Managing Credit Risk amongst Covid-19

Share on linkedin
Share on facebook
Share on twitter

In our previous article we discussed the development of credit scorecards as a way to fairly assess credit customers and allow credit providers to manage their risk portfolio. Credit scorecards are built on the understanding that consumers’ previous behaviour will predict future behaviour. However, black swan events, such as Covid-19, create large behavioural changes in a population decreasing the accuracy of the scorecard. Pre-emptive policy changes and ongoing monitoring can help a company manage these shifts and maintain the scorecard.

Incline has built scorecards for national and international clients over the past 4 years. The scorecards identify “risky behaviour” attributes which lead to a statistically proven higher likelihood of default, and return a score. The score (bucketed below into Applicant Risk categories) is then used to indicate the “riskiness” of the applicant, based on performance of similar applications. A higher score is indicative of lower risk, and therefore a high probability of exhibiting “Good” repayment ability.  

Risk bands showing increased likelihood of poor repayment in the lower bands

A well-developed scorecard will be built on data that is recent enough to be applicable, but broad enough that the scorecard is stable.  However, scorecard efficacy typically worsens through time due to population shifts. Fine-tuning and realigning maintenance on the scorecard can mitigate this decline. In the case of black swan events, there is widespread behavioural change that causes population jumps. These population jumps affect the stability of the scorecard and change the original risk distribution suddenly.  For example, COVID-19 has created large-scale turbulence, GDP has dropped by 17% YoY with 3 million lost jobs. These circumstances are beyond the control of scorecard development and need to be actively managed.

  • Pre-emptive: Adopt a conservative approach under current scorecards, pausing aggressive lending. This may take different forms, for example, a company may stop extending credit to risk bands which have higher risk of “bad behaviour”. Alternatively, the company will extend credit but offer a lower credit limit.  
  • Ongoing: Monitoring can be done at regular intervals, looking for early shifts in the population. Early indicators of problems can be identified through stability analysis and vintage curves.
Vintage curve showing increased "bad" accounts for 2020
  • Consider scorecard rebuilds: If the scorecard lacks any predictability, a rebuild may be needed . Either a standard rebuild or using a shortened time span for the observation and evaluation to accommodate only the data relevant to the black swan event (example timeline for a Short-outcome/ Strict-performance scorecard).

Despite the large impact Covid-19 has had on the population, companies have various tools available to manage their risk. First, conservative pre-emptive policy changes can prevent riskier accounts from entering the book. Monitoring of the scorecard, gives early warning of population changes. Fine-tuning and re-alignment can extend the life of a scorecard. And lastly if needed the scorecard can be rebuilt on newer data. To see how Incline could help your company manage its risk portfolio see our website here or how we helped a company with a credit health check here.