Dynamic Interpretation of Emerging Risks in the Financial Sector
Forthcoming Review of Financial Studies
77 Pages Posted: 11 Jun 2016 Last revised: 21 Mar 2019
Date Written: February 28, 2018
We use computational linguistics to develop a dynamic, interpretable methodology that can detect emerging risks in the financial sector. Our model can predict heightened risk exposures as early as mid 2005, well in advance of the 2008 financial crisis. Risks related to real estate, prepayment, and commercial paper are elevated. Individual bank exposure strongly predicts returns, bank failure and return volatility. We also document a rise in market instability since 2014 related to sources of funding and mergers and acquisitions. Overall, our model predicts the build-up of emerging risk in the financial system and bank-specific exposures in a timely fashion.
Keywords: Emerging Risk, Bank, Financial Institutions, Financial Crisis, Computational Linguistics, Text Analytics, Failure
JEL Classification: G21, G28, E66
Suggested Citation: Suggested Citation