Banks contribute positively to society by enabling simple, safe and efficient management of money, writes Pierre Venter. This is the third of a six-part series over successive days.

Probability of default (PD)

Banks use predictive models that are mostly based on the customer’s historical credit behaviour, in order to assess a customer’s propensity to repay the advanced loan diligently, or for the customer to default in repayment (assuming that affordability at inception is in order).

These probabilities vary across products, with unsecured products generally reflecting riskier business than secured loans, as banks can recover a portion of the debt for secured loans from the security offered by the customer in support of the loan (for example, a mortgage bond).

Loss given default (LGD)

Once a client has defaulted, a lender calculates the loss that it expects to incur as a percentage of the balance at the point of default. In the case of a mortgage loan, the loss incurred is the final write-off amount after the property was repossessed and a legal sale in execution was done, or the property was voluntarily sold.

LGD is therefore based on the actual property, whereas PD focuses on the probability of the client repaying the loan.

Table 1: An example of calculating loss.

Balance at the point of Default R400 000
Value realised at the auction R250 000
Less Credit Loss Cover R70 000

Based on the example in Table 1, a bank expects to lose R80 000 from a R400 000 balance, which is a 20% LGD (R400 000 divided by R80 000).

In the above example, loan cover (the credit loss cover), reduces the LGD by 17.5 %. Even though customers pay for the credit loss cover, they receive a better interest rate than a loan without this, as the LDG is lower. Such cover enables banks to increase their risk appetite and still be able to offer a more affordable interest rate.

Pierre Venter is the general manager, Human Settlements in Market Conduct Division at the Banking Association South Africa.