Pricing Covered Bonds

Covered bonds1 have emerged as a potentially preferred funding vehicle from the credit crisis, and they were already a major part of many European financial systems. Although some prices are directly available on, for example Bloomberg2, there is no detailed examination of how covered bonds should be priced taking into account the features that make them attractive to investors: i.e. over-collateralization of the bankruptcy-remote reference pool of assets, and covenants by the pool manager (issuer) on asset replacement. This study provides the first pricing methods for covered bonds based on a Triggered Refreshed CDO with Issuer Risk model that we introduce here. Since covered bonds are relatively new to the pricing literature we include a generic description in Section 2. We define a Refreshed CDO as one where the first m assets that default are replaced. A Triggered Refreshed CDO is one where replacement is triggered by a credit event other than default, e.g. downgrade, restructuring, etc. m expresses the liquidity available from the issuer to replace assets. A further step is to include issuer default not caused by the coverpool directly. We define this as a Triggered Refreshed CDO with Issuer Risk. Issuer Risk is important because, whilst the assets of the coverpool are bankruptcy-remote, the covenants on the manager are not, and the manager may default for reasons not directly related to the coverpool. We provide analytic pricing for Triggered Refreshed CDOs with or without Issuer Risk, i.e. covered bonds, based on the factor Copula model. The analytic pricing is based both on direct calculation of the underlying asset default distribution, and an alternate recursive characterization. Since both of these methods are numerically intensive, involving integration over high