Enhancing credit default swap valuation with meshfree methods

In this paper, we apply the meshfree radial basis function (RBF) interpolation to numerically approximate zero-coupon bond prices and survival probabilities in order to price credit default swap (CDS) contracts. We assume that the interest rate follows a Cox–Ingersoll–Ross process while the default intensity is described by the Exponential-Vasicek model. Several numerical experiments are conducted to evaluate the approximations by the RBF interpolation for one- and two-factor models. The results are compared with those estimated by the finite difference method (FDM). We find that the RBF interpolation achieves more accurate and computationally efficient results than the FDM. Our results also suggest that the correlation between factors does not have a significant impact on CDS spreads.

► In this study, we provide numerical evidence and show that the meshfree radial basis function (RBF) interpolation outperforms the finite difference method (FDM) in terms of accuracy and computational efficiency in approximating zero-coupon bond prices and survival probabilities in an effort to evaluate credit default swap (CDS) contracts. The advantage of adopting the RBF interpolation increases with the dimensional order of the problems.