State Street has completed its proof of concept for FIBO—a conceptual ontology designed to define financial industry terms, definitions and synonyms. The proof was successful, but there is still a long way to go before its wider industry adoption.
The executives at State Street knew they had to find a way to better understand the sea of data that the investment bank produces on a daily basis. What’s becoming increasingly apparent—not just at State Street, but across the industry—is that the best way to understand your data is if it’s speaking a common language.
“We all have a problem when it comes to our data: understanding what it represents, the regulations that surround it and the various technologies that we use to support it,” said David Saul, chief scientist at State Street, during his keynote speech in the North American Trading Architecture Summit held in April. [Saul also spoke with Waters after the event for this story.]
The quest for a common industry language has been talked about and argued over for years, and more banks and regulators are looking to test out ontologies for tracking financial data, especially since the industry wants to ensure that interconnections between firms are transparent.
The Financial Industry Business Ontology (FIBO) is seen as one of the possible answers to State Street’s—and many other banks’—data problems. The conceptual ontology, a joint effort between Object Management Group (OMG) and the Enterprise Data Management (EDM) Council, is currently being developed and tested by a select group of industry participants. The tests—or proofs of concept—are to see if FIBO is viable and can be used on a larger scale.
“The goal is to rip it to shreds,” Michael Atkin, managing director of the EDM Council, tells Waters. “We are actively looking for something to be wrong with FIBO.”
“The approach to the proof of concept was that we wanted to take FIBO as is—we didn’t want anything that wasn’t already freely available in the open on the EDM Council website—and we wanted to match that up with some technology and combine that with our own data.” David Saul, State Street
State Street was one of the first companies curious about FIBO and was interested in running a proof of concept. It conducted its FIBO proof of concept on interest-rate swaps to see if the ontology could handle live data generated by the bank, a process that began earlier this year.
Participants in the study were State Street, Cambridge Semantics and consultancy Dun & Bradstreet. The EDM Council provided some counsel, while Wells Fargo, which conducted its own proof-of-concept tests on FIBO, gave feedback based on its experience working with the ontology. Cambridge Semantics is one of State Street’s service providers and allowed the bank to use its Anzo analytics platform for the project. Dun & Bradstreet was involved in the study as an independent source of corporate entity and hierarchal data.
“The approach to the proof of concept was that we wanted to take FIBO as is—we didn’t want anything that wasn’t already freely available in the open on the EDM Council website—and we wanted to match that up with some technology and combine that with our own data,” Saul said.
Since the project was just a proof of concept, State Street did not connect FIBO to a production system, though the bank used data extracts. The scope was purposely kept small, though Saul noted that interest-rate swaps were the majority of transactions it handled. It matched FIBO data with its transactional system, FIS’ Front Arena platform.
Saul said the proof of concept had four steps: First, they loaded the FIBO model into the Anzo tool—the application that stores the FIBO model and matches it to State Street’s data—to see if the semantic data could be read. Then they took the extracted data and mapped it against FIBO. Next, they generated the graphical representation of transactions. Finally, they created analytics based on the data.
Loading the FIBO model into Anzo was a very simple, straight-forward process, because FIBO is represented in a standardized language, according to Saul. While the bank used Cambridge Semantics’ Anzo tool, Saul said he was not advocating one vendor over another, as FIBO could potentially work with other products.
With the model loaded into Anzo and each element mapped in Front Arena, State Street was now free to create a graphical representation of the data. Saul said it was an automated function to see what corporate entities the data was pointing to, so specialized skills were not needed to monitor that stage. This was supposed to show if there were any data elements that were present in the transaction but that had no intrinsic value to the final product; State Street’s map did not have any dangling data.
With all the previous steps finished, State Street was in a position where it could now start answering questions using the mapped data. The bank was able to generate reports without the need to build a data warehouse, Saul said. It also saw where there were concentrations of collateral in a particular entity or geography.
While the proof of concept largely went off without a hitch, there were a few hiccups. For example, there were some IT problems, mainly relating to connectivity issues, when the bank first began to load the FIBO model. But this wasn’t an indictment of FIBO—Saul explained that these were more traditional IT issues and had little to do with FIBO itself. Since the data used for the proof of concept was managed by State Street’s business operations group, working around their schedule also affected the project. Any potential problem during the proof of concept was largely unrelated to the technology that it was testing, Saul insisted.
There was one point during the process, though, where State Street had to call on Wells Fargo’s earlier FIBO experience. “We had one data element that we had not matched up. Our first reaction was, is the FIBO model incomplete?” Saul said. The firm then got in touch with Wells Fargo, which informed them that they were “looking in the wrong place in FIBO.” At the completion of the proof of concept, State Street found FIBO was complete for use with interest-rate swaps. “The good news is that it works,” Saul said.
Where Will FIBO Go?
This was, of course, not the first time FIBO has been tested. Other members of the EDM Council have mapped FIBO to other data and financial instruments. FIBO has been in development since 2011—in cooperation with the Object Management Group (OMG) standards consortium—and since then there has been a lot of developments in the program. Atkin, of the EDM Council, said in a January 21 webinar, that the Council is working on FIBO to create practical applications like scenario-based analytics, business process automation, and securities processing.
The EDM Council created a timeline to guide the public where FIBO is in terms of its development. It also established what it calls a “FIBO vocabulary” to get firms to understand financial instruments. Another milestone was the creation of a system of record called the web ontology language, which is currently being polished.
The EDM Council hailed the ease with which State Street handled the proof of concept. “It works as advertised,” Atkin tells Waters. “Mapping was easy and data can be harmonized through a common language and can be connected to a third party.”
So FIBO works for State Street, but where does it go from here? Most of FIBO is intuitive, according to Saul, although he noted that “as it builds, it’s going to get more complex, so some education will be required.”
As FIBO evolves and more financial instruments are connected to it, it might become more difficult to understand. “There is some learning associated with it and I think there’s a real opportunity here in the professional services industry for consulting on FIBO,” Saul said.
Saul explains that as FIBO becomes more complicated, the EDM Council will look into providing consultants to help firms using the ontology, although Atkin adds that it is not the job of the Council to be a consultant, saying that it is the domain of the banks’ subject matter experts.
Atkin says some domains covered by FIBO are now in the final stages of development, but there will always be a need for enhancement as technology and the industry evolve. He says the Council is working on expanding the reach of FIBO, and big banks are responding. He confirms that some are asking the group to expedite the process, and are promising to provide resources like subject matter experts for further testing and validation of FIBO.
The Way Forward
Dun & Bradstreet, which provided data for corporate entities and hierarchies for State Street’s proof of concept, says FIBO is crucial for the industry, which is still recovering from the impact of the 2008 financial crisis. “FIBO is where the industry is going,” Michael Lubansky, leader for Dun & Bradstreet’s global alliances and partnerships strategy, tells Waters. “During the financial crisis, it was challenging to see the interconnections of counterparties, but with FIBO we can get a full view of all relationships.”
State Street’s project was the first FIBO proof of concept for the data firm, although Lubansky says interest in helping to create a common language for the industry is high. Atkin says he agrees, adding that “everyone is interested in the results of all these proofs of concept, even the regulators. We couldn’t be more proud of FIBO.”
Other member firms of the EDM Council are already working on their own FIBO proofs of concept for other financial instruments, according to Atkin. Some are mapping it to physical data models. The Council also presented FIBO during an industry conference in April where vendors exhibited their capabilities for visualization and modeling.
The EDM Council hopes widespread adoption of the fully fleshed out ontology will be used by its members, as well as the rest of the industry, soon. It is showing FIBO to regulators to generate more interest in the ontology. Atkin says the Council is even working on rolling out FIBO for the retail world, such as for loans.
But widespread adoption may have to wait for a tipping point where curiosity about FIBO from investment banks turns into actual use-cases, Lubansky says. It’s a process that will take some time. He also says that it will be important for the regulators to help push along the adoption curve.
For the EDM Council, it maintains that interest is high enough that the industry is moving toward a single, logical model in terms of how it operates a function that FIBO serves. Many banks are determining how to fully use FIBO on their own, following proofs of concept.
Saul says in an interview with Waters that "literally no day goes by that we don't get an inquiry about duplicating what we did, or do some variation of the FIBO proof of concept."
State Street is moving on from interest-rate swaps to see how it can operationalize FIBO for its day to day business. Saul says the bank may also look to use FIBO for other kinds of swaps and scale up the number of transactions FIBO can read. And because FIBO has already mapped the data it needs, State Street no longer needs to redo many of the steps it previously went through for the first proof of concept.
"We will probably do some version of both. We're going to meet with consumers to ask where the FIBO reports could help them," Saul says.
As State Street’s FIBO proof of concept shows, there is a possibility that FIBO will be widely used due to its mapping ease. But it is still in the process of being refined, dismantled and rebuilt. Soon enough, the industry might have that common language, after all, and banks like State Street will no longer struggle to understand its data.
- State Street’s FIBO interest-rate swaps proof of concept was successful enough that the bank will expand its tests to other asset classes and instruments.
- According to State Street and the EDM Council, FIBO is intuitive and easy to load into third-party software in order to fully map the relationships of data and transactions.
- Widespread adoption of FIBO might take time, however; it is contingent on interest from banks and regulators to expedite the ontology.
Bryan Cross, who heads UBS Asset Management's QED group, joins to discuss alternative data and AI.Subscribe to Weekly Wrap emails