27 March 2014

Jonathan Crook, Director of the Credit Research Centre and Director of Research, gave an invited presentation on the benefits of survival analysis for credit risk modelling. Jonathan summarised the latest research on the application of survival modelling to probability of default prediction, combination scores, profit scoring and stress testing.

On 27 February Professor Jonathan Crook gave an invited presentation to the Credit Risk and Collections Forum in Bucharest, Romania. Jonathan explained to the audience the benefits of survival analysis techniques when modelling credit defaults compared with traditional methods. The talk drew from several papers he had published recently including work co-authored with Mindy Leow and Tony Bellotti. Survival models allow a bank to predict the probability that a borrower will default for the first in a particular month which an analyst can choose. The technique can be applied to calculating the expected profit from a loan to a particular borrower and also allows the bank to predict the effects of a change in the Economy on the probability that a particular borrower defaults. Survival models also allow a bank to stress test a loan portfolio by choosing values of items like the unemployment rate, the interest rate and other aspects of the Economy multiple times from past values and then finding an extreme value of the resulting probability of default distribution.