A1 Refereed original research article in a scientific journal
Predicting banking crises with artificial neural networks: The role of nonlinearity and heterogeneity
Authors: Kim Ristolainen
Publisher: Wiley-Blackwell
Publication year: 2018
Journal: Scandinavian Journal of Economics
Volume: 120
Issue: 1
First page : 31
Last page: 62
Number of pages: 32
ISSN: 0347-0520
DOI: https://doi.org/10.1111/sjoe.12216(external)
Web address : http://onlinelibrary.wiley.com/doi/10.1111/sjoe.12216/abstract;jsessionid=84E991400EDD7FF223D439DF392F1193.f02t03(external)
The early warning system (EWS) literature on banking crises usually relies on linear
classifiers, estimated with international datasets. We construct an EWS based on an artificial
neural network (ANN) model, and also account for regional heterogeneity in order to improve
the generalization ability of EWS models. All of the banking crises in our test set are then
predictable at a 24 month horizon, using information from earlier crises. For some countries,
estimation with a regional dataset significantly improves the predictions. The ANN
outperforms the usual logit regression, assessed by the area under the receiver operating
characteristics curve.