In this issue of The Journal of Credit Risk we present three research papers and one technical report.
The issue's first research paper is "Bounds for rating override rates" by Dirk Tasche. This paper examines an important question posed by banks and their regulators with respect to internal credit rating models: what frequency of rating overrides should one reasonably expect in the credit assessment process? The answer proposed by the author is simple and intuitive. The misclassification rate of a model, ie, a weighted combination of the type I and type II errors of the rating model, provides an upper bound on the reasonable frequency of rating overrides. The author then shows that this "natural error rate" of a rating model is closely related to a commonly used measure of the quality of the rating model, ie, the accuracy ratio.
The second research paper in this issue is "Calibration of structural and reduced-form recovery models" by Alexander Koivusalo and Rudi Schafer. This paper uses the Merton model as a reference system, with diffusion and jump diffusion as underlying processes. Three different models of recovery rates (constant rates, the reduced-form approach and the structural model) are calibrated based on the Monte Carlo simulation of the reference system. The comparison results show that the structural recovery model, which involves a special functional relation between recovery and default probabilities, is the most reliable and stable method in terms of both calibration and loss estimation.
In the third research paper, "Credit default swap spreads, fair-value spreads and interest rate dynamics" by Andy Yeh, evidence of the relationship between credit spread and interest rate movements is presented. Interest rates are modeled using Treasury rates and the swap interest rate. Regressions are carried out in log levels and log changes. Evidence supports a negative link between credit spread and interest rate movements. The main contribution of this paper is the use of the credit default swap in place of the plain credit spread, and the application of a so-called fair-value spread that turns out to be less important than assumed. The results show that swap and Treasury rates seem to convey different information about credit spread movements for nonfinancial and financial companies.
A technical report describes a particular practical technique and enumerates situations in which it works well and others in which it does not. Such reports provide extremely useful information to practitioners in terms of saved time and minimizing duplication of effort. The contents of technical reports complement rigorousconceptual and model developments presented in the research papers. A technical report can be a useful survey article as well.
The technical report in this issue is "Optimal structuring of collateralized debt obligation contracts: an optimization approach" by AlexVeremyev, Peter Tsyurmasto and Stan Uryasev. The paper formulates optimization problems that allow optimal attachment points to be found for tranches in synthetic collateralized debt obligations. Several variations are considered, including constant and time-varying attachment points, as well as the optimal composition of the pool of credit default swaps. Simplifications that allow the optimization problems to be solved more efficiently are discussed and a brief case study is presented.