An Empirical Analysis of Equity Default Swaps Ii: Multivariate Insights

Abstract

Equity Default Swaps (EDS) - contracts that trigger a payment when the underlying equity price falls below a predetermined level - have attracted much attention recently because of their similarities to credit default swaps (CDS) on the one hand, and American digital puts on the other. Particular interest has been received by Collateralized Debt Obligations (CDOs) referencing a portfolio of EDSs, which not only requires the univariate assessment of the risks inherent in EDSs, but also the analysis of dependencies between EDSs (and other asset classes). In this paper, we specifically address correlation or dependency aspects of EDSs, by applying techniques developed for estimating default correlation. Based on Standard & Poor's CreditPro and Compustat (North America) databases, extensive empirical research is presented. Amongst the main findings are that EDS correlations for standard strikes/barriers of 30% are significantly higher than default correlations, and increase in barrier level, but only for strikes above 50%. This indicates a barrier dependent correlation concept.

Keywords: EDS, Equity Default Swap, correlation, dependence

JEL Classification: C14, C81