Predicting U.S. Bank Failures with MIDAS Logit Models
Journal
Journal of Financial and Quantitative Analysis
ISSN
0022-1090
Type
journal article
Date Issued
2019-12
Author(s)
Abstract (De)
We propose a new approach based on a generalization of the logit model to improve prediction accuracy in U.S. bank failures. Mixed-data sampling (MIDAS) is introduced in the context of a logistic regression. We also mitigate the class-imbalance problem in data and adjust the classification accuracy evaluation. In applying the suggested model to the period from 2004 to 2016, we show that it correctly classifies significantly more bank failure cases than the classic logit model, in particular for long-term forecasting horizons. Some of the largest recent bank failures in the United States that had been previously misclassified are now correctly predicted.
Language
English
HSG Classification
contribution to scientific community
HSG Profile Area
SEPS - Quantitative Economic Methods
Refereed
Yes
Publisher
Graduate School of Business Administration
Volume
54
Number
6
Start page
2575
End page
2603
Contact Email Address
alexander.kostrov@unisg.ch
Eprints ID
258150