ECONOMIC AND REGULATORY CAPITAL

Transcription

1 ECONOMIC AND REGULATORY CAPITAL Bank Indonesia Bali 21 September 2006 Presented by David Lawrence

2 OpRisk Advisory Company Profile Copyright , OpRisk Advisory. All rights reserved. 2

3 DISCLAIMER All comments presented here represent the views of the speaker and are not necessarily those of any particular financial institution. All the examples have been prepared for teaching purposes only and the numbers therein are not necessarily those used by any industry participant. David Lawrence Executive Director OpRisk Advisory David Lawrence Executive Director Lawrence Risk Consulting

4 Economic Capital

5 Primary Objectives Risk Capital and Capital Allocation Establish a consistent and comprehensive risk management tool Quantify all major risk types Apply across all businesses and regions Consistent with Basel II requirements Establish a consistent and comprehensive return on capital framework Consistent, risk-adjusted measure of business performance Optimise investment of capital short-term and long-term Balance growth and returns Client selection and product management Acquisitions and divestitures 5

6 Economic Risk Capital Definition The amount of capital needed to cover unexpected economic losses during extreme events Measured as the potential unexpected economic losses over a one year time period at a 99.97% confidence level Extreme 3 in 10,000 risk event Economic losses include Losses on the income statement Further decline in the value of assets Further increase in the value of liabilities Expected losses are covered through profit margins 6

7 Economic Capital Frequency of loss 99.97% Confidence level Expected loss Unexpected loss Stress loss Hedging techniques Operating Expense Capital Financing Amount of loss Transfer / Accept 7

8 RISK TYPES The drivers of economic losses are risks,, which may be broadly categorized as: Market Risk Trading market risk Accrual (ALM) interest rate risk Proprietary Investment Risk Credit Risk Corporate lending risk Counterparty risk Retail credit risk Cross-Border/In Border/In-Border Country Risk Business Volume Risk (risk of customer revenue not covering fixed expenses) Operational Risk Insurance Liability Risk 8

9 Risk Capital and Capital Allocation SUMMARY Integrated risk-based capital management system Risk quantification more comprehensive than Basel II Consistent measure of risk-adjusted returns for internal growth Disciplined measure of capital investment for acquisitions Growth and returns balanced to maximise long-term shareholder value 9

10 Capital for Market Risk

11 VALUE AT RISK MULTIPLE MARKET FACTORS The Value at Risk is a statistical estimate of the amount that could be lost over a given period (e.g. one day) at a given confidence level (e.g. 99%), due to correlated changes in the underlying market factors. The volatilities and correlations of the market factors are calculated from three years historical data, except where recent markets movements dictate otherwise. 11

12 VALUE AT RISK VaR combines exposure with probability of market moves Exposure P&L Probability of P&L moves % moves in underlying frequency Probability of market moves frequency % moves in underlying P&L VAR 12

13 VALUE AT RISK 99% VaR is the maximum we expect to lose 99% of the time (99 out of 100 days)* frequency P&L 1 % of area VaR *However, 1% of the time we could lose a lot more! 13

14 B.I.S. CAPITAL Standard Model BANKING BOOK: FX, Commodities TRADING BOOK: TSIR, Equity, FX, Commodities STANDARD METHOD DEBT INSTRUMENTS Debt Securities Debt Derivatives Specific Risk General Market Risk EQUITY POSITIONS FOREIGN EXCHANGE RISK COMMODITY RISK 14

15 B.I.S. CAPITAL Internal Models BANKING BOOK: FX, Commodities TRADING BOOK: FX, Commodities, TSIR and Equity INTERNAL MODELS Variance-covariance covariance method Historic Simulation Monte Carlo Simulation CAPITAL REQUIREMENT VAR is the 99% confidence level 10-day VAR MAX [ VAR(Today), F x VAR(60-day average)] F = 3 to 4, depending upon backtesting. 15

16 FROM VALUE AT RISK TO CAPITAL RISK MANAGEMENT Set limits on VAR (99%, 1-day) 1 = 2.32 σ ECONOMIC CAPITAL EcCap (99.97%) = 1.47 * SQRT(250) x VAR(99%,1-day) = 54 σ REGULATORY CAPITAL VAR (99%, 10-day) = 3.16 x VAR(99%, 1-day) 1 = 7.37 σ MAX [ VAR(Today), F x VAR(60-day average)] F = 3 to 4, depending upon backtesting and specific risk model RegCap (99%) = 22 σ 16

17 Capital for Credit Risk

18 ECONOMIC CAPITAL FOR CREDIT RISK Monte Carlo simulation for a single counterparty At each point in time: Simulate the exposure of the counterparty portfolio Simulate a credit migration / credit default For credit migration, simulate the changes in credit spread For credit default, simulate the loss rate (net of recovery) Continue until required time horizon is reached Calculate the expected value of the economic loss Calculate the unexpected economic loss 18

19 Economic Capital Frequency of loss 99.xx% Confidence level Expected loss Unexpected loss Stress loss Hedging techniques Operating Expense Capital Financing Amount of loss Transfer / Accept 19

20 CREDIT RISK OF PORTFOLIO COMPLEX APPROACH Models to estimate the credit risk of the entire portfolio - all counterparties : PortfolioManager TM CreditMetrics TM CREDITRISK+ TM CreditPortfolioView TM KMV J.P.Morgan Credit Suisse McKinsey 20

21 EQUITY BASED APPROACH Used by CreditMetrics and KMV Default occurs if the value of the firm s s assets fall below the firm s s liabilities Market value of debt is the difference between the total value of the firm and its equity Simulation requires joint probabilities of default - estimated from the correlation of equity prices 21

22 Portfolio value in one year VaR 99% 16 VaR 99.9% 31 22

23 ACTUARIAL APPROACH This approach is represented by CreditRisk + developed by Credit Suisse It uses actuarial techniques to model the loss distribution function Joint default is modelled as a common behavior of multiple obligors 23

24 CREDIT DEFAULT LOSS DISTRIBUTION 1.8% 1.6% Expected loss Probability 1.4% 1.2% 1.0% 0.8% 0.6% 0.4% 0.2% 0.0% 99th Percentile loss level Loss Source: CSFB 24

25 ECONOMETRIC APPROACH This approach is represented by CreditPortfolioView developed by McKinsey & Co. It is based on a model that uses macroeconomic factors to determine the borrower s s default probability The portfolio loss distribution is constructed by Monte Carlo simulation methodology according to the historical covariance structure 25

26 Credit Risk Regulatory Capital

27 REGULATORY CAPITAL: BASEL I CREDIT RISK Risk Weighted Asset = Loan Principal X Counterparty Risk Weighting Counterparty Risk Weighting Government 0% OECD Bank 20% Other 100% 27

28 BASEL II : Pillar 1 approaches 28

29 Basel II: Standardised Risk Weightings CLAIM AAA/AA A BBB BB B <B Unrated Sovereign Banks (1) Banks (2) Corporates (1) Risk weighting based on risk weighting of sovereign. (2) Risk weighting based on the assessment of the individual bank. (3) Short dated claims (<= 3 months) one category more favourable. (4) Remove 50% cap for derivatives. [ BCBS 118: ] 29

30 Internal Ratings Based Approach FOUNDATION APPROACH PD 0.03% floor except sovereigns LGD EAD M 45% senior unsecured 75% subordinated unsecured 100% balance sheet 75% of undrawn commitments Off balance sheet: CEM / SM / IMM Maturity (optional) If no adjustment, then all assets are assumed to have a maturity of 2.5 years ADVANCED APPROACH PD 0.03% floor except sovereigns LGD Bank s s own estimate EAD Bank s s own estimate M Maturity MTM-based maturity adjustment Default-based adjustment Maturity capped at 5 years [BCBS 118: 244, 287, ] 324] 30

31 Basel II Requirement: Portfolio invariance The capital required for any given loan should only depend on the risk of that loan and must not depend on the portfolio it is added to This makes recognition of institution-specific specific diversification effects within the framework difficult Model is calibrated to a well-diversified bank, with concentration risk to be evaluated undwer Pillar II These are therefore ratings-based models, similar to the KMV approach for a single counterparty Only Asymptotic Single Risk Factor (ASRF) models are portfolio invariant, as these depend upon having a large number of relatively small exposures [Source: BIS ] 31

32 Corporate Asset classes Corporate Specialised lending Sovereign Bank Retail exposures secured by residential properties qualifying revolving retail exposures other retail exposures Equity [ BCBS 118: 215, 233 ] 32

33 RWA for Corporates,, Sovereigns and Banks R = 0.12 x (1 EXP(-50 x PD)) / (1 EXP(-50)) x [1 (1 EXP(-50 x PD)) / (1 EXP(-50))] b = ( x ln(pd))^2 K = [LGD x N[(1-R)^ R)^-0.5 x G(PD) + (R / (1-R))^0.5 x G(0.999)] - PD x LGD ] x (1 1.5 x b)^-1 1 x (1+(M-2.5) x b) RWA = K x 12.5 x EAD Where R is asset value correlation [BCBS 118: 272] N(x) is the cumulative normal distribution G(z) is the inverse cumulative normal distribution M is maturity b is a maturity adjustment factor 33

34 Adjustment for SMEs R = 0.12 x (1 EXP(-50 x PD)) / (1 EXP(-50)) x [1 (1 EXP(-50 x PD)) / (1 EXP(-50))] x (1 (S-5)/45) Where S is expressed as total annual sales in millions of Euros b = ( x ln(pd))^2 K = [LGD x N[(1-R)^ R)^-0.5 x G(PD) + (R / (1-R))^0.5 x G(0.999)] - PD x LGD ] x (1 1.5 x b)^-1 1 x (1+(M-2.5) x b) RWA = K x 12.5 x EAD where N(x) is the cumulative normal distribution [BCBS 118: 273] G(z) is the inverse cumulative normal distribution M is maturity b is a maturity adjustment factor 34

35 Correlation (R) = 0.15 Risk Weight Function Residential Mortgages Capital requirement (K) = LGD N[(1 - R)^-0.5 G(PD) + (R / (1 - R))^0.5 G(0.999)] - PD x LGD Risk-weighted assets = K x 12.5 x EAD [BCBS 118: 328] 35

36 Risk Weight Function - QRRE Correlation (R) = 0.04 Capital requirement (K) = LGD N[(1 - R)^-0.5 G(PD) + (R / (1 - R))^0.5 G(0.999)] - PD x LGD Risk-weighted assets = K x 12.5 x EAD [BCBS 118: 329] 36

37 Risk Weight Function Other Retail Correlation (R) = 0.03 (1 - EXP(-35 PD)) / (1 - EXP(-35)) [1 - (1 - EXP(-35 PD))/(1 - EXP(-35))] Capital requirement (K) = LGD N[(1 - R)^-0.5 G(PD) + (R / (1 - R))^0.5 G(0.999)] PD x LGD Risk-weighted assets = K x 12.5 x EAD [BCBS 118: 330] 37

38 Risk Weight vs Probability of Default 600% 500% 400% 300% 200% 100% 0% 0% 5% 10% 15% 20% Corp SME Retail Mtge QRRE 38

39 Capital for Operational Risk ECONOMIC AND REGULATORY

40 A Basic Indicator Approach Basel Approaches B Standardised Approach B1 Alternative Standardised Approach C Advanced Measurement Approach C1 Loss Distribution Approach C2 Risk Drivers and Controls Approach C3 Scenario-based Approach [ BCBS 118, Nov 2005 ] [ NewYorkFedConf,, May 2003] [ ] 40

41 A. Basic Indicator Approach K BIA = [ Σ (GI 1...n x α ) ] / n A fixed percentage (α)( ) of a single measure (Gross Income) GI n is the positive annual gross income in year n (n = 1 to 3) Basel 2 sets α at 15% Formula gives the average capital requirement over the previous three years for those years in which there was positive annual gross income [ BCBS 118: 649 ] 41

42 B. Standardised Approach K TSA = { Σ max [ Σ ( β i x GI i ), 0] } / 3 Each business line has its own multiplier (β( i ) of Gross Income Basel 2 set β in range 12% - 18% Business Line Multiplier Corporate finance 18 % Trading and sales 18 % Retail banking 12 % Commercial banking 15 % Payment and settlement 18 % Agency services 15 % Asset management 12 % Retail brokerage 12 % Negative capital charges in one business line can offset positive capital charges in another business line. The capital charge is i averaged over the previous three years, with a zero for any year in which the total capital charge was negative. [ BCBS 118: 654 ] 42

43 B1. Alternative Standardised Approach As for Standardised Approach, except for Retail Banking and Commercial Banking K RB = β RB x m x LA RB K RB is the capital charge for retail banking β RB is the beta for the retail banking line LA RB is the total outstanding retail loans and advances averaged over the past three years m = Similarly for commercial banking It is not envisaged that large diversified banks in major markets would use the ASA. [ BCBS 118: 652 footnote ] 43

44 Qualifying Criteria for STA Its board of directors and senior management, as appropriate, are actively involved in the oversight of the operational risk management framework. It has an operational risk management system that is conceptually sound and is implemented with integrity. It has sufficient resources in the use of the approach in the major business lines as well as the control and audit areas. [ BCBS 118: ] 44

45 The Standardised Approach Operational risk management function with clear responsibilities for developing strategies, for codifying policies and procedures, for implementing an assessment methodology and for implementing a risk-reporting reporting system Track operational risk data, including material losses by business line. Integrate the assessment system with the risk management processes Regular reporting of exposures, including losses System well documented System validated and independently reviewed Regular review by external auditors / supervisors [ BCBS 118: 663 ] 45

46 C. Basel Criteria for AMA Active involvement of board / senior management System is sound and implemented with integrity Sufficient resources in businesses, control and audit Independent operational risk management function Integral part of operational risk management, including economic capital Regular reporting to senior management / board Well documented system Reviewed by internal and external audit Validation by auditors and / or supervisors [ BCBS 118: 664, 666 ] 46

47 Basel Criteria for AMA Soundness comparable to IRB (99.9% one year) EL + UL, unless adequately capturing EL Sufficient granularity to capture the tail May use internally determined correlations, if they can be validated ated Internal data Relevant external data Scenario analysis Business environment and internal control factors Loss data history of five years; transition period three years May recognize risk mitigation through insurance, though with a capc Floor on total capital reduction versus Basel I Partial use; home-host host issues [BCBS 118: 667, 669, 672, , 678, 680 ] 47

48 Advanced Measurement Approach A A bank must be able to demonstrate that its approach captures potentially severe tail loss events. Whatever approach is used, a bank must demonstrate that its operational risk measure meets a soundness standard comparable to that of the internal ratings-based approach for credit risk, (i.e. comparable to a one year holding period and a 99.9th percentile confidence level). [ BCBS 118: 667 ] 48

49 COSO The Committee for Sponsoring Organisations of the Treadway Committee (COSO) released a framework that set the standards for enterprise-wide risk management. The essential steps in this frameworks are: Identification, definition and assessment of risks in all processes ses is carried by the businesses A typical risk-assessment implementation produces a huge catalogue of risks These are then ranked using the likelihood-impact impact method of risk assessment This methodology is severely flawed. [ ] 49

50 Modern ORM versus Traditional MODERN ORM (Basel II) TRADITIONAL (COSO) Frequency High (3) Med (2) Low (1) n/a n/a n/a Likelihood High (3) Med (2) Low (1) Phantom Risks Real Risks Low (1) Med (2) High (3) Impact Low (1) Med (2) High (3) Impact [Source: ORA ] 50

51 COSO Likelihood-Impact Likelihood Impact Loss 10 % 1, % 8, % 60, Note that this is analogous to Expected Loss, not Risk The high likelihood / low impact losses are usually well understood and well controlled The low likelihood / high impact losses are usually not well understood or well controlled 51

52 Loss Distribution Approach Simulate an aggregate potential loss distribution for operational risk using an actuarial method Drivers of the simulation model include: Probability distribution for N events [Frequency] Potential loss distribution given an event [Severity] These are obtained by fitting internal loss data Economic Capital requirements are calculated as the difference between the expected loss level and the potential loss level: At the target confidence Over the defined time horizon Split by business line and (if possible) by risk category Adjust for quality 52

53 Value-at at-risk Profile Annual Frequency Year # Losses Confidence Level Total Risk (Value-at-Risk Profile) Empirical distribution Poisson distribution Severity Distribution Aggregate Loss 53

54 Baseline OpRisk Economic Capital Frequency of loss 99.xx% Confidence level Expected loss Unexpected loss Stress loss Hedging techniques Operating Expense Capital Financing Amount of loss Transfer / Accept 54

55 Economic Capital compared to Regulatory Direct calculation of the 99.9% or 99.97% is not feasible, so calculation done at 95% and scaled up Expected Loss is subtracted to give Unexpected Loss, where regulatory is restricted to routine losses Correlation between major events is close to zero Diversification is taken into account, across geographies, business lines and risk types Home / host problems are based on management hierarchy rather than legal vehicle hierarchy 55

56 CORRELATION MATRIX BETWEEN RISK TYPES Add up over different risk classes and categories with correlation CREDIT MARKET ALM P&L OPRISK INS. Credit Market ALM P&L OpRisk Ins

57 International Best Practices TRADITIONAL VERSUS MODERN OPERATIONAL RISK MANAGEMENT

58 Event Frequency Traditional Under Modern ORM, total exposure has two components: expected loss (cost) and unexpected loss (risk); these definitions are consistent with Basel II. EXPECTED LOSSES High Frequency Low Impact Low Frequency High Impact UNEXPECTED LOSSES Effect Severity Catastrophic Probability 99.xx% Confidence level Modern Expected loss RISK Unexpected loss Catastrophic loss Hedging techniques Operating Expense Cumulative Annual Loss Capital Financing Transfer / Accept 58

59 Key differences between Traditional and Modern ORM. Traditional Definition: Risk is defined as a kind of unpleasant or undesirable event, such as a fraud or a system failure. Modern Definition: Risk is defined as a measure of exposure to losses at a specified confidence level, in excess of the mean. Risk Identification Process: Make a subjective determination about which risks are relevant. ( Risk can be causes, events or effects; no restriction on overlaps; no differentiation between risks and controls. ) Risk Identification Process: First define the risk universe, consisting of a finite (comprehensive) set of mutually exclusive (non-overlapping) risk classes. Use historical loss data to reveal where the losses (and hence the risks ) exist. Risk Assessment/Measurement Method: Calculate risk by multiplying likelihood and impact. (This yields the average or probability weighted severity.) Risk Assessment/Measurement Method: Calculate risk by measuring the aggregate exposure above the mean at a specified probability level (e.g., 99%). (This yields a kind of worst case aggregate loss. Analysis Period: Point in time analysis. What is measured: Loss from one specific incident (severity with frequency of 1). Analysis Period: Period of time analysis (e.g., one year). What is measured: What is measured: Aggregate loss from all incidents during one year (frequency and severity). 59

60 The path from Traditional to Modern ORM Adopt the Modern definition of risk. Don t t confuse risk types with control types. Recognize that identifying your risks means establishing a pre- defined universe of risks classes and letting the data tell you where the high risks exist. Conduct risk assessment independently from control assessment. Don t t use likelihood and impact analysis for risk assessment. Don t t define high risk to be high likelihood and high impact. Don t t confuse likelihood and frequency. Recognize that historical loss data is the starting point for risk assessment and control assessment. Recognize that ORM is a process that leads to more educated decision making, which can increase profitability, not a compliance exercise. 60

61 Why are banks having such a difficult time establishing ORM programs under Basel II? Many banks have developed ORM frameworks based on a blend of traditional and modern ORM. Traditional ORM and Modern ORM are based on entirely different definitions, approaches, processes and methodologies. These methodologies are highly subjective, resource intensive and generate a huge catalogue of unmanageable risks. risks. Most businesses do not see value in these ORM initiatives; As a result Basel II is often seen as a meaningless compliance exercise. In many cases a minor realignment in framework can result in significant benefits. 61

62 ECONOMIC AND REGULATORY CAPITAL Bank Indonesia Bali 21 September 2006 Presented by David Lawrence