There is an increasing need for effective monitoring of financial institutions, accentuated by the impact and lingering effects on economic activity of financial crises. Yet developing new methods for monitoring financial firms and markets is challenging given the intricate interconnectedness of financial institutions, which can function as a mechanism for propagation and amplification of shocks through the system and also in fostering systemic risk.

This grant award supported several research papers that developed methods for effective monitoring that also provide insights into the interactions between financial entities. For example, our newly developed change point detection methods can indicate a shift in the underlying dynamics of the interbank market while also pinpointing specific banks that are likely the cause of the change. Such interpretable and semi-automated methodology can be extremely useful to regulators, risk managers, and policy makers. Our research generally combines financial and economic data streams through network and statistical modeling, while the specific tools are diverse and include vector auto regressions, mixture models, constrained matrix factorization, and other statistical and econometric techniques.

From a financial economics perspective, most of the extant work in systemic risk and interconnectedness is theoretical in nature, because the required data for empirical work is not readily available or can only be imprecisely estimated from regulatory filings. The methods we developed help address this key issue by inferring properties of the balance sheet (e.g., holdings of a common asset) at much higher frequencies than regulatory filings allow. Our methods will also add to the toolkit that financial economists have to estimate financial networks, summarize them, and thus identify potential vulnerabilities within the financial system.

Last Modified: 12/30/2021
Modified by: Shawn Mankad