Professor Michael Hanke and Merlin Bartel from the University of Liechtenstein and Sebastian Petric from LGT Bank AG presented their latest research entitled "Crisis Identification and Prediction using Machine Learning: The Case of U.S. Regional Banks" at the Bank of England.
In their innovative study, the team develops a machine learning approach to identify and predict financial crises, particularly those of regional banks in the US. The research emphasizes the importance of early detection of crises, as they have a significant impact on the economy and society. The developed approach combines various machine learning methods such as clustering, ridge regression, and sequential feature selection to detect and predict crises.
Notably, the predictions can be used to minimize risk during negative market phases, leading to improved risk-adjusted returns compared to traditional investment strategies.
This research provides investors and policy makers with valuable tools to react early to impending financial crises.