Published: June 27, 2018
This paper outlines an approach to capturing information in the National Information Center database about bank holding companies and their subsidiaries, and their ownership and control relationships. We propose a detailed ontology that non-experts can use to analyze bank holding companies more effectively.
We consider the challenges and benefits of ontologies for information management for regulatory reporting from bank holding companies (BHCs). Many BHCs, especially the largest and most complex firms, have multiple federal supervisors who oversee a diverse array of subsidiaries. This creates a federated data management problem that disperses information across many firms and regulators. We prototype an ontology for the Federal Reserve’s public National Information Center (NIC) database. The NIC identifies all BHCs, their subsidiaries, and the ownership and control relationships among them. It is a basic official source on the structure of the industry. A formal ontology can capture this expert-curated knowledge in a coherent, structured format. This could assure data integrity and enable non-experts to more readily integrate and analyze data about complex organizations. We test the design and development of federated prototype ontologies in the Web Ontology Language (OWL) to provide and integrate the NIC data with precise semantics for transparency and consistency. Our preliminary results indicate that this is feasible in practice for data search and analysis, and that the ontologies can facilitate semantic integration and improve the integrity of data and metadata.
Keywords: Financial regulation; data integration; knowledge representation; ontologies; Web Ontology Language (OWL); data integrity; bank holding companies.