A two-stage logistic regression-ANN model for the prediction of distress banks: Evidence from 11 emerging countries

This paper proposes a new approach of two-stage hybrid model of logistic regression-ANN for the construction of a financial distress warning system for banking industry in emerging market during 1998-2006. The proposed two-stage hybrid model integrated the benefits of logistic regression and ANN while preventing computational complexity. Some innovative treatments are adopted so that the proposed approach was found to outperform than traditional models. First, the "optimal cutoff point" approach proposed by was adopted to determine the cutoff point for financial distress. Additionally, cross-validation was used to evaluate the prediction power of the proposed model. The results found that the factors of liquidity, capital, and asset quality are crucially related to the financial distress of banks in emerging market. In the prediction of financially distressed banks, the proposed two-stage hybrid model provided best fit measure based on the RMSE and R2, and demonstrated enhanced prediction power than conventional ones. Thus, the proposed hybrid model was demonstrated to be valuable and reliable for the construction of a financial distress warning system.