
In practice, the general tendency is to assign a higher relevance to statistical tech-
niques for small borrowers, which do not generate considerable volumes of business, and
to assign higher relevance to human judgment as the actual/potential business secured by
the borrower becomes bigger. This can be explained by a number of factors.
•
Small exposures may not justify the cost for a trained analyst to evaluate the
loan.
•
Larger borrowers may be too complex to be treated through generic statistical
techniques such as scoring models, and they cannot be evaluated through third-
party judgements only (e.g., agency ratings, EDF) if the bank want