This action might not be possible to undo. Are you sure you want to continue?
CreditMetrics was introduced in 1997 by JPMorgan and its co-sponsors (Bank of America, Union Bank of Switzerland, et al.) as a value at risk (VaR) framework to apply to the valuation and risk of non-tradable assets such as loans and privately placed bonds.2 Thus, while RiskMetrics seeks to answer the question, if tomorrow is a bad day, how much will I lose on tradable assets such as stocks, bonds, and equities? CreditMetrics asks, if next year is a bad year, how much will I lose on my loans and loan portfolio?3 With RiskMetrics (see Chapter 10) we answer this question by looking at the market value or price of an asset and the volatility of that asset’s price or return in order to calculate a probability (e.g., 5 percent) that the value of that asset will fall below some given value tomorrow. In the case of RiskMetrics, this involves multiplying the estimated standard deviation of returns on that asset by 1.65 and then revaluing the current market value of the position (P) downward by 1.65 . That is, daily value at risk (daily VaR at a 95 percent confidence level) is VaR 5 P 3 1.65 3 Unfortunately, since loans are not publicly traded, we observe neither P (the loan’s market value) nor (the volatility of loan value over the horizon of interest—assumed to be one year for loans and bonds under CreditMetrics). However, using (1) available data on a borrower’s credit rating, (2) the probability of that rating changing over the next year (the rating transition matrix), (3) recovery rates on defaulted loans, and (4) yield spreads in the bond market, it is possible to calculate a hypothetical P and for any non-traded loan or bond and thus a VaR figure for individual loans and the loan portfolio. Consider the example of a five-year, fixed-rate loan of $100 million made at 6 percent annual interest.4 The borrower is rated BBB.
On the basis of historical data collected by S&P, Moody’s, and other bond analysts, it is estimated that the probability of a BBB borrower’s staying at BBB over the next year is 86.93 percent. There is also some probability that the borrower of the loan will be upgraded (e.g., to A), and there is some probability that it will be downgraded (e.g., to CCC) or even default. Indeed, there are eight possible transitions the borrower can make over the next year, seven of which involve upgrades, downgrades, and no rating changes and one that involves default. The estimated probabilities are shown in Table 12A–1.
The effect of rating upgrades and downgrades is to impact the required credit risk spreads or premiums on loans and thus the implied market value (or present value) of the loan. If a loan is TABLE 12A–1 One-Year Transition Probabilities
for BBB-Rated Borrower Rating AAA AA A BBB BB B CCC Default Transition Probability 0.02% 0.33 5.95 86.93 5.30 1.17 0.12 0.18
Most likely to stay in same class
This appendix, which contains more technical topics, may be included in or dropped from the chapter reading depending on the rigour of the course.
See JPMorgan, CreditMetrics—Technical Document, New York, April 2, 1997; and A. Saunders and L. Allen, Credit Risk Measurement: New Approaches to Value at Risk and Other Paradigms, 2nd ed. New York: John Wiley & Sons, 2002, Chapter 6.
3 In 2002, JPMorgan introduced a third measure of credit risk, CreditGrades. The CreditGrades model establishes a framework linking the credit and equity markets. The model employs approximations for the asset value, volatility, and drift, which are used to value credit as an exotic equity derivative. This model is similar in approach to the KMV model described in the chapter. See CreditGrades: Technical Documents, RiskMetrics Group, Inc., May 2002. 4
This example is based on the one used in JPMorgan, CreditMetrics—Technical Document, April 2, 1997.
05 4.0372 2 11. they reflect forward rates from the current Treasury yield curve—see discussion in Chapter 11—and si are annual credit spreads for loans of a particular rating class of one year. if the loan borrower is upgraded from BBB to A.64 51. Table 12A–2 shows the hypothetical values of rt and st over the four years. then (measured in millions of dollars): P561 1 6 6 1 11 1 r1 1 s1 2 11 1 r2 1 s2 2 2 6 106 1 11 1 r3 1 s3 2 3 11 1 r4 1 s4 2 4 That is.75 0.e.13 107.55 102.40 st 0. three years.0493 2 Value of loan 51. Then the present value or market value of the loan to the FI at the end of the oneyear risk horizon (in millions of dollars) is: P561 6 6 6 106 1 1 1 2 3 11. (This is the value the FI would theoretically be able to obtain if it “sold” the loan. The minimum value is the estimated recovery value of the loan if the borrower declares bankruptcy. Note that the loan has a maximum market value of $109.00% 3. to another FI at the end of year 1 horizon at the fair market price or value. two years. and so on.. 6 percent loan at the end of the first year after a credit event has occurred during that year. As can be seen.0532 2 4 11.12A-2 Appendix 12A CreditMetrics downgraded.37 109. Suppose the borrower is upgraded during the first year from BBB to A. two years.) Table 12A–3 shows the value of the loan if other credit events occur.13 where the ri are the risk-free rates on T-bonds expected to exist one year. the $100 million (book value) loan has a market value to the FI of $108. with the accrued first year coupon of 6. into the future (i.57 4.66 107. at the end of the first year. TABLE 12A–2 Risk-Free Rates on T-Bonds and Annual Credit Spreads Year 1 2 3 4 rt 3. TABLE 12A–3 Value of the Loan at the End of One Year under Different Ratings Year-End Rating AAA AA A BBB BB B CCC Default Loan Value ($) (including first-year coupon) 109.37 = Mean 5 $108. the loan rate in our example is fixed at 6 percent) so that the present value of the loan to the FI should fall.02 98.66 million. the value FIGURE 12A–1 Distribution of Loan Values on a Five-Year BBB Loan at the End of Year 1 Probability The first coupon or interest payment of $6 million in the above example is undiscounted and can be viewed as being similar to the accrued interest earned on a bond or a loan since we are revaluing the loan at the end (not the beginning) of the first year of its life.10 83. the reverse is true for a credit rating upgrade. Technically.37 (if the borrower is upgraded to AAA) and a minimum value of $51.88 0.66 . and four years to maturity (the latter are derived from observed spreads in the corporate bond market over Treasuries).09 109.72% 0.13 if the borrower defaults.92 The probability distribution of loan values is shown in Figure 12A–1. since we are revaluing the fiveyear $100 million. the required credit spread premium should rise (remember.19 108.0432 2 11.
99 Assuming Normal 5% VaR 5 1.99..97 million.09 2 $102.99 *5% VaR approximated by 6.12 0.Appendix 12A CreditMetrics 12A-3 of the loan has a fixed upside and a long downside (i. then.3592 0. which is the sum of each possible loan value at the end of year 1 times its transition probability.07 million. it demonstrates a negative skew or a long-tail downside risk.3 percentile equals 83.94777 Year-End Rating AAA AA A BBB BB B CCC Default = Standard deviation = $2.93 5.9477 (squared) and its standard deviation or volatility is the square root of the variance equal to $2.47% VaR (i. as shown in Figure 12A–1.36 6.77% VaR (i.18%) and 1% VaR approximated by 1.96) Probability Weighted Difference Squared 0. In particular. implying an approximate 5 percent actual VaR of over $107.10 and the 0.1853 1.10 1.15 1.17 0.18%). Thus the 5 percent VaR for the loan is 1. Based on the normal distribution of loan values.46 (5.02 0. These actual VaRs could be made less approximate by using linear interpolation to get the exact 5 percent and 1 percent VaR measures. and that there is a 1.18 New Loan Value plus Coupon ($) $109. at year 1.07 b Distribution* 1% VaR 5 99% of actual distribution 5 $107.28 2.e.41 1.6598 5.02 = $5. the FI is concerned about losses or volatility in value.9446 0.0010 0.02% 0. Assuming that loan values are normally distributed.33 $2.99.6358 Variance = 8.09 (also see Figure 12A–1). Using the actual distribution of loan values and probabilities.e.93 million.09 $2.. .37 109. The first step in calculating VaR is to calculate the mean of the loan’s value.12% + 0.66 107.02 5 $5. Based on the actual distribution of loan values.09 2 $98.09 Probability Difference of Weighted Value Value from Mean ($) ($) $ 0. 2. if next year is a bad year. or its expected value. In particular.77 percent probability that the loan value will fall below $102. the mean value of the loan is $107.0146 0. a negative skew). the distribution of the loan’s value is clearly nonnormal.10 5 $8.e.47 percent probability that the loan value will fall below $98.57 0.. For example. As can be seen.02 98.13 Mean = $107.45) (55.06) (8.64.65 3 b Distribution 1% VaR 5 2.10. 5.47 percentile equals 98.09 − $98.99 = $6.64 51. However.97 Assuming Actual 5% VaR 5 95% of actual distribution 5 $107. CALCULATION OF VaR Table 12A–4 shows the calculation of the VaR based on each approach for both the 5 percent worst-case and the 1 percent worst-case scenarios. the variance of loan value around its mean is $8. Thus CreditMetrics produces two VaR measures: 1. this is likely to underestimate the actual or true VaR of the loan because.3% + 1.09 − $102.19 108.47 93. However.49 5.30 1. we can see from Table 12A–4 that there is a 6.99) (23. using linear interpolation.95 86. implying an approximate 1 percent actual VaR of over $107.1474 0.17% + 0. It is clear that the value of the loan is not symmetrically (or normally) distributed.93 5 $6. while the 1 percent VaR is 2.10 0.12% + 0.10 83.17% + 0. 1. since the 1.65 $2.02. the TABLE 12A–4 VaR Calculations for the BBB Loan Probability of State (%) 0.33 3 5 $4.55 102.33 5.10 = $8. how much can it expect to lose? We could define a bad year as occurring once every 20 years (the 5 percent VaR) or once every 100 years (the 1 percent VaR)—this is similar to market risk VaR except that for credit risk the horizon is longer: 1 year rather than 1 day.99 = $4.
00 st% 0. a.60 114.82 114.75 4.30 .60 0. From Table 12A–1.43 86.12 Forward Rate Spreads at Time t t 1 2 3 4 rt% 3. As discussed in Chapter 20.96 1. what is the probability of a loan upgrade? A loan downgrade? a. In our example of a $100 million face (book) value BBB loan.00 percentile equals $92. Questions and Problems 1.03 108.97 million would be required (i. How do the capital requirements of the CreditMetrics approach differ from those of OSFI? 2. depending on the approach adopted under Basel II.31 1.20 New Loan Value plus Coupon $ $114.48 5. What is the volatility of the loan value at the end of year 1? d. less than OSFI’s requirement).. How is the discount rate determined after a credit event has occurred? c. assuming a normal distribution of values.16 1. the probability distribution given in the table below has been determined for various ratings upgrades. Based on hypothetical historical data.12A-4 Appendix 12A CreditMetrics 1. while using the 1 percent VaR based on the iterated value from the actual distribution. and default possibilities over the next year. a $14.e. What is the mean (expected) value of the loan at the end of year 1? c. Using the 1 percent VaR based on the normal distribution.09 − $92.55 98. Calculate the 5 percent and 1 percent VaRs for this loan.80 million capital requirement would be required (which is much greater than OSFI’s capital requirement). Information also is presented reflecting the forward rates of the current Treasury yield curve and the annual credit spreads of the various maturities of BBB bonds over Treasuries.00% 3. downgrades. A five-year fixed-rate loan of $100 million carries a 7 percent annual interest rate. It should be noted that under the CreditMetrics approach. the capital requirement would be $8 million. This contrasts to the two market-based VaR measures developed above.72% 0. What is the present value of the loan at the end of the one-year risk horizon for the case where the borrower has been upgraded from BB to BBB? b. Estimate the approximate 5 percent and 1 percent VaRs using the actual distribution of loan values and probabilities. each loan held by a Canadian bank may carry an individual capital requirement based on its credit rating. CAPITAL REQUIREMENTS It is interesting to compare these VaR figures with the capital reserves against loans currently required by OSFI.80. How do the capital requirements of the 1 percent VaRs calculated in parts (d) and (e) above compare with the capital requirements of OSFI? Rating AAA AA A BBB BB B CCC Default Probability Distribution 0. Why does the probability distribution of possible loan values have a negative skew? d.82 54.45 6.29 = $14.40 3. The borrower is rated BB. status quo.01% 0. f. a capital requirement of $6.05 85. every loan is likely to have a different VaR and thus a different implied capital requirement.90 0.29. This suggests an actual 1 percent VaR of $107. e. What is the impact of a rating upgrade or downgrade? b.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue listening from where you left off, or restart the preview.