RISK MANAGEMENT IS IT NECESSARY?

Transcription

1 RISK MANAGEMENT IS IT NECESSARY? Credit Risk Management - Fundamentals, Practical Challenges & Methodologies

2 While financial institutions have faced difficulties over the years for a multitude of reasons, the major cause of serious banking problems continues to be directly related to lax credit standards for borrowers and counterparties, poor portfolio risk management, or a lack of attention to changes in economic or other circumstances that can lead to a deterioration in the credit standing of a bank s counterparties. This experience is common in both G-10 and non- G-10 countries. Basel Committee on Banking Supervision

3 CREDIT GROWTH HH credit to GDP, and GDP per capita, 2005 Estimated Growth Rates, Source: Global Financial Stability Report, 2006, IMF Credit penetration seems to be linked to GDP per capita

4 RISKS TO BANKS Large capital flows can create asset bubbles Large unbanked sections with limited information Herd behaviour in banks Borrowers more exposed to macro and micro factors Mis-pricing risks for new customer segments Differentiation of risk a challenge due to lack of granular information

5 WHAT IS CREDIT RISK? Risk of loss due to default by a borrower / counter-party in meeting its obligations Reduction in portfolio value arising from actual / perceived deterioration in credit quality of borrower / counter-party Default could be due to inability, incapacity, unwillingness to honour commitments Commitments typically can be in relation to lending, trade financing, trading, hedging, settlement, commitments, guarantees, other financial transactions Made up of concentration risk and intrinsic risk from individual exposure Dependent on factors internal and external to a bank In respect of an individual, corporate, bank, financial institution or sovereign

6 OBJECTIVES OF RISK MANAGEMENT Contain volatilities in cash flow expected when assuming a risk position Contain volatilities in equity value of an asset Contain the above volatilities within the risk appetite of a bank Risk appetite typically related to capital available with the bank Managed at portfolio level as well as individual exposure level Risk should be assumed when there is compensating returns There should be economic value addition to a bank Ensure long term sustenance of a bank

7 HOW IS CREDIT RISK MANAGED? Governance and Effective Risk Architecture Risk Appetite Statement, Strategy and Policies Risk Identification Risk Assessment and Measurement Risk Monitoring Risk Control Risk Mitigation Risk-Return Trade Off

8 ARCHITECTURE Board of Directors Risk Management Committee (Board Level) Risk Management Group

9 TYPICAL FUNCTIONS OF RISK MANAGEMENT Identify, measure, monitor and control credit risk Formulate credit policies / processes / strategy and enforce implementation Design and validate risk measurement systems, capital allocation models and pricing frameworks Monitor & protect portfolio quality Independently assess individual risk exposures assumed by business units Research on economy, industries, sectors and carry out impact analysis of major policy changes / events

10 POLICIES Desired Balance sheet quality and composition defined based on risk appetite Target market identified Acceptable underwriting standards / screening parameters laid down Prudential exposure ceilings and portfolio concentration levels set Products and quantum of exposure dispensable risk related specified Delinquency recognition and provisioning norms stipulated

11 Management RISK IDENTIFICATION Ownership and Board Composition Track Record Competency Industry & Business Financials Demand-Supply Gap Government Policy Competition & Market Position Input availability & Operating Capacities Gearing Liquidity Profitability Operating Efficiency

12 RISK MEASUREMENT Rating / Risk Scoring Models Single point indicator of obligor s creditworthiness Risk Components Probability of Default Loss Given Default Exposure At Default Portfolio Risk Measurement / Capital computation Default Risk Value at Risk Pricing Models RAROC framework

13 RISK MONITORING Transaction analysis Customer calls Plant visits Quarterly financial review Market / industry developments Tracking external ratings Annual review Analysis of Portfolio Distribution

14 RISK CONTROL Approvals for risk positions given under multiple signatures or committee, with specified dispensation powers Ceiling on high risk positions, group/industry/geographic concentration Exposure limit set based on degree of risk Independence of risk approvals from business origination Independence of operations/administration from business origination Prompt remedial action based on early warning signals Periodic reviews/audit

15 CREDIT RISK MITIGATION Eligible Financial Collateral (subject to haircut) Cash Gold Securities issued by Central/State Govt. KVP, NSC, Life insurance policy Rated debt securities which would attract a risk weight of atleast 100% Unrated debt securities issued by a bank, listed on a recognised stock exchange, classified as a senior debt and all rated issues by the issuing bank are at least BBB(-) or PR3/P3/F3/A3 Units of Debt Mutual Funds Guarantees issued by sovereign, sovereign entities, banks other entities rated AA (-) or better

16 RAROC FRAMEWORK Risk Adjusted Return On Capital = Earnings less Cost of Funds less Overheads less Risk Premium Divided by Economic Capital Risk Premium = Expected Loss Economic Capital = multiple of Unexpected Loss based on desired confidence level

17 EVOLUTION OF RISK MANAGMENT TOOLS Stage I : Notional Amounts Risk measured by total notional amount Capital established using a multiplier (~ 8%) to this amount Ignores variations in probability of default Stage II : Risk Weighted Amounts Basel 1988 Rough categorization of credit risk by class; Risk weights to scale each notional amount Created incentives to alter portfolio to maximize returns adverse selection Stage III : External / Internal Credit Ratings Basel New Capital Accord Better representation of credit risk; More in line with economic measures Credit measured on a stand-alone basis Stage IV : Portfolio Credit Risk Management Models Benefits of diversification

18 DRIVERS OF CREDIT RISK Probability of Default (PD) Discrete state default / no default What is the probability that a counter-party would default over a defined horizon? Exposure At Default (EAD) Credit Exposure Economic value of claim at time of default What amount / percentage of the committed line would be outstanding at the time the counter-party defaults? Loss Given Default (LGD) Fractional loss due to default / economic value of loss What percentage of the exposure at the time of default will the bank end up losing forever?

19 CREDIT LOSS DISTRIBUTION Expected Loss PD x LGD x EAD Risk Premium Provision / Earning Spread Unexpected Loss Variance about Expected Loss Derive Credit Value at Risk (VaR) at some confidence level Economic Capital Frequency of loss Expected Loss Unexpected Loss Stress Loss 0 Amount of loss

20

21 DEFAULT RISK MODELS Default Risk Primary component of credit risk Represents the probability of default (PD) as well as the loss given default (LGD) Can be measured by two approaches Actuarial Methods Objective measures of default rates Usually based on historical data Describes actual or natural probability of default Market-Price Methods Infers the market s assessment of default risk From traded prices of debt, equities, credit derivatives Incorporates all the news about an entity s prospects Contaminated by effect of risk premium and hence do not exactly measure default probabilities

22 Credit Event A discrete state Bankruptcy; failure to pay; cross default; cross acceleration; restructuring Credit Rating / Score ACTUARIAL DEFAULT RISK MODELS Evaluation of creditworthiness / likelihood to default opinion of the future ability, legal obligation and willingness of a bond issuer or other obligor to make full and timely payments on principal and interest due to investors Moody s Typically a number of criteria used, select accounting ratios, demographics, account behaviour Statistical tools used to test relevance and correlation to select model parameters and relationship Population needs to be reasonably homogeneous and of statistically meaningful size Population should be representative of real life market Model to be built on sample characteristically identical to population Should have distinct dimensions obligor rating and facility rating

23 ACTUARIAL DEFAULT RISK MODELS (contd) Historical Default Rates Ratings migration of individual credits tracked Transition matrix built for given horizon Probability of migration across ratings, including to default grade for each rating Omitting or adding a few borrowers could skew default rates if data points are scarce Paucity of default data points could pose problem. Estimation of default rates for low probability events can be very imprecise Defaults are correlated with economic activity Movement of entity into and out of the portfolio to be accounted for Treatment of withdrawn / exited entities to be formalized Periodic and prompt review of rating essential, particularly flagging a default Transition matrix to be refreshed periodically

24 RATING TRANSITION

25 ACTUARIAL DEFAULT RISK MODELS (contd) Recovery Rates Fraction recovered after default ; 1 recovery rate yields LGD Credit event will decide when default exposure frozen and recovery clock starts ticking Function of status and seniority of the debt, state of the economy and industry, estimated cost of recovery, timing of recovery, means of payment Recovery rates, as per international studies, vary widely. Recovery rates are observed to be negatively related to default rates This correlation creates bigger losses, which extends the tail of the credit loss distribution

26 EXPOSURE AT DEFAULT & LOSS GIVEN DEFAULT Bank s internal Historical Default & Loss Data (For each defaulted capture exposure at time of default, recoveries made subsequently (discounted), expenses incurred on recovery, industry type, credit product,etc) Exposure at time of default divided by aggregate commitment Recoveries made discounted to date of default reduced by expenses on recovery, etc Exposure At Default % Loss Given Default %, i.e. (100 PV of adjusted recovery%)

27 Market Indicator Market prices of securities are affected by default / expectation of credit weakening Expected to be more up-to-date Presumed to be accurate measure of credit risk because financial markets have access to large amount of information Corporate Bond Prices MARKET PRICE DEFAULT RISK MODELS Yield on a corporate bond is broken down into a default probability, a recovery rate and a risk-free yield P* = P / (1 + y*) = [ P / (1 + y ) ] x (1 - p) + [ f x P / (1 + y)] x p simplifies to y* = y + p (1 f) P : bond value at maturity P* : initial price of bond y : risk free yield y* : market determined yield f : fractional recovery

28 MARKET PRICE DEFAULT RISK MODELS (contd) Corporate Bond Prices Issues Requirement of well developed bond market Reference bond may not be actively traded Counterparty may not have outstanding publicly traded bond or bond may have a Equity Prices call feature Advantage more widely available, considered of better quality than bond prices Equity is viewed as a call option on the value of the firm, with strike price equal to the value of the debt Merton Model (1974) Default probability inferred from value of this option (using Black Scholes) B = Min ( V, K) ECL = N(-d 2 ) [K V e N(-d 1 ) / N (-d 2 )] = p x [exposure x LGD] B : Derived value of the bond / loan V : Total value of the firm K : Face value of the bond / loan N(d) : cumulative distribution function for standard normal distribution

29 MARKET PRICE DEFAULT RISK MODELS (contd) Equity Prices - Issues Validation in domestic environment needed Markets to be efficient and mature Cannot be used to price sovereign credit risk Fundamental drawback is that it relies on static model of firm s capital and risk structure debt level assumed to be constant over the horizon Debt could mature at various points in time Management could undertake new projects, which could increase equity value as well as volatility Fails to explain magnitude of credit spreads Hence, found most useful in tracking changes in Estimated Default Frequencies (EDFs) over time

30 PORTFOLIO CREDIT RISK MODELS Background Portfolio theory pushed for viewing risk in the context of contribution to the total risk and not in isolation Pricing of risk is markedly lower when considering portfolio effects Technological advances has helped aggregate risk close to real time Models too have benefited from technology Exponential growth in new products securitizations, credit derivatives, which allow better management of credit risk Developments in regulatory and financial markets are laying greater emphasis on credit risk

31 PCRM - APPROACHES Top-Down vs Bottom-Up Models Top-down models group credit risks using single statistics. Aggregate many sources of risk viewed as homogeneous into an overall portfolio risk appropriate for retail Bottom-up models account for features of each asset. Most useful to take corrective action, because risk structure can be reverse-engineered to modify risk profile Default Mode vs Mark-to-Market Models Default mode models consider only outright default as a credit event MTM models consider changes in market values and ratings changes including default Conditional vs Unconditional Models of Default Probability Conditional models incorporate changing macroeconomic factors into default probability Unconditional models have fixed default probabilities; focus on borrower information Structural vs Reduced Form Models of Default Correlations Structural models explain correlations by joint movement of assets Reduced form assume functional relationship between default and background factors

32 STRESS TESTING Definition Looking at the effect on the portfolio of large predefined moves in a single variable Addresses the question How much could the bank lose based on a major change in a particular dimension Deals with outlier events large moves that lie beyond day-to-day risk monitoring Takes a good hard look at variables that drive the value of a portfolio (vs scenario analysis which typically starts with hypothesis about an alternative state of the world) Steps Picking what to stress (choice of variable, range of stress, usefulness of stress information vs data overload) Identifying assumptions (will correlations hold, does underlying financial model hold good) Revaluing the portfolio Deciding on action steps (reporting, cross checks on model, action plan for dealing with actual catastrophe situation

33 BACK TESTING Definition Models are designed to reflect reality. Backtests compare realized results with model generated risk measures Evaluate a new model and reassess accuracy and consistency of existing models Steps Record realized results with as comprehensive a detailing possible Compare with model estimates Interpret back test results (variance, magnitude of excess) Investigate and question cause for variance Carry out corrective action on model (parameters, methodology, assumptions, variables, data type, data period, etc) Validate model

34 BASEL II APPROACHES Menu of approaches for computing capital Moves along sophistication of a bank s risk management systems Each approach demands increased compliance with minimum requirements Credit Standardized Foundation Internal Ratings Based Advanced Internal Ratings Based Simple Simplicity Sophisticated Less reflective of actual risk Risk Sensitivity More reflective of actual risk Low level of detail Data & Documentation High level of detail

35 STANDARDISED APPROACH - PRESCRIPTION Use of external credit ratings or asset class type to determine risk weights More risk differentiation, thus addressing the main concern from Basel I More recognition of risk mitigation techniques Relatively easier for banks with average risk management systems

36 INTERNAL RATINGS BASED APPROACH Rely on own internal estimates of Risk Components Meet minimum conditions & disclosure requirements Obtain Supervisory approval Capital Requirement as per Internal Ratings Based Approach Probability of Default (PD) Loss Given Default (LGD) Exposure At Default (EAD) Effective Maturity (M)

37 FOUNDATION & ADVANCED APPROACHES Corporate Sovereign Foundation Bank s own PD estimate Supervisory EAD, LGD, M Advanced Bank s own PD, EAD, LGD, M estimates Bank Retail Bank s own PD, EAD, LGD estimate No distinction between foundation & advanced Equity Market-based approach PD/LGD approach

38 RISK WEIGHT FUNCTION Correlation (R) = 0.12 (1 EXP(-50 PD)) / (1 EXP(-50)) [1 - (1 - EXP(- 50 PD))/(1 - EXP(-50))] Maturity Adjustment (b) = ( ln(pd))^2 Capital Requirement (K) = [LGD N[(1-R)^-0.5 G(PD)+(R/(1- R))^0.5 G(0.999)] PDxLGD] x (1-1.5xb)^-1 (1+(M-2.5) b) Risk-Weighted Assets (RWA) = K x 12.5 x EAD Ln denotes natural logarithm N (x) denotes cumulative distribution function for standard normal random variable (i.e. probability that a normal random variable with mean 0 and variance of = / < 1 or x) G (z) denotes inverse cumulative distribution function for standard normal random variable (i.e. value of x such that N(x) = z)

39 If everything seems to be going well, you have obviously overlooked something. Thank You