Skip to main content
Alfa-Bank's Automated Model Monitoring Improves Management

Automated Model Monitoring Improves Capital Utilization

Alfa-Bank, the largest universal privately-owned bank in Russia, has used FICO® Decision Central to automate monitoring of its predictive models, in order to meet requirements from the Central Bank of Russia.

The project has reduced report preparation time by more than 90 percent, created complete transparency into model performance and recommendations for improvement, and will help the bank qualify for an internal ratings-based (IRB) approach to Basel II, which should reduce the capital the bank needs to set aside on loans.

You can read more about this story in the full media release.

“Alfa-Bank is using more and more predictive models based on different data types,” said Vera Perevitskaya, head of the validation at Alfa-Bank. “Because of this, we implemented FICO Decision Central to make the monitoring of models more transparent and comfortable. We can now prepare the reports nearly two times faster and are fully compliant with Russian Central Bank regulation.”

Alfa-Bank also worked with FICO to develop a wide range of custom statistical and qualitative tests covering different types of models, including PD, EAD and LG, for retail and corporate business models, and special tests for low-default portfolios.

For its achievements, Alfa-Bank has won the 2021 FICO® Decisions Award for regulatory compliance.

Alfa-Bank has provided an excellent example of how model monitoring should be done. This kind of work is absolutely critical, not just to meet regulations but to ensure the quality of the models that drive a bank’s lending decisions. As an economy changes — such as in the current pandemic — continuous validation of model performance is even more important.

“Model maintenance or the model lifecycle is a very hot area in the world of analytics,” said David Dittmann, vice president, data & analytics, P&G and one of the FICO Decisions Awards judges. “I was impressed that Alfa Bank was pushing forward with new ways to think and manage the model lifecycle that addressed the top risk and compliance challenges.”

Automated Model Monitoring

 

related posts