The main goal of this diploma thesis is to price the dependence of extreme events (triggers) among different asset classes and regions on the portfolio level. For this means, a new product group, namely cross asset portfolio derivatives, is introduced and explained under the light of already existing related products and their pricing methods. Three models for the pricing of the underlyings are chosen: The BNS model serves as model for the stock market and the CIR model is used for all other asset classes except for the Japanese deposit rate, which is modelled by the Vasicek model. The distribution of triggers is attained by a Monte Carlo simulation under the risk neutral measure, which allows for the pricing of the underlying univariate trigger products. The dependence modelling between the univariate trigger products is done via the concept of copulas, allowing for an independent modelling of the marginal and the joint distribution. The estimation of the copula parameters is carried out using the historical trigger dependence of Kendall's rank correlation. Besides the market standard Gauss copula and the Student t-copula, several Archimedean and partly nested Archimedean copulas are applied. For the considered copulas, estimation as well as simulation algorithms are presented. Two examples illustrate the influence of the different copulas used in the pricing routines on the fair price (spread) of cross asset portfolio derivatives.
«
The main goal of this diploma thesis is to price the dependence of extreme events (triggers) among different asset classes and regions on the portfolio level. For this means, a new product group, namely cross asset portfolio derivatives, is introduced and explained under the light of already existing related products and their pricing methods. Three models for the pricing of the underlyings are chosen: The BNS model serves as model for the stock market and the CIR model is used for all other ass...
»