More than one-third of banks are failing to tackle data quality issues when it comes to stress tests.
A total 35% of banks find it difficult to manage data requirements and aggregate data, found a SAS-sponsored study of more than 100 senior banking officials across Europe and the US.
That included the quality of information they collected during stress tests, introduced by the European Banking Authority and the US Federal Reserve following the 2008-09 financial crisis to try and prevent a repeat.
Brooke Upton, SAS’s senior product marketing manager for risk, said: "Banks are really struggling with data quality and a lot of that has to do with aggregating the data, and combining that data to present it in an accurate and reliable format to produce good results."
More than a quarter told SAS they had only a basic stress testing framework, but 34% said theirs were well-developed, yet still many found the process challenging.
The second-most cited issue was finding the right skills for stress testing, despite Citigroup bolstering its regulatory and compliance teams by 30,000 staff this year and Deutsche Bank doubling its compliance budget and adding 500 resources including staff.
Upton said this resulted in a lack of data sharing across different teams within banks.
"Banks are having a difficult time acquiring and retaining human capital and resources," she said. "They don’t have enough skills and expertise in the regulatory space. There’s a lot of manual intervention to actually collaborate the data across lines of business, there’s a lot of manual intervention and reliance on spreadsheets."
She added banks must make automating data sharing a priority, before using stress test data to inform business decisions, rather than seeing it simply as something they have to do – which 41% of respondents were guilty of.
Upton said: "They’re using it primarily as a compliance tool and 44%, which is alarming, do not involve their executive management at all. It’s critical analytical tools are used to help banks along the maturity curve."
