Title: Measuring asset holdings in the banking sector with balance sheet driven probability factorization
Authors: Celso Brunetti - Bocconi University and Federal Reserve Board (United States) [presenting]
Abstract: The response of governments around the world to recent financial crises has been to provide regulators with access to several new financial and economic data streams with the promise that the increased data availability will lead to more effective risk monitoring. A major challenge then for regulators is to integrate these data together for new and timely insights into the potential channels for the propagation of risk that could endanger the overall financial system. We provide novel methodology that regulators can use to monitor the response of the interbank and stock markets, yielding insight into the balance sheets of banks at a higher frequency than standard disclosures allow, and complimenting other approaches that build on network science to assess systemic risk levels. The approach is based on a probability matrix factorization that is developed from a rigorous accounting framework to integrate the data sources together. Using data from before, during and after the 2008 financial crisis, we study the behavior of the banking sector to find evidence that measures derived from our approach leads multiple risk related variables published by the European Central Bank.