Description
We briefly present a multi-factor Gaussian copula portfolio model for
default risk. The model assumes three types of systematic factors
driving the asset returns, hence the value of each company. These
factors represent the state of the global economy and the economic
conditions of different geographical regions and industries. The
corresponding factor loadings play a key role in the model, as they
capture the correlation structure between the asset returns of
different companies and therefore influence the joint probabilities of
default. Higher correlation between the returns of different companies
in a portfolio increases the likelihood that multiple companies will
default simultaneously, thus increasing the likelihood of extreme
losses in the portfolio. Hence, accurately measuring these
correlations is essential for the identification of portfolio risk.
We describe a possible methodology for measuring the correlations
between asset returns of different companies, which can be used for
calibrating the corresponding factor loadings. The approach relies
upon single-name CDS spread data. We will also analyze of the
structure of correlations obtained using this methodology.