The importance of models in managing credit risk

You need only read the latest news to know that the Australian economy is experiencing great instability. Australian consumers are under ever increasing cost pressures and interest rate rises intended to bring inflation down are adding to the consumer burden. Economists may continue to debate what impacts this uncertainty will have in the medium to long term, but what is certain is that many organisations have significant credit exposures that will be affected by whatever happens next and the decisions these organisations make now to manage their credit risk.

Good credit decisions require credit models that are fit for purpose and based on sound assumptions. However, credit model assumptions, particularly those about consumer behaviour, become outdated over time even in relatively calm economic conditions. Such ‘model drift’ occurs because of slow but steady changes in consumer behaviours, the effects of competition and new products such as Buy Now Pay Later entering the market. Significant bouts of economic uncertainty and disruption also tend to accelerate model drift. As models drift, the quality of credit decisions deteriorates.

The COVID-19 pandemic is a recent example of what can happen to credit modelling when macro-level uncertainty hits. The rapid onset of the pandemic saw a recalibration of credit decisioning to tighten standards, payment holidays and a pause in collections activity. Credit models were significantly affected, at times overridden, or even turned off.

There is no such thing as ‘normal’. Now more than ever business leaders must be vigilant of changing conditions and consumer behaviours. Credit models should be closely monitored for whether the assumptions on which they are based are holding true over time. If organisations do not detect a drift in real-world behaviour from a model’s assumptions in a timely fashion, it can develop into a weakness in credit decision making and ultimately pose financial risks to the organisation as well as leading to poor customer outcomes.

Those with carriage of credit exposures should be asking the following questions:

1. Am I comfortable with how regularly I monitor the performance and stability of each of my credit models beyond their date of implementation, especially given increased economic uncertainty?
2. Am I continually looking ahead and challenging myself with what information and data is available to measure the credit risk of my customers?
3. Am I challenging myself with my modelling assumptions and identifying if they are still relevant today and in the future?
4. Can I take faster pulse checks on adherence of customer behaviours to model assumptions early in a credit model’s deployment and/or during times of macro-economic uncertainty?
5. Do I have the latest tools to identify and understand the drivers behind instances of model drift or model deterioration early on, and where to focus efforts to address these?

As Australia faces into tougher times, it is imperative that organisations equip themselves with the tools they need to successfully navigate their credit portfolios through this next period of change and uncertainty.

Co-author
Rob Till
Director in Risk Advisory
Connect on LinkedIn