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.
Transcript of Why Stress Testing Is in the News
1. Regulatory Requirementsfor 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 TestingRequirements for Capital Adequacy
Liquidity Threats to Viability Systemic Risk Disclosure Other Risk
Management Purposes
4. Liquidity
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 StressScenario 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
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 Firmsas of Q3 2012 120100 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. Whos In the 2012-2013 Cycle? CCAR 19
15. Whos In the 2012-2013 Cycle? CapPR 11 Other US BHCs
>$50B in Assets
16. D-F/CCAR Stress Testing Cycle NovemberOctober 1 January 5
March15 15Fed & Banks Run Fed Runs Biggest PublicBanks
Scenarios Own Reporting BanksDevelop Analyzes of ResultsScenarios
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
19. From macro shocks to gainsand losses Crucial! Two ways to
go: implied relationship historical relationshipSee, e.g., CCAR
2012: Methodology and Results for Stress ScenarioProjections,
Appendix A
20. Market Shocks SevereAdverse - 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%
21. Supervisory Scenario Market Shocks to FX Vol
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 dont work with shocks Documentation
Must use standard templates must document and justify
deviations
24. Disclosure & SystemicRisk Assessment
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
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 TestingDefined 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 SupervisoryCRMPG 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 TestingIllustrationMirzai & Mller. 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 TestingIllustrationMirzai & Mller. 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