CMStatistics 2017: Start Registration
View Submission - CFE
A0694
Title: Adjustable network reconstruction with applications to CDS exposures Authors:  Axel Gandy - Imperial College London (United Kingdom)
Luitgard Veraart - London School of Economics (United Kingdom) [presenting]
Abstract: The problem of reconstructing weighted directed networks from the total in- and out-weight of each node is considered. This problem arises, for example, in the analysis of systemic risk of partially observed financial networks. Typically, a wide range of networks is consistent with this partial information. We develop an empirical Bayesian methodology that can be adjusted such that the resulting networks are consistent with the observations and satisfy certain desired global topological properties such as a given mean density. Furthermore, we propose a new fitness based model within this framework. We apply our methodology to a novel data set containing 89 financial networks of credit default swap exposures. The performance of the reconstruction methodology is very good under a wide range of performance criteria and also compared to other existing reconstruction methods. In particular, we are able to reconstruct the degree distribution of the underlying networks with remarkable precision if a good estimate of the true density of the underlying network is available.