As financial systems become more complex, the need has arisen for more proactive, robust and comprehensive data to allow regulators to implement effective supervision. The Central Bank of the Philippines (BSP) supervises 571 banks, large and small, and an array of other financial institutions. Its use of a prototype regtech system has helped it overcome the drawbacks of its earlier processes.
In 2017, the BSP required banks to submit, at various frequencies, 240 Excel-based reporting templates with more than 100,000 data points. These templates were manually populated, took significant time to complete and were subject to human error. Furthermore, the templates were emailed, creating further security concerns. While an automatic validation process was in place, staff at the BSP and the institutions it regulates were still required to cross-validate and reconcile data.
As a result, the BSP was unable to devote time to analytics or generate information that could be used for a more prompt regulatory response, hampering policy-making. The BSP developed a prototype regtech solution – a prudential reporting system based on an application programming interface (API) that streamlines the transmission, processing, warehousing and analysis of a regulated entity’s prudential reports.
“The API provides a machine-to-machine link between the banks’ core banking systems and BSP’s system, eliminating human intervention and human error,” says Vincent de Villa, director of the fintech subsector. The BSP tested the system on two pilot banks, with the results showing great success.
Data no longer has to be manually populated into the numerous templates. Instead, data is now be extracted directly from the banks’ databases, converted into a single XML file and fed directly into the BSP’s system for processing. “Through the API-based architecture, a larger amount of data can be transferred for processing in a shorter amount of time, resulting in more robust and granular data for analysis,” says De Villa.
Email transfer has been replaced by data transfer, which the BSP has ensured uses international standards of security and encryption. The regulated entity now also receives instant validation on receipt of the file submission. Feedback is now transmitted via the API in case of validation errors, with the previous average processing time of 1,800 seconds being cut to just 10.
The number of data points has also been reduced by 53% – from 107,000 to 50,000 – which has allowed the layered reporting packages to be converted into a single reporting package. “After processing, this streamlined dataset is automatically placed in a centralised and secure data warehouse where we are able to extract data using pre-scripted SQL queries without requiring further intervention,” says De Villa. Analytical tools, dashboards and statistical software can be applied by various departments across the bank who need to use the data.
Scaling up this framework has now become a priority for the BSP, which is in the process of pilot testing with 18 other institutions.