The main purpose of this study is to examine
a dataset of
32
matched pairs of failed and
non
-
failed Eurozone banks
, according to the size of total assets,
over the period 2008
-
2015.
Logistic Regression and Multiple Discriminant Analysis (M
DA), based on the
CAMEL rating
system, were
employed
on yearly
report data for one, two and three
years prior to
failure in order to determine whether
reliable failure prediction models
for Eurozone banks can be developed.
Logistic Regression Analysis outperformed Multiple Discriminant Analysis when yearly
data for one year prior to failure were employed. Notably, the logit model yielded an
overall correct classification accuracy of 82.81% compared to 81.25% for the MDA. On
the other hand, MDA was superior
when yearly data for two and three years were
employed. It yielded an overall correct classification accuracy of 73.44% and 64.06%
respectively, compared to 71.88% and 59.38% for the logit model.
Collections
Show Collections
