## EXECUTIVE SUMMARY ## Source details **Canonical URL:** [EXECUTIVE SUMMARY](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16164.pdf) ## Other formats - [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16164.pdf.md) - [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16164.pdf.json) --- ### Assessment scope and purpose - Summarizes the assessment of interconnectedness and systemic risk for the United Kingdom (U.K.) financial system conducted as part of the Financial Sector Assessment Program (FSAP). - Analysis consists of three parts: - Motivation for monitoring cross-sector interconnectedness as part of financial system resilience assessment. - Description of selected empirical methods to analyze interconnectedness. - An illustrative analysis based on a definition of the financial system that incorporates U.K. banking and life insurance sectors. - Time span for the illustrative analysis: August 2007–January 2016. - PoD inputs: PoD estimates backed out from five-year CDS spreads for each financial entity; these data are available on a daily frequency. ### Key findings on systemic risk and sector resilience - Overall systemic risk in the U.K. has declined to pre-crisis levels. This reflects, to a significant degree, the improved sector-specific resilience established under sector-by-sector approaches applied in this FSAP. - Improved banking sector resilience is verified using metrics spanning market-based as well as balance sheet information. - The likelihood of distress spilling over among individual banks is currently subdued, with little heterogeneity in this measure across banks. - Interconnectedness within the banking sector has been declining over recent years. - Banking–life insurance sector comparisons: - Using the likelihood of distress spillovers between sectors, the banking–life insurance interconnectedness shows variability. - Using alternative methodologies based on analyzing spillovers of asset price volatility, evidence of steadily increasing interconnectedness between these sectors is found. - The analysis is essentially market-data based and illustrative, covering banks and life insurers only. ### Data coverage—entities included - Banks included: Barclays, HSBC, RBS, Lloyd’s, and Standard Chartered. - Insurers (life only) included: Aviva, Legal & General, Prudential Plc, and Standard Life. ### Monitoring cross-sector interconnectedness: motivation and channels of spillovers - Rationale: - Firm-by-firm or sector-by-sector approaches analyze resilience at a point in time; structural market changes continually give rise to new channels of interconnectedness that can transmit or amplify shocks. - Interconnectedness channels: - Direct: counterparty transactions (e.g., interbank lending) creating counterparty credit risk where an insolvent firm imposes losses on counterparties. - Indirect: common exposures to risk factors (e.g., asset price declines) causing mark-to-market losses for all holders, potentially triggering asset fire sales and amplification of volatility. - Collateral-driven spillover channels via cross-sector participation in wholesale financing markets (derivatives trading, securities lending, repo): - Collateral mark-to-market losses can trigger margin calls requiring additional collateral. - Asset fire-sales to meet margin calls can amplify price declines and volatility. - Procyclical investment behavior in sectors (e.g., insurance and pension funds) can further amplify shocks. - Role of market-based data and methodological choices: - Cross-sector approaches typically rely on market-based data (performance indices, CDS spreads) as a useful complement to regulatory and qualitative data. - Limitations of market-based data: reduced trading volumes and liquidity can limit information content. - The analysis compares results across various methods to ensure consistency. - The analysis is illustrative and limited to banking and life insurance sectors. ### Modeling the system’s portfolio multivariate density - Structural Approach (SA) conceptualization: - The financial system is conceptualized as a portfolio of financial entities (FEs) spanning sectors. - The Structural Approach models a borrower’s asset value evolving stochastically; distress/default is triggered when asset value falls below a threshold determined by the FE’s financial structure. - Probability of Distress (PoD) of an entity is the likelihood its asset value falls below its distress threshold. - Portfolio Multivariate Density (PMD) and CIMDO: - A portfolio multivariate density (PMD) describing the joint likelihood of distress of all FEs is recovered via Consistent Information Multivariate Density Optimization (CIMDO). - Key facets of CIMDO: - Uses a prior parametric distribution (e.g., multivariate t) consistent with economic intuition and the SA. - Imposes consistency-constraint equations on marginal densities so the posterior matches empirically observed PoDs. - Recovers the distress-dependence copula simultaneously when inferring the posterior density; no separate copula calibration is required. - The dependence structure is allowed to dynamically adjust to changes in PoDs; increases in PoDs are reflected in increased dependence, potentially nonlinearly. - PoDs are exogenous inputs to the CIMDO framework and can be estimated using Merton-type models, CDS, out-of-the-money option prices, or supervisory information. In this analysis PoDs are backed out from five-year CDS spreads. - Simulation of systemic losses and tail measures: - Once the PMD is recovered, a systemic loss distribution is simulated. - Tail-risk measures computed from the simulated loss distribution (as used in this analysis): - Value at Risk (α-VaR): the smallest number ξ such that the probability that the loss exceeds ξ is not greater than α. - Expected Shortfall (ES): the expected value of the α-tail distribution. In the ES computation for this analysis α = 0.01 (1 percent). - Measures derived from PMD and simulated losses used in the illustration: - Joint Probability of Distress (JPoD): likelihood that all FEs in the portfolio/system are in distress. In the empirical illustration, distress thresholds delineate 1 percent tail of the distribution. - Conditional Probability of Distress Spillovers (Distress Dependence Matrix, DiDe): at each point in time reports the probability of the entity specified in the row falling into distress given the entity specified in the column is in distress. DiDe can be generalized to sectoral-level conditional probabilities. Conditional probabilities do not imply causation but provide insights into interconnectedness. - Tail Risk Index (TRI): systemic ES recorded at each point in time and normalized between zero and unity by deflating by the maximum ES recorded over the period (the peak level during the financial crisis is taken as the point of normalization). - Implementation notes: - The PMD modeling is robust under restricted data environments according to probability integral transform criteria. - The CIMDO objective minimizes Kullback (1959) minimum cross-entropy (minimum probabilistic divergence) from the prior subject to marginal PoD constraints, yielding a posterior multivariate density consistent with available PoD information. ### Box 2 — Monte Carlo simulation of systemic loss distribution (procedure and loss mapping) - Once the PMD is recovered, a Monte Carlo simulation is performed to generate the system’s loss distribution. - In addition to the binary case (distressed or not distressed), an FE can also experience losses if its risk quality deteriorates with respect to its current state. Losses are mapped to returns if they fall into a decay zone (lower risk quality zone). - If a return falls in the decay zone, then a loss will be assigned to this return, which is proportional to the severity of the return. - Loss definition and inputs: - For each FE i loss is defined as: Li,t = LGDi,t × EADi,t, - where LGDi,t is the loss given default, and EADi,t represents the exposure at default of the system to a given FE i. - The EADi,t is quantified by the total assets of each FE i, at a given time t. ### Key empirical findings from the chapter (systemic risk and sector interconnectedness) - Overall systemic risk has declined to pre-crisis levels; the Tail Risk Index is close to 40 percent of its peak value reached around the global financial crisis (GFC). - SRISK at the end of the sample period stands around US$300 billion, close to half of its GFC level. - The (average) probability that distress in any one bank spills over to the rest of the banking sector is currently around four times less than peaks reached around the global financial crisis and also the euro area crisis. - At end-2013, the share of securities issued by banks represented 16 percent of (life) insurers’ corporate bond portfolios and 10 percent of their equity portfolio. - The insurance sector’s likelihood of going into distress given distress in the banking sector has remained roughly stable over the past five years after peaking during the global financial crisis. - The likelihood of the banking sector going into distress conditional on distress in the insurance sector peaked around the euro area crisis and has displayed considerable variability since the financial crisis. - The prolonged low interest rate environment is identified as an overarching structural factor increasing vulnerability of both banks and insurers. - Using the Diebold and Yilmaz (2015) connectedness framework, the percentage of forecast error variation of banking (insurance) sector volatility explained by shocks from the insurance (banking) sector has been trending upward over time. - A similar upward trend in volatility spillovers was found using daily log returns in equity prices as inputs for the framework. *Prepared by IMF staff as part of the United Kingdom FSAP technical analysis.* ### EXECUTIVE SUMMARY ___________________________________________________________________________ 4 ### EXECUTIVE SUMMARY ### Assessment scope and purpose - Summarizes the assessment of interconnectedness and systemic risk for the United Kingdom (U.K.) financial system conducted as part of the Financial Sector Assessment Program (FSAP). - Analysis consists of three parts: - Motivation for monitoring cross-sector interconnectedness as part of financial system resilience assessment. - Description of selected empirical methods to analyze interconnectedness. - An illustrative analysis based on a definition of the financial system that incorporates U.K. banking and life insurance sectors. - Time span for the illustrative analysis: August 2007–January 2016. ### Key findings on systemic risk and sector resilience - Overall systemic risk in the U.K. has declined to pre-crisis levels. This reflects, to a significant degree, the improved sector-specific resilience established under sector-by-sector approaches applied in this FSAP. - Improved banking sector resilience is verified using metrics spanning market-based as well as balance sheet information. - The likelihood of distress spilling over among individual banks is currently subdued, with little heterogeneity in this measure across banks. - Interconnectedness within the banking sector has been declining over recent years. - Comparable measures between banking and life insurance sectors have displayed more variability: - Using the likelihood of distress spillovers between sectors, the banking–life insurance interconnectedness shows variability. - Using alternative methodologies based on analyzing spillovers of asset price volatility, evidence of steadily increasing interconnectedness between these sectors is found. - The analysis is essentially market-data based and illustrative, covering banks and life insurers only. ### Data coverage—entities included - Banks included: Barclays, HSBC, RBS, Lloyd’s, and Standard Chartered. - Insurers (life only) included: Aviva, Legal & General, Prudential Plc, and Standard Life. - PoD inputs: PoD estimates backed out from five-year CDS spreads for each financial entity; these data are available on a daily frequency. --- ### MONITORING CROSS-SECTOR INTERCONNECTEDNESS: SOME BACKGROUND ### Motivation and channels of spillovers - Firm-by-firm or sector-by-sector approaches analyze resilience at a point in time; structural market changes continually give rise to new channels of interconnectedness that can transmit or amplify shocks. - Interconnectedness channels: - Direct: counterparty transactions (e.g., interbank lending) creating counterparty credit risk where an insolvent firm imposes losses on counterparties. - Indirect: common exposures to risk factors (e.g., asset price declines) causing mark-to-market losses for all holders, potentially triggering asset fire sales and amplification of volatility. - Cross-sector participation in wholesale financing markets (derivatives trading, securities lending, repo) creates collateral-driven spillover channels: - Collateral mark-to-market losses can trigger margin calls requiring additional collateral. - Asset fire-sales to meet margin calls can amplify price declines and volatility. - Procyclical investment behavior in sectors (e.g., insurance and pension funds) can further amplify shocks. ### Role of market-based data and methodological choices - Cross-sector approaches typically rely on market-based data (performance indices, CDS spreads) as a useful complement to regulatory and qualitative data. - Limitations of market-based data: reduced trading volumes and liquidity can limit information content. - The analysis compares results across various methods to ensure consistency. - The analysis is illustrative and limited to banking and life insurance sectors. --- ### MODELING THE SYSTEM’S PORTFOLIO MULTIVARIATE DENSITY ### Structural Approach (SA) conceptualization - The financial system is conceptualized as a portfolio of financial entities (FEs) spanning sectors. - The Structural Approach models a borrower’s asset value evolving stochastically; distress/default is triggered when asset value falls below a threshold determined by the FE’s financial structure. - Probability of Distress (PoD) of an entity is the likelihood its asset value falls below its distress threshold. ### Portfolio Multivariate Density (PMD) and CIMDO - A portfolio multivariate density (PMD) describing the joint likelihood of distress of all FEs is recovered via Consistent Information Multivariate Density Optimization (CIMDO). - Key facets of CIMDO: - Uses a prior parametric distribution (e.g., multivariate t) consistent with economic intuition and the SA. - Imposes consistency-constraint equations on marginal densities so the posterior matches empirically observed PoDs. - Recovers the distress-dependence copula simultaneously when inferring the posterior density; no separate copula calibration is required. - The dependence structure is allowed to dynamically adjust to changes in PoDs; increases in PoDs are reflected in increased dependence, potentially nonlinearly. - PoDs are exogenous inputs to the CIMDO framework and can be estimated using Merton-type models, CDS, out-of-the-money option prices, or supervisory information. In this analysis PoDs are backed out from five-year CDS spreads. ### Simulation of systemic losses and tail measures - Once the PMD is recovered, a systemic loss distribution is simulated. - Tail-risk measures computed from the simulated loss distribution (as used in this analysis): - Value at Risk (α-VaR): the smallest number ξ such that the probability that the loss exceeds ξ is not greater than α. - Expected Shortfall (ES): the expected value of the α-tail distribution. In the ES computation for this analysis α = 0.01 (1 percent). - Measures derived from PMD and simulated losses used in the illustration: - Joint Probability of Distress (JPoD): likelihood that all FEs in the portfolio/system are in distress. In the empirical illustration, distress thresholds delineate 1 percent tail of the distribution. - Conditional Probability of Distress Spillovers (Distress Dependence Matrix, DiDe): at each point in time reports the probability of the entity specified in the row falling into distress given the entity specified in the column is in distress. DiDe can be generalized to sectoral-level conditional probabilities. Conditional probabilities do not imply causation but provide insights into interconnectedness. - Tail Risk Index (TRI): systemic ES recorded at each point in time and normalized between zero and unity by deflating by the maximum ES recorded over the period (the peak level during the financial crisis is taken as the point of normalization). ### Implementation notes - The PMD modeling is robust under restricted data environments according to probability integral transform criteria (referenced methodology). - The CIMDO objective minimizes Kullback (1959) minimum cross-entropy (minimum probabilistic divergence) from the prior subject to marginal PoD constraints, yielding a posterior multivariate density consistent with available PoD information. --- *Prepared by IMF staff as part of the United Kingdom FSAP technical analysis.* ### Box 2. Monte Carlo Simulation of Systemic Loss Distribution ### Box 2. Monte Carlo Simulation of Systemic Loss Distribution ### Monte Carlo simulation procedure and loss mapping - Once the PMD is recovered, a Monte Carlo simulation is performed to generate the system’s loss distribution. random numbers are generated, and for every simulation two cases are considered: - If ௜௜ then the FE is in distress, and ቀ ିஶ,௑ ೏ ೔ ቃ . - If ௜௜ then the FE is not in distress, and ቀ ିஶ,௑ ೏ ೔ ቃ . - In addition to the binary case (distressed or not distressed), an FE can also experience losses if its risk quality deteriorates with respect to its current state. Losses are mapped to returns if they fall into a decay zone (lower risk quality zone). - If a return falls in the decay zone, then a loss will be assigned to this return, which is proportional to the severity of the return. If the decay threshold for a given FE is defined as ௜ ௗ௘௖௔௬ then the random variable ௜ can be assigned as follows: - Here, ௜ is the cumulative distribution function of the returns of the FE. - Example: Loss threshold ### Loss definition and inputs - For each FE i loss is defined as: ௜௜௜ , - where ௜ is the loss given default, and ௜ represents the exposure at default of the system to a given FE i . - The ௜ is quantified by the total assets of each FE i , at a given time t. ### Key empirical findings from the chapter (systemic risk and sector interconnectedness) - Overall systemic risk has declined to pre-crisis levels; the Tail Risk Index is close to 40 percent of its peak value reached around the global financial crisis (GFC). - SRISK at the end of the sample period stands around US$300 billion, close to half of its GFC level. - The (average) probability that distress in any one bank spills over to the rest of the banking sector is currently around four times less than peaks reached around the global financial crisis and also the euro area crisis. - At end-2013, the share of securities issued by banks represented 16 percent of (life) insurers’ corporate bond portfolios and 10 percent of their equity portfolio. - The insurance sector’s likelihood of going into distress given distress in the banking sector has remained roughly stable over the past five years after peaking during the global financial crisis. - The likelihood of the banking sector going into distress conditional on distress in the insurance sector peaked around the euro area crisis and has displayed considerable variability since the financial crisis. - The prolonged low interest rate environment is identified as an overarching structural factor increasing vulnerability of both banks and insurers. - Using the Diebold and Yilmaz (2015) connectedness framework, the percentage of forecast error variation of banking (insurance) sector volatility explained by shocks from the insurance (banking) sector has been trending upward over time. - A similar upward trend in volatility spillovers was found using daily log returns in equity prices as inputs for the framework. *Source: Box 2 and related text, "CROSS-SECTOR INTERCONNECTEDNESS AND SYSTEMIC RISK: AN EMPIRICAL ILLUSTRATION", UNITED KINGDOM, INTERNATIONAL MONETARY FUND.* --- _Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16164.pdf_