Abstract
In this paper, a normal inverse Gaussian factor model is developed to describe the fat-tailed feature of the default distribution of reference entities in order to study basket default swaps pricing. Based on this model, the explicit formula for the distribution of the kth default time is accurately obtained by making use of order statistics, and the closed forms of the price of BDS at the kth default and m out of n default entities are calculated using the risk-neutral pricing principle.