Machine learning methods for predicting failures of US commercial bank

In this paper, we attempt to study the effectiveness of various simple machine learning methods in the prediction of bank failures. From a raw dataset of 10,938 US banks during the period of 2000-2020, we find that machine learning approaches do not really outperform the benchmark of conventional st...

Full description

Saved in:
Bibliographic Details
Published inApplied economics letters Vol. 31; no. 15; pp. 1353 - 1359
Main Authors Tuan, Le Quoc, Lin, Chih-Yung, Teng, Huei-Wen
Format Journal Article
LanguageEnglish
Published London Routledge 01.09.2024
Taylor & Francis LLC
Subjects
bank failure
Bank failures
banking system
Commercial banks
financial indicators
financial safety
Machine learning
United States > US
Online AccessGet full text

Cover

More Information
Summary:In this paper, we attempt to study the effectiveness of various simple machine learning methods in the prediction of bank failures. From a raw dataset of 10,938 US banks during the period of 2000-2020, we find that machine learning approaches do not really outperform the benchmark of conventional statistical method, logistic regression. However, using PCA to retain relevant variance in variables significantly improve the performance of machine learning methods and raise the out-of-sample accuracy of those method to over 70% to over 80%. Of all the machine learning methods used in this paper, the simple KNN seems to be the best model in forecasting bank failure in the United States.
ISSN:1350-4851
1466-4291
DOI:10.1080/13504851.2023.2186353