Kesu - Fotolia
The failure of investment banking firm Lehman Brothers that helped precipitate the global financial crisis in 2008 showed, among other things, that vital data was lacking when crucial business decisions needed to be made by other organizations with possible financial ties to Lehman. "It was hard for companies to access their risk exposure," recalled David Saul, senior vice president and chief scientist at Boston-based State Street Corp., which did have operations that were affected by the Lehman meltdown.
Measuring the potential risks of a Lehman failure was just as hard for federal treasury officials, partly due to the lack of clearly identified connecting lines between Lehman's many distinct but intertwined global business entities and other financial services firms. That meant the feds also had to guess, Saul told attendees at a meeting of The Data Warehouse Institute's Boston chapter held in Waltham, Mass., earlier this month. "In hindsight, they guessed wrong," he said, adding that wider use of semantic data, including an emerging legal entity identifier (LEI) reference code, could help reduce such guesswork and enable organizations to better identify and mitigate business risks.
Probably the most notable champion of semantic data identification has been Tim Berners-Lee, inventor of the Web and now director of the World Wide Web Consortium. As described by Saul, Berners-Lee pushed forward the idea that an ontology of well-defined, universally-agreed-to identifiers could stand to represent company names and other objects that might appear in a wide variety of data formats.
Hooked on semantics
Saul and others propose that this semantic approach be applied to different technologies, working like "metadata on steroids," as he put it. For example, he said that the Enterprise Data Management (EDM) Council, the Object Management Group (OMG) and other organizations are working to make standard LEIs a reality.
Some of that LEI work has occurred as part of a joint initiative between the OMG and the EDM Council to develop a Financial Industry Business Ontology. LEIs, which provide unique alphanumeric codes tied to a company's name, address, corporate lineage and so on, are likely to move ahead more quickly than some other data semantics efforts because of regulatory requirements.
Participants in financial derivatives markets -- which enable organizations to trade specific financial risks to other companies, in sometimes dizzyingly complex transactions -- are required by the federal Dodd-Frank Wall Street Reform and Consumer Protection Act, signed into law in 2010, to obtain proper legal entity identifiers from a handful of approved sources.
Map data instead of moving it
Saul said the use of LEIs can be looked at as an adjunct to a traditional data warehouse, allowing "ad hoc integration" without having to create a new database. "The LEI is a natural for semantic mapping," he remarked. "Within State Street, we're leaving the information in [existing] systems and mapping that against the LEIs. Now we can look at dissimilar data without having to move from system to system."
It would be very useful for such traceability to extend across organizations and financial systems around the globe, said OMG CEO Richard Soley.
"A major source of the problems that caused the 2008 financial meltdown was the lack of ability to track risk flows through contracts [between financial services firms]," Soley said. "Some of the risk relationships from the meltdown are still being tracked. The only way to simplify and automate that problem is to have standardized, worldwide identifiers for corporations so you can put those in contracts easily."
Saul's colleague David Blaszkowsky, senior vice president and head of data governance at State Street, has been pursuing the notion of the LEI for more than 10 years -- ever since his days leading XBRL-based data undertakings at the U.S. Securities and Exchange Commission. Blaszkowsky told attendees at the TDWI chapter meeting that financial services players that stand to benefit from taxonomies for complex securities need to take first steps of a semantic kind.
"What we are urging is that people discover their data in order to do data governance," he said. "You have to know where data is, what it means and where it is weak." But, he noted, easy-to-use tools for managing semantic data are still in their early days. LEIs will find wider use when the software to implement them is as simple as something like TurboTax, Blaszkowsky said, referring to the widely-used tax preparation package.
Read David Saul's take on the future of semantic databases
Find out how semantic technology works with big data