Analytical indicators in discriminant analyzes of credit files

Abstract

The discriminant analysis technique for credit decision on loan files represents the interest of being able to classify and compare populations of an opposite character. Therefore, to be able to “discriminate” or separate to the best the “good” of “bad” files.
This separation would be accessible through a linear discriminating function, composed of financial weighted ratios, according to their predictive power.
A panorama of the different one-dimensional and multidimensional methods applied shows that the use of accounting information emanating from corporate financial documents, is achieved in the form of known traditional ratios and used in classic financial analysis.
What about the analytical performance indicators? We mention the concepts of break-even-point, safety margin, safety index, sampling index, operational leverage, for example, may have an important failure prediction powers.
Would it be relevant to incorporate them as part of these studies?
Only tests done on the discriminant functions to establish, would provide answers to these questions.