- Average balance with regard to value dates
- Most favorable balance for the borrower
- Most unfavorable balance for the borrower
- Credit turnover
- Debit turnover
- Bill of exchange credits
- Check credits
- Transfer credits
- Cash deposits
- Bill of exchange debits
- Check debits
- Transfer debits
- Cash payouts
- Limit
Profile analysis, dichotomous classification, and linear discriminant analysis were
the three techniques applied on the data. All three methods revealed important dif-
ferences between the bad and the good companies. Linear discriminant analysis pro-
vided the best results. The function contained the following variables:
- (Most favorable balance for the borrower)/limit
- (Most favorable balance for the borrower)/debit turnover
- Check debits/debit turnover
- Debit turnover/limit
- Bill of exchange debits/debit turnover
- Transfer credits/credit turnover
The classification results (from von Stein and Zeigler [1984]) on the development
sample are shown in Exhibit 10.3
The third phase of the study attempted to identify the characteristics and concrete
behavioral indications that distinguish the failed firms from the solvent ones. The au-
thors used a psychological technique named “nomethetical assessment” and the
“principle of simultaneous vision.” The latter term is taken to mean that the authors
looked for factors consistently found in the failed group that are consistently absent
10.4 GERMANY 10 • 11Group Year Before Fixed Date Correct Classification
5 71.4%
4 78.2
Bad cases 3 86.6
2 89.9
1 95.0
Good cases 1977 83.7
Exhibit 10.2. Classification Results—Fix/Hodges Nonparametric Model.