggregate results disclosure support rigorous assessment of translation from scenario to outcomes. - Timeline for BoE concurrent stress test process: from release of the scenario (late March) to publication of results (beginning of December) takes eight months. Given the balance sheet cut-off date is the prior year-end, disclosed results represent bank resilience from 11 months ago. *Source: UNITED KINGDOM — INTERNATIONAL MONETARY FUND, _cr16162 - 23. The 2015 stress scenario focussed on exploring foreign risk factors. Recognizing* ### 50. Bank-specific results are largely limited to summary measures like post-stress test ### _cr16162 - 50. Bank-specific results are largely limited to summary measures like post-stress test ### Bank-specific results and disclosures - Bank-specific results are largely limited to summary measures like post-stress test ratios and thereby capital depletion due to the stress scenario. - Banks are allowed to take some mitigating actions, and their impact is reflected in the disclosures. - The capital impact is shown with and without “strategic” management actions such as expense reductions attributable to cuts in staff costs, to allow the public to judge flexibility under stress. - This disclosure allows for evaluation of the degree of conservatism or discipline the supervisor has exercised over the banks on whose calculations these results are based. - Example illustration: Table 2 in BoE (2015b) reproduced as Figure 3 in the source. ### Aggregated disclosures: asset quality, profitability, and trading - Bank-level results are complemented by more aggregated disclosures that lend insight into asset quality and profit-generation capacity. - With information about portfolio composition for a given bank, market participants can infer a bank’s projected loss profile from aggregate results. - Examples from the 2015 exercise: - Disclosure on impact by asset class or geography: - Asset-class-level results can hint at scenario impact on nonparticipating banks since some portfolio composition information is usually available. - See Box 3 for illustrative charts and commentary. - Disclosure of profits: - Profit generation capacity in a stressful environment provides important insights into bank resilience. - See Box 4 for commentary: profitability projections under baseline and stress, and impairment projection by major geographic region. - Disclosure of impact on trading: - Some banks have significant capital markets activities; the stress test must accommodate those business models and probe vulnerabilities. - See Box 5 for points to consider. ### Box 3 — Disclosure on impact by asset class or geography (highlights) - Impairment presentation methods: - “Chart 12”: by borrower category (individuals, nonfinancial businesses) and U.K./non-U.K. across banks; showed both absolute (GBP) and relative (percent) impairments. - “Chart 13”: by bank for a given borrower category (lending to individuals) split between mortgage and nonmortgage. - “Chart A”: impairments by major region aggregated across individuals and businesses, showing bars for impairment in GBP and diamonds for impairment rate. - Key observations: - Absolute and relative impairments were higher from overseas exposures than from the U.K.; possible drivers include harsher scenario in non-U.K. geographies (especially East Asia) and portfolio risk differences. - For U.K. retail exposures, nonmortgage impairments swamped mortgage losses in absolute terms; drivers include product mix and relative size of shocks. - Geographic impairment rates averaged across retail and commercial portfolios: baseline to stress was just under two times for the U.K., just over three times for China and Hong Kong SAR, reflecting differing home price shocks (U.K. -20 percent; China -35 percent; Hong Kong -40 percent). ### Box 4 — Disclosure of profits (highlights) - Profitability modeling under stress is less developed than credit and market risk stress loss modeling but can substantially affect final results. - Disclosing baseline versus stress profitability projections helps demonstrate the severity of the stress scenario; the stress case was projected to be even worse than observed during the financial crisis. ### Box 5 — Disclosure of trading assets (highlights) - Six of the seven firms were required to execute a trading shock. - Four components to traded losses (BoE, 2015b, p. 28): - direct marked to market impact (“market risk”); - counterparty credit risk and credit value adjustment (CVA); - prudential value adjustment (PVA) for valuation uncertainty (complex and/or very illiquid positions). - Disclosures provided enough detail to give insights on trading shock sensitivities and category materiality without revealing market-sensitive firm proprietary information. - Disclosures by bank did not decompose loss components; components were shown in aggregate across banks. - Note: Nationwide is not permitted to undertake trading activity except for managing liquidity. ### Misconduct cost component - In 2015, a separate misconduct cost component was added to the scenario to capture uncertainty in fines and legal settlements related to crisis-era misconduct. - The BoE published only an aggregate stressed misconduct cost projection of GBP 40 billion. - The disclosure noted that at year end-2014, banks had already paid GBP 30 billion in fines and related costs since 2009, and had provided for an additional GBP 13 billion. - The stressed misconduct cost projection was not presented as a central forecast of misconduct costs and provisions in the period covered by the stress test. ### Qualitative disclosures and supervisory integration - The disclosure document included qualitative commentary on data quality, model risk-management capabilities, governance processes around interest income projections, and documentation standards. - As bank capitalization improves, qualitative, microprudential aspects of the stress testing program are expected to gain importance. - Supervisory conclusions are informed by the supervisory review and evaluation processes (SREP) and other relevant supervisory information to inform PRA Board decisions on capital adequacy. ### Governance of the BoE concurrent stress testing program - The BoE governance process is thorough but time-intensive, bringing together broad expertise; bank staff operate under guidance of the FPC and PRA Board. - Bank-specific results are approved by the PRA Board. - Stress test results inform sizes of buffers, particularly the UK countercyclical buffer rate and PRA buffers. - Sequencing in decision-making: - FPC considers system-wide buffer adjustments first. - PRA then considers bank-by-bank microprudential PRA buffer sizing, accounting for heterogeneity across firms. - Three governance challenges the BoE faces: - (a) balancing micro and macroprudential mandates; - (b) avoiding group think; - (c) arriving at an objective and independent assessment of bank submissions and overall stress test results. - Mechanisms to mitigate these challenges: - Major decisions (stress scenarios, final results, disclosures) taken together after debate across multiple parts of the bank. - Broad participation in scenario development (monetary analysis, international directorate, PRA supervisors and risk specialists). - Cross-bank challenge sessions by major area (retail, wholesale, net interest income, etc.) and inclusion of external members on FPC and PRA Board for additional challenge. - Use of BoE-built models and analytical tools to provide independent assessments and reduce incentives for banks to understate stress impacts. ### Timeline, resources, and process logistics - The governance process follows the BoE’s stress test timeline lasting eight months from scenario release to disclosure of results. - Expected timeline for 2016 (data cut-off and scenario dates): - data cut-off for bank financials: December 31, 2015; - scenario and guidance release: 29 March, 2016. - Banks have three months (April to June) to complete the stress tests. - Work during submission period includes expansion of the stress scenario and mapping scenario to stressed outputs: losses; profitability; balance sheet; capital ratios. - Complexity of scenario specification noted: 57 macro variable projections across several geographies plus 11 yield curve paths. - Submissions often include extensive supporting documentation, often exceeding 1,000 pages. - Quality assurance by BoE staff begins after submission with regular contact to minimize data errors and misinterpretations. - Results phase (typically October/November): FPC and PRA Board discuss staff analysis and policy responses; final decisions made the day before publication. - Publication is typically around the end of November/beginning of December. - Timing considerations and potential efficiencies: - By the time results are disclosed they represent bank resilience from 11 months prior (balance sheet cut-off is prior year-end). - Banks typically have internal financials within two to three weeks of a quarter-end; if they received the scenario by the end of January, they could submit by end-April, potentially cutting two months off the timeline. - As the process matures and the data model stabilizes, further production efficiencies may be achievable. - Footnote reference: matching Federal Reserve CCAR scenario release timing which occurred this year on January 28. ### Conclusion and linkages to resolution - Stress testing is an important regulatory tool at the BoE through the concurrent stress test and ICAAP; the concurrent stress test adds a simultaneous stress with published results, while ICAAP is staggered and not public. - Room for improvement exists to make the approach both additive to objectives and proportional. - The BoE suggests that banks outside the concurrent stress test should consider using the annual cyclical scenario as part of ICAAP, and could consider greater synergies between the two processes. - Focus areas for improvement: - improving value to market participants of disclosed information; - increasing speed/timeliness to make results more market-relevant; - potentially increasing extent of disclosure. - Resolution regime linkage: - If failure of a large institution can occur without major systemic disruption, need for “thick” capital buffers becomes less critical. - The linkage between stress testing and resolution could benefit from more clarity: plausibility of wind-downs in stressed environments; potential contagion or knock-on effects of subsidiary or nonbank failures on the U.K. financial system. - The BoE’s stress testing framework, supported by analytical infrastructure, could provide insights into these questions. *International Monetary Fund — UNITED KINGDOM section (excerpt)._cr16162 - 50.* --- _Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16162.pdf_