Neural Networks: A New Tool for Predicting Thrift Failures

Abstract

A neural network model that processes input data consisting of financial ratios is developed to predict financial institution failure. The network's ability to discriminate between healthy and failed institutions is compared to a traditional statistical logit model. We find that the neural network has performed as well or better than logit for failure prediction. We also observe that, in some cases, when the cutoff point was lowered, the reduction in Type I errors committed was accompanied by greater increases in Type II errors for the logit model than for the neural network. This may be an important result when examiners factor in the cost of committing Type I and Type II errors. The neural network, which uses the same financial data, requires fewer assumptions, achieves a higher degree of prediction accuracy, and is more robust.