The main purpose of this Dissertation is to examine a sample data of 108
matched
pairs of failed and non
-
failed banks of Eastern and Western Europe., according to the
size of total assets, over the 2006
-
2016 period. Techniques such as Multiple Logistic
Regression and Multiple Discriminant Analysis based on Camel Rating System
were
applied on report data for one, two and three years prior to failure so as to determine
the robustness of bankruptcy prediction models for European Banks.
The logit Model predicts bank failure with 81,75% accuracy in comparison with the
79,55% for the
MDA Model, one year prior to bankruptcy. Nonetheless, MDA Model
outperforms the Logit Model showing 72,44% accuracy two years prior to failure and
67,06% accuracy three years prior to failure.
Collections
Show Collections
