CASE STUDY
Stress testing a banks’ credit loss model
How we helped a major bank improve its risk management practices and better prepare for potential economic shocks
We worked with a major bank that wanted to ensure its credit loss models were robust and accurately capturing the potential losses that could result from a severe economic downturn.
As a result of our work, the bank was able to improve its credit loss models and better understand the potential credit losses that could result from severe economic downturns. In turn, this helped them improve their risk management practices and ensure that it was adequately capitalised to withstand economic shocks. And the bank was able to maintain its reputation as a responsible and reliable financial institution.
Key results
Comprehensive scenario analysis process implemented
​
More rigorous model validation process developed
​
Improved risk management practices
Our approach
​
Review
Our team began by reviewing the bank’s existing credit loss models, including its methods for estimating credit losses under both ‘normal’ and ‘stressed’ economic conditions. Based on this review, we identified several areas where the bank’s models could be improved.
Scenario analysis
The bank’s existing models relied on a limited number of economic scenarios, which may not have adequately captured the potential range of outcomes that could result from severe economic downturns. We recommended they use a broader range of scenarios, including scenarios that were more severe than those used in the past.
Data quality
We found that the historical data used to estimate credit losses may not have been adequate for predicting losses under severe economic conditions. By reviewing and improving its data quality and considering using alternative data sources, such as macroeconomic indicators, the bank could supplement its historical data to better capture potential losses.
Model validation
As the models had not been rigorously validated to ensure they were accurate and reliable, we recommended that the bank conduct a model validation exercise to identify any weaknesses in its models and improve their accuracy.
Deliverables
Based on our recommendations, we worked with the bank to develop new credit loss models and helped the bank to implement a more comprehensive scenario analysis process, including more severe scenarios and comprehensive data inputs. We also worked with them to develop a more rigorous model validation process, including back-testing and sensitivity analysis.
​
Find out more
Our work demonstrated the value of engaging a third-party consultancy to conduct stress tests on credit loss models. Our expertise and guidance helped the bank to identify weaknesses in its existing models and develop new models that were more robust and accurate.
Whether you’re migrating, implementing or enhancing your models, we can support you by developing tools and strategies to provide insights and controls.
​