How COVID-19 is changing credit risk models

Credit model owners should focus on certain to ensure that the output of their models remains valid and robust under the unprecedented change.

Faced with the unprecedented pace and magnitude of economic disruption from the COVID-19 pandemic, risk modeling teams are challenged to develop a Now, Next and Beyond response:

• Current models in the prudential domain were built for an economic downturn, but not a sudden halt in both supply chains and demand side of economic activity. The acuteness of this impact is beyond anything in history, so risk modeling teams must carefully question how and when historical data can be relevant to forward-looking credit analysis.
• The varied social distancing policies implemented by governments and inherent attributes of COVID-19 that we still do not fully understand mean that this pandemic is developing in an asynchronous manner across the world. Consumers responses are partially guided by psychological fear, making it difficult to predict otherwise rational decisions, such as labor supply and consumption of services, involving close proximity to others. The impact of COVID-19 and the path to recovery will vary widely by sector and geography and will be further exacerbated by the interlinked character of the global economy. Sector and region dynamics are also influencing unemployment demographics, a critical driver for assessing consumer credit risk.
• Government stimulus activities that aim to alleviate both individual and business financial distress are without historical precedent. In the corporate credit space, government-backed lending programs may mitigate defaults in the short to midterm, but they will increase leverage, which in turn will further compound widespread downgrades. In the consumer space, payment holidays and new guidelines on forbearance are masking the traditional delinquency indicators such as the days-past-due metrics.