The multivariate option iPoD framework: assessing systemic financial risk

We derive multivariate risk-neutral asset distributions for major US financial institutions (FIs) using option implied marginal risk-neutral asset distributions (RNDs) and probabilities of default (PoDs). The multivariate densities are estimated by combining the entropy approach, dynamic copulas and rank correlations. Our density estimates yield information about the conditional distributions of the individual FIs, and we propose several financial distress measures based on default scenarios in the financial sector. Empirical results around the period of the US sub-prime crisis show that the proposed risk measures identify in a timely manner: i) the most distressed FIs in the system; ii) the systemically most important FIs; iii) the implicit bailout guarantees given to some FIs; and iv) a "too connected to fail" problem in the US financial sector throughout the year 2008.