Adjustable Network Reconstruction with Applications to CDS Exposures

We develop an empirical Bayesian methodology to reconstruct weighted directed networks from the total in- and out-weight of each node. This problem arises in the analysis of systemic risk of partially observed financial... [ view full abstract ]

We develop an empirical Bayesian methodology to reconstruct weighted directed networks from the total in- and out-weight of each node. This problem arises in the analysis of systemic risk of partially observed financial networks. Importantly, our methodology can be adjusted such that the generated networks satisfy certain desired global topological properties such as a given mean density. We apply our methodology to a novel data set containing 89 financial networks of credit default swap exposures. Our methodology performs well under a wide range of performance criteria and also compared to other existing methods.