The purpose of this talk is to illustrate the importance of network-based models in financial industry since the 2007-08 crisis. I will present our novel method to learn a (Bayesian) network using default dependence of financial institutions and show how financial distress can propagate via such a network. I will share some examples where we extensively use NetworkX and bnlearn packages.
The financial crisis of 2007-08, in which the default on debt rose to $430 billion up from just $8 billion, had catastrophic effects on global financial stability. It demonstrated that financial institutions are highly connected to each other --- for example, the default of Lehman Brothers had repercussions for the whole economy. Since then modelling default dependence of financial institutions via network-based approach has gained significant importance.
I will present our novel method to learn a (Bayesian) network using default dependence of financial institutions and show how financial distress can propagate via such a network. I will share some examples where we extensively use NetworkX and bnlearn packages. Finally, I will show applications of these measures in the banking industry.