nonfailed firms. The original function was built based on a matched (by industry,
size, and year) sample of 50 failed and nonfailed entities from 1974 through 1979.
Collongues (1977), Mader (1975 and 1979) also have attempted to combine fi-
nancial ratios with data from failed and nonfailed French firms. Mader’s studies were
descriptive of firms in difficulty and the utility of ratios as risk measures. These have
led to several multivariate studies performed by the Banque de France in their “cen-
trale de bilans” group. Collongues did utilize discriminant analysis in his analysis of
small- and medium-size firms with some success.
The application of statistical credit scoring techniques in the French environment
appears to be problematic, but the potential remains. One problem usually is the qual-
ity of data and the representativeness of them. But this is a problem in all countries
and is not unique to France. The government has gone on record on several occasions
as intending not to keep hopelessly insolvent firms alive artificially but to try to as-
sist those ailing firms prior to total collapse. An accurate performance predictor
model could very well help in this endeavor.
10.9 SPAIN
(a) Fernández (1988). This study describes an empirical model to objectively evalu-
ate and screen credit applicants. The work consists of the determination of the model
with two objectives: (1) to check the validity of financial ratios as prediction tools,
and (2) to predict a firm’s collapse.
The research sample consisted of 25 failed and 25 non-failed firms, with an addi-
tional 10 each being set aside for validation testing. Data pertaining to two years pre-
ceding the failure was collected. Only data pertaining to 1978–1982 was permitted in
order to eliminate the possible distortion caused by the natural changes in ratios
caused by the business cycle. The ratios were examined using three techniques:
1.Univariate analysis
2.Factor analysis by principal components
3.Discriminant analysisThe author concludes that univariate analysis is not practical given the volume of
the ratios to be considered and the possible interactions among the ratios. In addition,
the univariate ratio analysis has to be performed in the context of the market in which
the firm operates, thus the ratios show only relative position of the company. Lastly,
multivariate ratios can improve analyst productivity and free him/her to concentrate
on other equally important matters such as the credit terms, maturity, guarantees, and
so on.
When there are a large number of variables to be considered, principal component
analysis is a way to eliminate the variables that carry the same information and re-
duce the observation to a handful of factors or “principal components.” Each princi-
pal component is a linear combination of one or more of the underlying variables.
The coefficient of the underlying variable in the factor equation is called the “factor
loading.” In this study the author conducted factor analysis in two ways: (1) without
rotation of the factors and (2) using varimax rotation to ensure the independence of
the resulting factors.
The second way is believed to produce more desirable (i.e., stabler) results when
used as independent variables in regression or discriminant analysis.
10 • 20 BUSINESS FAILURE CLASSIFICATION MODELS