Predicting U.S. Bank Failures with MIDAS Logit Models
| Item Type | Journal paper |
| Abstract | 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. |
| Authors | Audrino, Francesco; Kostrov, Alexander & Ortega, Juan-Pablo |
| Journal or Publication Title | Journal of Financial and Quantitative Analysis |
| Language | English |
| Subjects | economics finance |
| HSG Classification | contribution to scientific community |
| HSG Profile Area | SEPS - Quantitative Economic Methods |
| Refereed | Yes |
| Date | December 2019 |
| Publisher | Graduate School of Business Administration |
| Volume | 54 |
| Number | 6 |
| Page Range | 2575-2603 |
| ISSN | 0022-1090 |
| Publisher DOI | https://doi.org/10.1017/S0022109018001308 |
| Contact Email Address | alexander.kostrov@unisg.ch |
| Depositing User | Alexander Kostrov |
| Date Deposited | 17 Oct 2019 18:01 |
| Last Modified | 20 Jul 2022 17:39 |
| URI: | https://www.alexandria.unisg.ch/publications/258150 |
DownloadFull text not available from this repository. (Request a copy)CitationAudrino, Francesco; Kostrov, Alexander & Ortega, Juan-Pablo (2019) Predicting U.S. Bank Failures with MIDAS Logit Models. Journal of Financial and Quantitative Analysis, 54 (6). 2575-2603. ISSN 0022-1090 Statisticshttps://www.alexandria.unisg.ch/id/eprint/258150
|
