Next generation of Stress Testing: Using as a tool and BAU integration

By Liang-Khoon Koh, Director Stress Testing, Silicon Valley Bank

What, for you, are the benefits of attending a conference like ‘Stress Testing USA’?

The biggest benefit of attending a conference like this is the opportunity to learn from the subject matter experts in the field. While Stress Testing results are public information, the methodology and process are not. It is definitely an invaluable opportunity to be able to get to make contacts with experts directly and hear their thoughts and their methodology even at very high level.

What do you see as the future of Stress Testing and integrating into BAU?

Given the past few years of intensive effort in preparing for DFAST and CCAR, the stress testing mentality is now much better embedded into organizations and has become part of most banks’ DNA. The future of Stress Testing will continue to evolve into a more streamlined, integrated, and end-to-end processes. Organization weaknesses uncovered during stress testing will continue to drive improvement in BAU. I expect BAU to continue to leverage stress testing models, methodology and scenarios. We expect to see more collaboration between FP&A teams and stress testing teams.

How can institutions best balance moving Stress Testing into BAU whilst maintaining the level of rigor to ensure effective execution?

I don’t foresee this will be a big challenge as long as the organization’s key resource still focus around Stress Testing processes and methodologies. A full review from top to bottom level in Stress Testing process will help to maintain the level of rigor to ensure effective execution.

What are the main challenges and opportunities of BAU integration?

Business strategy integration is one of the bigger challenges. This also presents an opportunity to enhance stress testing models. For example, stress test models might not already integrate at a very detailed level the kind of strategies that the BAU typically involve in. Cross function communications can always be a big challenge to such integration due to differences in skill set and focus. Automation continues to be a challenge especially for small and mid size organizations.

In your opinion, what will the next generation of Stress Testing look like?

The next generation of Stress Testing will be more streamlined and integrated across organizations. Automation will continue to play an important role in order to make this happen.
Fed has designed the 2019 severely adverse scenario to be even tougher than any prior years. We expect the Fed to continue doing this. As such, Stress Testing models will continue to evolve to deal with such tougher scenarios. For example, the Fed itself has incorporated some adjustments into its Stress Testing model in this year’s DFAST examination. The Fed adjusted the way it calculates projected defaults and projected losses for auto loans. It also fully phased in changes to the way it estimates credit card losses. Both of those alterations resulted in higher projected losses for banks with auto loan and credit card exposure. The Fed also refined its treatment of corporate loans for which there is incomplete data, adopting a “conservative” model that resulted in a “small decrease in overall losses” over prior iterations of the model. We expect the next generation of Stress Testing to have more enhanced models to deal with new stress, tougher scenarios, and further quantify risk across organizations that are not currently already done.