Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
SlideShare a Scribd company logo

Why Stress Testing Is in the News

•Download as PPTX, PDF•
•
Barry Schachter
Barry Schachter

Regulators have seized on stress testing as central to controlling bank risk and systemic risk. Find out what financial firms have to do, what regulators hope to accomplish, and where the efforts may fall short.

1 of 39
Regulatory Requirements for Stress Testing Barry Schachter Courant Institute, NYU (March 5, 2013)
Required Risk Measurement Stress Reverse Tests Stressed VaR Scenario Stress Tests Internal Standard Model (VaR) Model
Regulatory Stress Testing Requirements for Capital Adequacy Liquidity Threats to Viability Systemic Risk Disclosure Other Risk Management Purposes
Liquidity
Basel III Liquidity Stress Test  Liquidity Coverage Ratio High quality liquid assets Cash outflows under stress (30 days)
Assets Liquid Under Stress  Able to sell, borrow against with no loss even under severe market stress.  Level 1: Cash, central bank reserves, securities backed by sovereigns and CBs  In unlimited amounts  Level 2: covered bonds, corporate debt, RMBS, some equities, some government securities  In limited amounts with haircuts
Elements of Liquidity Stress Scenario  Run on retail deposits  Partial loss of wholesale funding  Margin calls from 3 step credit downgrade  Drawdowns of committed lending lines
Liquidity Stress Scenario  Set out as “minimum” supervisory requirement  LCR requirement must be satisfied “continuously” in normal times  Banks must do additional scenarios  Reflect business-specific activity  Over longer time horizons  Share results with supervisors Basel III: International framework for liquidity risk measurement, standards and monitoring (December 2010/January 2013) (bcbs188 and bcbs238)
Operational Considerations  Assumptions – size of run-offs and other impacts  Data – detailed collateral scenario forecasts  Operational risks – integrating systems to aggregate liquidity impacts  Liquidity adjustments to pricing  Consistency with liquidity assumptions in VaR
Capital Adequacy
Rules Apply to Who? Dodd-Frank (section 165): Financial HCos., Banks, S&Ls, SIFIs >$10 billion assets, SIFI >$50 billion CCAR: Bank HCos. >$ 50 billion assets >$100
What Tests? CCAR and D-F  Dodd-Frank (Sect. 165)  Annual supervisory stress tests for “SIFIs”, Banks>$50B in assets  Annual company-run stress tests for Banks and Finc. HCos. >$10B  Semi-annual company-run stress testing if >$50B in assets  Company-run scenarios include a severely adverse scenario  Minimum of three supervisory scenarios  Mandated public disclosure  CCAR  Annual supervisory severe adverse stress test for large, complex bank HCos., >$100B in assets originally, now >$50B
Number of Affected Firms as of Q3 2012 120 100 Bank HoldCos > $100B 80 Fin HoldCos >$10B 60 *Fin HoldCos >$50B 40 Banks & S&Ls >$10B 20 SIFIs 0 Categories Source: FFIEC, FDIC, Federal Reserve
Who‟s In the 2012-2013 Cycle? CCAR 19
Who‟s In the 2012-2013 Cycle? CapPR 11 Other US BHC‟s >$50B in Assets
D-F/CCAR Stress Testing Cycle November October 1 January 5 March15 15 Fed & Banks Run Fed Runs Biggest Public Banks Scenarios Own Reporting Banks‟ Develop Analyzes of Results Scenarios Capital Mid- cycle Process
Stress Test Creation  Shocks to 26 (for 2012-2013) macroeconomic variables,  Translated into impact on items of income and expense, and  Flow those imacts through quarterly net income  Add to forecast changes in the capital account to estimate impact on regulatory capital  Trading books subject to a separate, instantaneous shock
Macroeconomic Shocks 2013 Supervisory Scenarios 2012-2013 DJ Total Stock Market 20000 18000 16000 14000 12000 11042.3 10000 8000 7221.7 6000 4000 2000 0
From macro shocks to gains and losses  Crucial! Two ways to go: implied relationship historical relationship See, e.g., CCAR 2012: Methodology and Results for Stress Scenario Projections, Appendix A
Market Shocks – Severe Adverse - Equity Equity by Geography 0% -15% New Zealand Switzerland United Kingdom Sweden Chile -30% Germany Mexico Philippines Turkey Australia Euro Stoxx 50 Malaysia India United States 600 South AfricaIndex FranceJapan Stoxx Europe Poland South Korea MSCIMSCI World China Hong Kong EAFE Belgium Finland Canada MSCI All Country -45% Denmark Hungary Taiwan Index Netherlands Czech RepublicBrazil Indonesia IsraelWorld Index Singapore Greece (ACWI) MSCI EM Index Italy Argentina Norway MSCI EMEA -60%Austria MSCI EM Latin Spain Index America Index -75% Bulgaria Russia Ireland Portugal Ukraine, -84.30% -90%
Supervisory Scenario – Market Shocks to FX Vol
ATM USDMXN Vol History Supervisor Shock (vol pts)
Stress Testing under CCAR & D-F  The details  9 quarter horizon, instantaneous shock for trading books  As-of dates  9/30 (and 3/31) for annual (mid-cycle) tests  Random end-Q4 as-of date for trading books  Implementation decisions  Interpolation, extrapolation, proxying shocks  Bank baseline scenario – adopt supervisory baseline?  Models don‟t work with shocks  Documentation  Must use standard templates  must document and justify deviations
Disclosure & Systemic Risk Assessment
Disclosure under D-F  From banks to the supervisor  FR Y-14M (and 14A and 14Q)  From the supervisor to the public  Details of shocks  Using both D-F and bank capital plan assumptions  Worst quarterly and end-of-horizon capital ratios for severe adverse scenario  Using D-F capital plan assumptions  Total 9 quarter net revenue and income, loan losses, trading and counterparty credit losses
Disclosure under D-F  From Banks to the public  types of risks included;  description of company scenarios (key variables, such as GDP, unemployment rate, housing prices);  description of the methodologies to estimate losses, revenues, and changes in capital; and  net revenue and income, pro forma capital and capital ratios over the planning horizon.
Disclosure under Form PF  Private Funds (SEC and CFTC regulated)  Investment advisor or CTA/CPO  >$150MM “regulatory assets” (or >$1.5 Billion “large funds”)  Reporting  Large funds – 60 days after end of quarter  Others – 120 days after end of year
Disclosure under Form PF  May use internal methodologies  Must not be inconsistent with instructions  Must be consistent with internal/client information  Do not net longs and shorts  Value is notional amount (delta-adjusted)
Form PF – Stress Shocks – Rates Yield in Pct. Stress Tested US Yield Curve Using SEC Form PF Shocks 4.5 4 3.5 3 2.5 2 Base Case (3, March 2013) up 25bps 1.5 dn 25bps 1 up 75bps 0.5 dn 75bps 0 0 36 72 108 144 180 216 252 288 324 360 -0.5 -1 Months
Threats to Viability
Reverse Stress Testing  Supervisory Guidance (SR 12-7)  Financial HCos., Banks, S&Ls >$10 billion assets  CEBS (GL32)  FSA (CP 08/24, PS 09/20)
Reverse Stress Testing  New concept for banks  FSA (UK) leading in requiring this  Regulators like it but are silent on method
Reverse Stress Testing Defined  “assume a known adverse outcome…  “then deduce the types of events that could lead to such an outcome.” (Federal Reserve Board SR 12- 7, p. 12)  Sounds simple, but…
Reverse Stress Testing History FSA FSA Proposed BIS Principles Requirements US Supervisory CRMPG III Rules for Sound UK banks go Formalized Guidance 2008-08 (CP08/24) Stress Testing live 2011-12 (PS09/20) 2012-11 2008-12 2009-05 2010-12
Reverse Stress Testing Illustration Mirzai & Müller. 2013. On Reverse Stress Testing. Intelligent Risk, 8-11.  Pick critical loss level („CLL‟)  Obtain 30K samples from joint distribution of risk factor returns  Revalue portfolio 30K times  Only look at those samples where portfolio loss > CLL By CaitlinJo [CC-BY-3.0 (http://creativecommons.org/licenses/by /3.0)], via Wikimedia Commons
Reverse Stress Testing Illustration Mirzai & Müller. 2013. On Reverse Stress Testing. Intelligent Risk, 8-11.  Can we find economic meaning in these large loss samples?  First, look for statistical commonalities using k-means clustering  Second, relate cluster behavior to market risk factors through heuristic analysis
Reverse Stress Testing Illustration Mirzai & Müller. 2013. On Reverse Stress Testing. Intelligent Risk, 8-11.
Reverse Stress Testing In sum: immature art Recent research collected at www.Gloria- Mundi.com; search on keyword phrase reverse stress.
Why Stress Testing Is in the News

More Related Content

Why Stress Testing Is in the News

  • 1. Regulatory Requirements for Stress Testing Barry Schachter Courant Institute, NYU (March 5, 2013)
  • 2. Required Risk Measurement Stress Reverse Tests Stressed VaR Scenario Stress Tests Internal Standard Model (VaR) Model
  • 3. Regulatory Stress Testing Requirements for Capital Adequacy Liquidity Threats to Viability Systemic Risk Disclosure Other Risk Management Purposes
  • 5. Basel III Liquidity Stress Test  Liquidity Coverage Ratio High quality liquid assets Cash outflows under stress (30 days)
  • 6. Assets Liquid Under Stress  Able to sell, borrow against with no loss even under severe market stress.  Level 1: Cash, central bank reserves, securities backed by sovereigns and CBs  In unlimited amounts  Level 2: covered bonds, corporate debt, RMBS, some equities, some government securities  In limited amounts with haircuts
  • 7. Elements of Liquidity Stress Scenario  Run on retail deposits  Partial loss of wholesale funding  Margin calls from 3 step credit downgrade  Drawdowns of committed lending lines
  • 8. Liquidity Stress Scenario  Set out as “minimum” supervisory requirement  LCR requirement must be satisfied “continuously” in normal times  Banks must do additional scenarios  Reflect business-specific activity  Over longer time horizons  Share results with supervisors Basel III: International framework for liquidity risk measurement, standards and monitoring (December 2010/January 2013) (bcbs188 and bcbs238)
  • 9. Operational Considerations  Assumptions – size of run-offs and other impacts  Data – detailed collateral scenario forecasts  Operational risks – integrating systems to aggregate liquidity impacts  Liquidity adjustments to pricing  Consistency with liquidity assumptions in VaR
  • 11. Rules Apply to Who? Dodd-Frank (section 165): Financial HCos., Banks, S&Ls, SIFIs >$10 billion assets, SIFI >$50 billion CCAR: Bank HCos. >$ 50 billion assets >$100
  • 12. What Tests? CCAR and D-F  Dodd-Frank (Sect. 165)  Annual supervisory stress tests for “SIFIs”, Banks>$50B in assets  Annual company-run stress tests for Banks and Finc. HCos. >$10B  Semi-annual company-run stress testing if >$50B in assets  Company-run scenarios include a severely adverse scenario  Minimum of three supervisory scenarios  Mandated public disclosure  CCAR  Annual supervisory severe adverse stress test for large, complex bank HCos., >$100B in assets originally, now >$50B
  • 13. Number of Affected Firms as of Q3 2012 120 100 Bank HoldCos > $100B 80 Fin HoldCos >$10B 60 *Fin HoldCos >$50B 40 Banks & S&Ls >$10B 20 SIFIs 0 Categories Source: FFIEC, FDIC, Federal Reserve
  • 14. Who‟s In the 2012-2013 Cycle? CCAR 19
  • 15. Who‟s In the 2012-2013 Cycle? CapPR 11 Other US BHC‟s >$50B in Assets
  • 16. D-F/CCAR Stress Testing Cycle November October 1 January 5 March15 15 Fed & Banks Run Fed Runs Biggest Public Banks Scenarios Own Reporting Banks‟ Develop Analyzes of Results Scenarios Capital Mid- cycle Process
  • 17. Stress Test Creation  Shocks to 26 (for 2012-2013) macroeconomic variables,  Translated into impact on items of income and expense, and  Flow those imacts through quarterly net income  Add to forecast changes in the capital account to estimate impact on regulatory capital  Trading books subject to a separate, instantaneous shock
  • 18. Macroeconomic Shocks 2013 Supervisory Scenarios 2012-2013 DJ Total Stock Market 20000 18000 16000 14000 12000 11042.3 10000 8000 7221.7 6000 4000 2000 0
  • 19. From macro shocks to gains and losses  Crucial! Two ways to go: implied relationship historical relationship See, e.g., CCAR 2012: Methodology and Results for Stress Scenario Projections, Appendix A
  • 20. Market Shocks – Severe Adverse - Equity Equity by Geography 0% -15% New Zealand Switzerland United Kingdom Sweden Chile -30% Germany Mexico Philippines Turkey Australia Euro Stoxx 50 Malaysia India United States 600 South AfricaIndex FranceJapan Stoxx Europe Poland South Korea MSCIMSCI World China Hong Kong EAFE Belgium Finland Canada MSCI All Country -45% Denmark Hungary Taiwan Index Netherlands Czech RepublicBrazil Indonesia IsraelWorld Index Singapore Greece (ACWI) MSCI EM Index Italy Argentina Norway MSCI EMEA -60%Austria MSCI EM Latin Spain Index America Index -75% Bulgaria Russia Ireland Portugal Ukraine, -84.30% -90%
  • 22. ATM USDMXN Vol History Supervisor Shock (vol pts)
  • 23. Stress Testing under CCAR & D-F  The details  9 quarter horizon, instantaneous shock for trading books  As-of dates  9/30 (and 3/31) for annual (mid-cycle) tests  Random end-Q4 as-of date for trading books  Implementation decisions  Interpolation, extrapolation, proxying shocks  Bank baseline scenario – adopt supervisory baseline?  Models don‟t work with shocks  Documentation  Must use standard templates  must document and justify deviations
  • 25. Disclosure under D-F  From banks to the supervisor  FR Y-14M (and 14A and 14Q)  From the supervisor to the public  Details of shocks  Using both D-F and bank capital plan assumptions  Worst quarterly and end-of-horizon capital ratios for severe adverse scenario  Using D-F capital plan assumptions  Total 9 quarter net revenue and income, loan losses, trading and counterparty credit losses
  • 26. Disclosure under D-F  From Banks to the public  types of risks included;  description of company scenarios (key variables, such as GDP, unemployment rate, housing prices);  description of the methodologies to estimate losses, revenues, and changes in capital; and  net revenue and income, pro forma capital and capital ratios over the planning horizon.
  • 27. Disclosure under Form PF  Private Funds (SEC and CFTC regulated)  Investment advisor or CTA/CPO  >$150MM “regulatory assets” (or >$1.5 Billion “large funds”)  Reporting  Large funds – 60 days after end of quarter  Others – 120 days after end of year
  • 28. Disclosure under Form PF  May use internal methodologies  Must not be inconsistent with instructions  Must be consistent with internal/client information  Do not net longs and shorts  Value is notional amount (delta-adjusted)
  • 29. Form PF – Stress Shocks – Rates Yield in Pct. Stress Tested US Yield Curve Using SEC Form PF Shocks 4.5 4 3.5 3 2.5 2 Base Case (3, March 2013) up 25bps 1.5 dn 25bps 1 up 75bps 0.5 dn 75bps 0 0 36 72 108 144 180 216 252 288 324 360 -0.5 -1 Months
  • 31. Reverse Stress Testing  Supervisory Guidance (SR 12-7)  Financial HCos., Banks, S&Ls >$10 billion assets  CEBS (GL32)  FSA (CP 08/24, PS 09/20)
  • 32. Reverse Stress Testing  New concept for banks  FSA (UK) leading in requiring this  Regulators like it but are silent on method
  • 33. Reverse Stress Testing Defined  “assume a known adverse outcome…  “then deduce the types of events that could lead to such an outcome.” (Federal Reserve Board SR 12- 7, p. 12)  Sounds simple, but…
  • 34. Reverse Stress Testing History FSA FSA Proposed BIS Principles Requirements US Supervisory CRMPG III Rules for Sound UK banks go Formalized Guidance 2008-08 (CP08/24) Stress Testing live 2011-12 (PS09/20) 2012-11 2008-12 2009-05 2010-12
  • 35. Reverse Stress Testing Illustration Mirzai & Müller. 2013. On Reverse Stress Testing. Intelligent Risk, 8-11.  Pick critical loss level („CLL‟)  Obtain 30K samples from joint distribution of risk factor returns  Revalue portfolio 30K times  Only look at those samples where portfolio loss > CLL By CaitlinJo [CC-BY-3.0 (http://creativecommons.org/licenses/by /3.0)], via Wikimedia Commons
  • 36. Reverse Stress Testing Illustration Mirzai & Müller. 2013. On Reverse Stress Testing. Intelligent Risk, 8-11.  Can we find economic meaning in these large loss samples?  First, look for statistical commonalities using k-means clustering  Second, relate cluster behavior to market risk factors through heuristic analysis
  • 37. Reverse Stress Testing Illustration Mirzai & Müller. 2013. On Reverse Stress Testing. Intelligent Risk, 8-11.
  • 38. Reverse Stress Testing In sum: immature art Recent research collected at www.Gloria- Mundi.com; search on keyword phrase reverse stress.

Editor's Notes

  1. a U.S. BHC that is owned and controlled by a foreign bank that is an FHC that the Board has determined to be well-capitalized and well-managed will not be required to comply with the Board's capital adequacy guidelines.” E.g., Deutschebank.
  2. Capital Plan Review (CapPR)Annual Company-Run Stress Test Requirements Are Delayed Until Fall of 2013 Fornational banks, savings associations and state non-member banks with average total consolidated assets of less than $50 billion; (no reporting for another year after that)BHCs with average total consolidated assets of less than $50 billion U.S.-domiciled BHC subsidiaries of FBOs relying on SR 01-1 for which implementation begins 2015; andstate member banks with average total consolidated assets greater than $10 billion (other than state member bank subsidiaries of SCAP BHCs).
  3. 38,975 separate shocks, benchmarked to 2nd half 2008. what about basket products?
  4. Model conflicts; vol smile impact; flat extrapolation
  5. RAUM are the gross assets under management, without the subtraction of borrowings, short sales or other forms of leverage.
  6. Parallel; Negative rates; Definition of risk-free
  7. Reasons given for RST:Outside normal business thinkingExplore hidden vulnerabilitiesChallenge assumptions about threats to business viability