In this issue, I present four full-length research articles and one technical report. As always, I am very fortunate to have contributors who are pioneers in financial research like Robert Engle, Nobel laureate, and John Hull and Alan White as well as contributors who are highly respected in their particular areas of expertise.
The first article, “Adjusting corporate default rates for rating withdrawals” by R. Cantor and D. Hamilton, addresses the treatment of withdrawn ratings (WR) or, equivalently, not-rated (NR) names in the computation of default probabilities. The authors show that adjusting for WR can matter a lot. Estimated default rates using the adjusted method are never smaller than unadjusted rates, and they diverge as the horizon increases. They also show that the assumption of random censoring, namely that the process of withdrawal is non-informative, is reasonable.
The second article, “The underlying dynamics of credit correlations”, is by A. Berd, R. Engle, and A. Voronov. In this article, the authors bring sophisticated time series methodology into the portfolio credit risk setting. The portfolio credit loss-generating model is derived under different dynamic specifications. More specifically, the authors consider a structural credit model where the latent asset variables evolve according to a one-factor multivariate asymmetric GARCH model. The authors show that the portfolio loss distribution derived from the model can produce implied correlation skews similar to ones observed in the synthetic CDO market. They also assess the ability of various credit loss-generating models to produce realistic pricing of CDO tranches.
The third article, “Forwards and European options on CDO tranches”, is by J. Hull and A. White. By utilizing a new change of measure argument, the authors show how the well-known Black (1976) European option pricing model can be adapted to price forwards and European options contracts on CDO tranches. As cash and synthetic CDOs have rapidly grown into deep markets in recent years, it is a matter of time before trading of options and forwards on CDO tranches becomes widespread. The model presented in this article addresses, in anticipation, the need to value such emerging products.
The fourth article, “Modeling exposure at default, credit conversion factors and the Basel II Accord” by R. Taplin, H. Minh To, and J. Hee, is on the treatment of exposure at default (EAD) for undrawn lines of credit. Little has been written on EAD in the publicly available literature, and empirical research on EAD is almost nonexistent in the literature. As a result, practitioners and regulators are left wanting for more conceptual development and empirical estimation of EAD. Although most of this article is in the context of retail lines of credit (in particular business credit cards), the concepts developed and points made are much more general and apply to treatment of all lines of credit.The last piece in this issue is a technical report. 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 duplication of efforts. The contents of technical reports complement rigorous conceptual and model developments presented in the research articles and provide a lot of value to practitioners.
In the technical report, “Benchmarking low-default portfolios to third-party ratings,” C. Franz and C. Lawrenz provide and compare known techniques and different sets of statistics used for benchmarking default models for low-default portfolios. The authors’ recognition of each of the techniques’ applicability to the benchmarking exercise in hand and their relative ease of implementation will be useful to practitioners, particularly given the benchmarking requirements that the Basel II Accord imposes on banks that are subject to its regulations.The last piece in this issue is a technical report. 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 duplication of efforts. The contents of technical reports complement rigorous conceptual and model developments presented in the research articles and provide a lot of value to practitioners.