Regulatory Data Harmonization: A NoETL Approach

fintech courses

By FINTECH Books Contributor, Steffen Krause
Follow: @surreyhobbit

Regulatory reporting in banking is congested, leaving little  room for innovative ideas. There are plenty of comprehensive software  packages that cover everything from valuations and risk functions to  stress testing, statistical and supervisory reporting. These systems rely on their own harmonized data pool so adopting them requires significant effort in the integration of source data via ETL. Once a new regulatory requirement appears, the data integration rarely covers everything and the process begins again.

Large organizations making long-term commitments to specific providers can benefit from this comprehensive approach but smaller banks and capital market firms struggle. A modular approach with last mile reporting such as offered by the Big Four are better for smaller budgets but do not solve the question of data preparation and harmonization before mapping to the regulatory data points, usually leaving spreadsheet solutions and manual intervention as the only option.

Graph virtualization offers an interesting alternative here.  Data can be combined by its business meaning using knowledge graphs (ontologies) which has potential to reduce integration cost enormously. This approach is also more adaptable to changed regulation because an ontology as the integration point does not pre-suppose the questions that will later be asked of the data.

The BIRD project by the ECB and member banks provides an integrated reporting dictionary and is aimed at the upcoming AnaCredit regulation of reporting credit risk information from financial institutions in granular form. It uses a common business language and interpretation rules that can easily be expressed as an ontology.

If you would like to learn more about FinTech please look at our video courses at