Pricing credit linked financial instruments. Theory and empirical evidence (Q5960178)

This monograph is based on the author's Ph.D. thesis written under the supervision of Professor Wolfgang Schmid at the University of Frankfurt/Oder. The author, now Head of Credit Analytics from Risklab, Munich, develops a new three-factor defaultable term structure model for pricing credit linked financial derivatives. For the mathematical modelling of credit risk in Section 2, the essential risk components are first identified and it is shown that default risk models can be based on default probabilities, i.e. the probability that an obligor will default on his contractual obligations, on recovery rates, that is the proportion of value still delivered after the default, and on transition probabilities, quantifying the fact that a debtor can improve or deteriorate his quality of creditworthiness. Possible approaches of modelling transition and default probabilities, like the historical method, the asset-based method, or the intensity-based approach, are discussed together with various possibilities of modelling recovery rates. In Section 3, the main part of this monograph, a new hybrid model to describe and analyze defaultable term structures is introduced. This three-factor model provides an integrated tool for a joint market and risk management by directly modelling the short rate credit spread as depending on some uncertainty index and assuming that the non-defaultable short rate process has a particular mean-reverting structure. Mathematically, the three components, i.e. nondefaultable short rate, the short rate credit spread, and the uncertainty index are modelled as a three-dimensional SDE admitting a unique weak solution. Section 3.2 describes the basic set-up, i.e. model definitions and existency results, followed by some general valuation formulas for contingent claims. In Section 3.3, the pricing of defaultable fixed and floating rate debts is discussed in some detail, including the investigation of defaultable discount bonds, (non-callable) fixed rate debt, callable fixed rate debt, floating rate debt, and interest rate derivatives. Section 3.4 is devoted to deriving (closed form) solutions for the pricing of credit derivatives, such as credit options, credit spread options, default swaps and default options, and some other credit derivatives. In order to allow for a numerical analysis, a discrete-time version of the three-factor model is discussed in Section 3.5, modifying an earlier two-stage procedure of Hull and White (1994). The discrete model not only allows for a pricing of default related derivatives which do not have any closed form solutions, but it also provides a computational tool for extensive numerical simulations of the new model. As an add-on, the author also discusses the possible fitting of his model to market data. Two methods are suggested for estimating the parameters of the underlying interest rate dynamics, the method of least squares minimization and a more sophisticated Kalman filtering methodology, taking time series of market values of bonds into account. As an example, a number of in-sample and out-of-sample tests are applied to confirm that the proposed model is able to explain e.g. market data on Greek and Italian credit spreads to German Government bonds. Finally, based on the discrete-time version of the defaultable term structure model, portfolio optimization can be achieved via Monte Carlo simulation of the process for a given set of bonds and a future time horizon. It is shown how the portfolio composition can be optimized by maximizing the expected final value or retun of the portfolio under given constraints. Again, a case study of German, Italian, and Greek sovereign bonds illustrates the applicability of the new method. In conclusion, this monograph provides a new and successful approach to the analysis of credit markets, in particular to the pricing of default related instruments and to an integrated credit risk management. It is a useful source for academics, finding a rigorous mathematical treatment of the underlying models, as well as for practitioners who are provided with analytical and numerical solutions for their real-life data analysis.