Prediction Bankruptcy for European Banks in 2008-2012
This paper examines a sample of 57 banks that failed between 2008 and 2012, comparing them to a random sample of banks that had not failed as of 2012 by matching the asset size of healthy and bankrupt banks. We performed logit regression and multiple discriminant analysis on annually report data for one year, two years and three years prior to bank failure to determine whether the failure could have been predicted before. Our logit model outperform the MDA model and is statistically significant at 1% level and predicts bank failures with 82.5% accuracy one year prior to bankruptcy, 75% two years prior to failure and 66.25% three years prior to default event. Finally, the models are validated using a holdout sample ex-ante test procedure where in both logit and MDA approach provides encouraging results.I have been fortunate that my advisor was Prof. Andreas Charitou and I would like to thank for the help and guidance Dr. Fragiskos Archontakis has given me.