FIN 683 Financial Institutions Management Credit Risk

Transcription

1 FIN 683 Financial Institutions Management Credit Risk Professor Robert B.H. Hauswald Kogod School of Business, AU Individual-Loan Characteristics Analysis and measurement of credit risk: specific loans Pricing loans and bonds Setting limits on credit risk exposure Credit Quality Problems 1980s: LDC loans, and residential and farm mortgage loans Late 1990s, credit card and auto loans debt crises in Argentina, Brazil, Russia, and South Korea : Mortgage delinquencies and subprime loans Emphasizes importance of managing credit risk 2

2 Credit Quality Problems Over the early to mid 1990s, improvements in NPLs for large banks and overall credit quality starting from a very low base after the S&L crisis Late 1990s and early 2000s: Telecom companies, tech companies Mid 2000s, economic growth accompanied by reduction in NPLs artificial prosperity Mortgage crisis: Countrywide, WaMu since then: increased emphasis on credit risk evaluation 3 Nonperforming Asset Ratio for U.S. Commercial Banks 4

3 Types of Loans C&I loans: Secured and unsecured Syndication Spot loans, loan commitments Decline in C&I loans originated by commercial banks and growth in commercial paper market: disintermediation Effect of financial crisis on commercial paper market RE loans: Primarily mortgages Fixed-rate, ARM Mortgages can be subject to default risk when loan-tovalue declines Consumer loans: competition with credit cards 5 ARMs: Share of Total Loans Closed,

4 Consumer Loans Individual (consumer) loans: personal, auto, credit card Nonrevolving loans Automobile, mobile home, personal loans Growth in credit card debt Visa, MasterCard vs. proprietary cards such as Sears, AT&T Consolidation among credit card issuers Bank of America & MBNA Risks affected by competitive conditions, predatorylending and usury legislation (interest rate ceilings) Bankruptcy Reform Act of Annual Net Charge-Off Rates on Loans 8

5 Other Loans Farm loans Lending to other banks: interbank market: role only recently appreciated repo market: collateralized lending Nonbank FIs Broker margin loans Foreign banks and sovereign governments State and local governments 9 Return on a Loan Factors: interest rate, fees, credit risk premium, collateral, and other requirements such as compensating balances and reserve requirements Return = inflow/outflow 1+k = 1+(of + (BR + m ))/(1-[b(1-RR)]) Expected return: 1 + E(r) = p(1+k) where p equals probability of repayment missing term: (1 p)k where k is default return Note that realized and expected return may not be equal 10

6 Access to Credit At retail Usually a simple accept/reject decision rather than adjustments to the rate Credit rationing If accepted, customers sorted by loan quantity For mortgages, discrimination via loan to value rather than adjusting rates At commercial creditworthy: Yes or No? simple accept/reject terms (what are they) and pricing adjustments 2/16/ Individual Credit Risk Robert B.H. Hauswald Measuring Credit Risk Availability, quality, and cost of information are critical factors in credit risk assessment Facilitated by technology and information Qualitative models Borrower specific factors are considered as well as market or systematic factors Specific factors include reputation, leverage, volatility of earnings, covenants, and collateral Market specific factors include business cycle and interest rate levels 12

7 Credit Scoring Models Linear probability models: n PD i = β j X i, j + error j= 1 Statistically unsound since the Z s obtained are not probabilities at all There is little justification for employing linear probability models superior statistical techniques are readily available 13 Other Credit Scoring Models Logit models Overcome weakness of the linear probability models use a simple transformation (logistic function) that restricts the probabilities to the zero-one interval Other alternatives include Probit and other variants with nonlinear indicator functions Quality of credit scoring models in predicting default has improved to a degree that better than traditional credit analysis large banks can enter local credit markets: SBL 2/16/ Individual Credit Risk Robert B.H. Hauswald

8 Altman s Linear Discriminant Model Z=1.2X X X X X 5 Critical value of Z = 1.81 X 1 = Working capital/total assets X 2 = Retained earnings/total assets X 3 = EBIT/total assets X 4 = Market value equity/ book value LT debt X 5 = Sales/total assets Totally outdated but still around variant of linear probability model 15 Altman s Z Score Only considers two extreme cases (default/no default) Weights need not be stationary over time Ignores hard to quantify factors including business cycle effects Database of defaulted loans is not available to benchmark the model 16

9 Term Structure Based Methods Infer the default probability from the risk premium expected return equals risk free rate after accounting for probability of default: p (1+ k) = 1+ I risk premium can be computed using Treasury strips and zero coupon corporate bonds May be generalized to loans with any maturity or to adjust for varying default recovery rates The loan can be assessed using the inferred probabilities from comparable quality bonds 17 Mortality Rate Models Similar to the process employed by insurance companies to price policies the probability of default is estimated from past data Marginal Mortality Rates: MMR 1 = (Value Grade B default in year 1) (Value Grade B outstanding yr.1) MMR 2 = (Value Grade B default in year 2) (Value Grade B outstanding yr.2) Similar problems associated with credit scoring sensitivity to the period chosen to calculate the MMRs 18

10 RAROC Models Risk adjusted return on capital One of the most widely used models RAROC = (one year net income on loan)/(loan risk) Loan risk estimated from loan default rates, or using duration are the two risks interchangeable? related? 19 Using Duration to Estimate Loan Risk For denominator of RAROC, duration approach used to estimate worst case loss in value of the loan: DLN /LN = -DLN x (DR/(1+R)) where DR is an estimate of the worst change in credit risk premiums for the loan class over the past year RAROC = one-year income on loan/dln 20

11 Option Models Employ option pricing methods to evaluate the option to default equity in the presence of risky debt and limited liability resembles what? Used by many large banks to monitor credit risk: extract default prob. from share prices KMV Corporation markets this model quite widely: ingenuous use of data Problem? 21 Applying Option Valuation Model Merton showed value of a risky loan: F(τ) = Be -iτ [(1/d)N(h 1 ) +N(h 2 )] Expressed as a yield spread: k(τ) - i = (-1/τ)ln[N(h 2 ) +(1/d)N(h 1 )] where k(τ) = Required yield on risky debt ln = Natural logarithm i = Risk-free rate on debt of equivalent τ = remaining time to maturity 22

12 Summary Credit risk: managed individually better safe than sorry before: screen and assess repayment prospects after: monitor Statistics revolutionized loan approval credit scoring: rescued profitability of whole segments of the credit market does not work for large firms a model is only as good : subprime crisis 23 Sovereign Risk Debt repudiation Since WWII, only China, Cuba, and North Korea have repudiated debt Recent steps to forgive debts of most severe cases conditional on reforms Rescheduling and restructuring Most common form of sovereign risk: EM markets, peripheral EU (Greece) South Korea, 1998 Argentina, 2001

13 Credit Risk vs. Sovereign Risk Governments can impose restrictions on debt repayments to outside creditors Loan may default even though borrower had a strong credit rating at origination of loan Legal remedies are very limited: sovereignty Emphasizes the need to assess credit quality and sovereign risk The Past Late 1980s and early 1990s: Expanding investments in emerging markets Peso devaluation and subsequent restructuring More recently: Asian and Russian crises Turkey and Argentina Argentina s focus on fiscal surplus Economic growth in the 2000s and reduction in external debt The Graduates: BR(I)CS

14 The Present Late 2000s, economies faltered Developed countries faced some of the worst declines in GDP ever experienced Spain s unemployment rate at 17.3 % World Bank and IMF pledged $250 billion Crisis in Greece spread to Portugal, Spain, and Italy Multiyear restructuring agreements (MYRAs) Large experiment in sovereign debt restructuring underway in the EU: wasting tax payers money Debt Rescheduling More likely with international loan financing rather than bond financing Loan syndicates often comprised of same group of FIs versus large numbers of bondholders facilitates rescheduling Cross-default provisions Specialness of banks argues for rescheduling but creates incentives to default again if bailouts are automatic: Greece vs the Eurozone

15 Country Risk Evaluation Outside evaluation models: The Euromoney Index The Economist Intelligence Unit ratings Highest risk in countries such as Iraq, Zimbabwe, and Sudan. Institutional Investor Index 2009 placed Switzerland at least chance of default and Zimbabwe at highest U.S. not the lowest risk: why? Country Risk Evaluation Internal Evaluation Models Statistical models Country risk-scoring models based on primarily economic ratios The selected variables are tested for predictive power in separating rescheduling countries from non-rescheduling countries using past data Rating models: qualitative vs. quantitative

16 Statistical Models Commonly used economic ratios: Debt service ratio: (Interest + amortization on debt)/exports Import ratio: Total imports / Total FX reserves Investment ratio: Real investment / GNP Variance of export revenue = σ 2 ER Domestic money supply growth = M/M Discriminant function: p=f(dsr,ir, INVR, ) probability of default estimation Problems with Statistical CRA Models Measurements of key variables Population groups Finer distinction than reschedulers and nonreschedulers may be required Political risk factors may not be captured Strikes, corruption, elections, revolution Corruption Perceptions Index 15-32

17 Other Shortcomings Portfolio aspects Many large FIs with LDC exposures diversify across countries Diversification of risks not necessarily captured in CRA models Rarely address incentive aspects of rescheduling Borrowers and Lenders: Benefits vs. Costs Stability: Model likely to require frequent updating Using Market Data to Measure Risk Secondary market for LDC debt Sellers and buyers Market segments Sovereign bonds Performing LDC loans Nonperforming LDC loans Small market for brady bonds 15-34

18 Managing Sovereign Risk Exposure Debt-equity swaps Citibank sells $100 million Chilean loan to Merrill Lynch for $91 million Merrill Lynch (market maker) sells to IBM at $93 million Chilean government allows IBM to convert the $100 million face value loan into pesos at a discounted rate to finance investments in Chile Throw good money after bad: promise new loans for reform and repayment of old ones MYRAs Aspects of MYRAs: Fee charged by bank for restructuring Interest rate charged Grace period Maturity of loan Option features Concessionality 15-36

19 Other Mechanisms Loan sales Bond for loan swaps (brady bonds) Transform LDC loan into marketable liquid instrument Usually senior to remaining loans of that country Pertinent Websites Bank for International Settlements Heritage Foundation Institutional Investor International Monetary Fund The Economist Transparency International World Bank