What would be a suitable metric to assess credit risk on a relative basis when making a relative value decision?

What will be an ideal response?

A suitable metric to assess credit risk on a relative basis when making a relative value decision is a return-risk metric. Default probabilities are used along with credit spreads in this metric which measures the ratio of the credit spread (the return) relative to the default probability (therisk). The metric of interest is the ratio of the credit spread to the maturity-matched default probability.The investor can then select the issue that offers the highest credit spread per basispoint of default risk for each maturity. More details are given below.

The focus of the early models was on generating forecasts of default rates, credit ratings, and credit spreads (measured relative to U.S. Treasury securities). From a portfolio perspective, the assumption was that credit default risk for corporate bonds is idiosyncratic, and therefore can be diversified away in large portfolios containing corporate bonds. In credit risk modeling for individual corporate bonds, some risk-return measure (such as the Sharpe ratio) was used to evaluate the observed credit spreads. Since the mid-1990s, more sophisticated approaches to credit risk modeling have been proposed and made commercially available to portfolio managers. For one of these approaches that is commercially available, the theoretical foundation underlying the model dates back to the early 1970s.

To estimate the fair marketcredit spread, a credit risk modelrequires (1) a model that estimates recovery if a default occurs, (2) a model that shows thecredit spread that investors want in order to acceptsystematic credit risk and idiosyncraticrisk—as the spread is largely driven by uncertainty of the timingof default for this onecompany,and (3) a model of the risk-free rate.

The Kamakura model is not limited tocredit risk but also integrates interest-rate risk.The model calculates the probability of default for a corporate bond issuer using eitherthe structural approach based on the Merton model or the reduced-form approach. Thechoice of the model to use in an application is up to the user. The Kamakura analysis estimated BAC's annualized default probabilities. Default probabilities are used along with credit spreads to obtain a return-risk metric. Thatmeasure is the ratio of the credit spread (the return) relative to the default probability (therisk). There is a default probability by maturity as indicated above. So, more specifically,the metric of interest is the ratio of the credit spread to the maturity-matched default probability.The investor can then select the issue that offers the highest credit spread per basispoint of default risk for each maturity.

What is needed is an estimate of the credit spread for each maturity for each bond issue.This involves two steps. First, the following yields are obtained from TRACE-reportedtrades for each BAC bond traded on that day: the lowest yield, the trade-weighted averageyield, and the maximum yield. In addition, a yield consistent with the best-fitting trade-weightedcredit spread is computed from the trades.The next step is to compute the credit spread based on some benchmark.

The analysisperformed in this chapter is an application of how a particular credit risk model can be used in identifyingrelative value among bond issues by the same issuer. The analysis can be extended toassess relative value among bonds of other issuers. This is done in the case of BAC's bonds bydetermining whether the return-to-risk ratios are "normal"—that is, determining whetherthey are above or below average. To address this question, it is necessary to compare thereturn-risk ratios to the credit spread to default probability ratios to a universe of similarbonds issued by other corporations. The chapter's analysis suggests that on a relative value basis none of the BAC bondissues is attractive.