Abstract
M.Comm. (Financial Management)
If the creditworthiness of applicants for overdraft facilities is assessed inaccurately, future
defaults on repayment commitments may threaten the cash flow and profitability of the
bank issuing the credit. The main purpose of this study is to assess how reliable the
inputs used by one large South African bank to determine the creditworthiness of
individual clients applying for overdraft facilities on personal current accounts are in
predicting the future behaviour of the client. The findings of this study should inform the
bank in question (and other credit providers) of the extent to which the original
application-based score (OABS), is an accurate predictor of a client's future behaviour
after the account is opened. It will also help them to identify which of those variables are
the best predictors, and so reduce the risk of default. A quantitative research
methodology is employed, using secondary data on individual applications processed at
one large South African banking institution between July 2006 and July 2009. The data
was used to establish whether any (and which) of the input variables captured by the
bank when the client originally applies for credit, are strongly associated with the
behavioural risk indicator (BRI) assigned to them after the first six months (BRI 1) and
twelve months (BRI 2) after being granted a loan, respectively. The findings of this study
revealed that non-financial characteristics (biographical and demographic information) are
not considered in the OABS; while finance-related characteristics (segment, income
bands, overdraft taken up category and overdraft taken up as a percentage of gross
income categories) are. The study found that there is also an association between the
OABS and the behavioural scores (BRI 1 and BRI 2) allocated thereafter. In particular,
the number of days the client went into excess during the initial six and subsequent six
months is strongly associated with the BRI 1 and BRI 2 scores. This finding implies that,
despite the utilisation of non-financial measures to determine creditworthiness scores, a
client’s current behaviour is still the best predictor of future behaviour. The study
concluded that, despite its low predictive power, the OABS is, however, associated with,
and to a certain extent predicts, the future behaviour of clients.