Correlation breakdown, copula credit default models and arbitrage

TL;DR

This paper investigates the correlation breakdown in credit default swap models, arguing it reflects a fundamental market inconsistency that can lead to arbitrage opportunities, and presents a framework for constructing arbitrage portfolios.

Contribution

It provides a novel analysis of correlation breakdown as a market inconsistency and introduces a method for arbitrage portfolio construction under such conditions.

Findings

Correlation breakdown indicates market inconsistency.

Models require correlations over 100% to fit market prices.

Arbitrage portfolios can be constructed under correlation breakdown conditions.

Abstract

The recent "correlation breakdown" in the modeling of credit default swaps, in which model correlations had to exceed 100% in order to reproduce market prices of supersenior tranches, is analyzed and argued to be a fundamental market inconsistency rather than an inadequacy of the specific model. As a consequence, markets under such conditions are exposed to the possibility of arbitrage. The general construction of arbitrage portfolios under specific conditions is presented.