Risk Managers in Singapore Discuss Covid Challenges in Model Risk Management

The Singapore Chapter of the Risk Management Association (RMA) hosted a “Thought Leadership Discussion” led by Julian Philips, COO and Head of Data Science, SPIN Analytics, on Model Risk Management in the Age of COVID-19.

Essential Takeaways from the Thought Leadership Discussion:

COVID-19 was a crisis of speed and accuracy for risk management models.

The pandemic featured previously unseen levels of market turbulence and government intervention.

Backward-looking risk management models rapidly became obsolete, yielding inaccurate results.

Expert human judgement proved key to managing risk, yet current modelling tools left the experts overwhelmed.

The crisis highlighted the need for banks to accelerate the development of forward-looking models, including
new sources of information, using new tools and techniques.

Automation is essential to allow experts to focus more on using their knowledge, experience, and judgment.

“As with previous crises, expert judgement played a crucial role during the COVID-19 pandemic, and risk management technology must keep human risk management experts at the center and in control.” Frankie Phua Peng Yeo, Head, Group Risk Management, UOB, and Chairman of The Singapore Chapter of the Risk Management Association.

“Singapore is at the forefront of innovation in the financial sector, and the RMA play a vital role by raising awareness in use of new technologies in areas like Model Risk Management, and by discussing potential challenges around use of AI in risk related processes.” Gautam Mukharya Chief Risk Officer, HSBC Singapore, and Director of The Singapore Chapter of the Risk Management Association.

“Solving the speed and accuracy crisis in credit risk modeling requires new tools which go beyond statistical modelling to cover the multiple dimensions of credit risk including data preparation, regulations, documentation, and expert judgement.” Julian Phillips, COO and Head of Data Science, SPIN Analytics