A study on the dependence among spanish banks via quantile regression forests and network analysis

The present work is focused on the estimation of Covar and Delta-Covar, two measures that are used to quantify the effects of a financial institution or a financial system in distress on the VaR of another financial institution. Firstly, the studied case consists of five Spanish banks, namely: BBVA, Bankia, Bankinter, CaixaBank, Santander and, in the second part, the MSCI Europe index is used to analyze the financial system effects. The method adopted is the quantile regression forests and the results are compared with the already well-known quantile regression. The last part of the thesis studies the relations among the banks and the system as a network. Considering additional European banks, a community detection analysis is performed.