CMStatistics 2022: Start Registration
View Submission - CMStatistics
B2032
Title: Decomposing systemic risk and the roles of contagion and common exposures: A structural approach Authors:  Ruben Hipp - Bank of Canada (Canada) [presenting]
Grzegorz Halaj - ECB (Germany)
Abstract: A derivative function is used to estimate contagion based on granular and confidential data. This function allows for the endogeneity of the banks capital via value functions of assets and liabilities. With a structural regression similar to that in the structural VAR literature, we infer how the success and risk of one bank have been transmitted to other banks. We accomplish this by allowing banks to be correlated solely via contagion -- from direct exposures, fire sales, and market-based sentiment -- or via common exposures -- from portfolio overlaps. We apply this model to confidential network data of the Canadian banking market, which consists of six large banks and five smaller ones. We discover that contagion has fluctuated over time, with the highest levels around the Great Financial Crisis (GFC) in 2008 and lower levels for the pandemic periods. Finally, we decompose contagion and common exposures by channels. We find that some channels' strength has notably changed since the GFC, hinting that interbank contagion has structurally changed.