Title: Consistency of bank defaults in financial networks with continuous credit default swaps
We investigate financial networks consisting of both simple debt ("bonds") and credit default swap (CDS) derivative contracts. We are interested in the problem of consistency: whether there even is an answer to the question who's in default and who is not and whether that answer is unique. We showed in a first simple model that this problem is not at all trivial: inconsistency is common to occur, detecting it is NP-complete, and preventing it is not easily achieved by existing policies. We further developed a framework for dependency analysis that can help detect the "problematic spots" in a network.
Our first model made an important simplification: CDSs would only depend on whether or not a default occurred while real CDSs depend continuously on the exact rate of recovery of the reference entity. In this talk, I would like to discuss with you a refined version of our model in which we can model the "continuous" variant as well as consequences for our results and presentation.