Family reunions inevitably miss a few relatives, like when nobody remembered to invite Billy's wife's cousin's nephew's mother-in law, Aunt Edna. While a missed invitation to the reunion may yield a few hurt feelings, in data departments on Wall Street, a forgotten link can cost a firm millions of dollars.
Any firm that held investments in subsidiaries of energy-giant Enron learned that valuable lesson the hard way, when the company's 2001 crash had chief financial officers with investments in the energy trader clutching their purses in fear. Getting rid of any Enron investment was just the beginning -divestment from any Enron subsidiary or holding was a necessity, too. However, without an organized data hierarchy of Enron's "relatives," the connections among these companies were not immediately clear.
"If you want to figure your overall exposure to one part of an institution, you have to understand all of its holdings and subsidiaries, and map that back to one profile," says Michael Haney, a senior analyst at Celent Communications.
While the Enron case brought credit risk to the front page of every newspaper, that singular collapse was not the only driver for enhanced credit examinations at firms. "The securities industry went through a huge wave of default," adds Haney. "Credit risk management has hit trading desks and spilled over everywhere in the firm. That itself is going to force you to improve your data."
Compliance measures have propelled interest in connecting the data dots for credit-risk management, says Haney, with Basel II, the Patriot Act and know-your-customer regulations looming overhead. Mergers and acquisitions have only made the task more daunting.
Barclays Capital is one firm that has done its data-mapping homework. In an initiative that began eight years ago, Barclays' determined the importance of managing its data to evaluate its credit risk. "Senior management had a requirement for more timely and accurate reporting against credit crises. We had no quick way of being able to gather the exposures in a timely manner," explains Wendy Brierley, head of global reference data at the firm. "We wanted to set up a relational identifier, so that before anything could be traded, the client had to be identified."
The firm's first challenge was getting 37 relevant systems, all taking real-time updates, to fall in line with the initiative. Brierley says that all existing users were mandated to integrate with the centralized database within two years. In addition, any newly developed or purchased system would be linked to the framework, she says.
Hierarchy building for credit exposure begins with linking clients, Brierley continues. The firm first examines each new client against each database to see if that client already has relatives in the Barclays system. If that's the case, the new client is fitted into the group hierarchy of its relatives. Otherwise, a stand-alone record is set up.
Having gained relative control over client data in a three-year period, the firm's next step was to create a central instrument static database, which would contain linkages of all securities to their issuers and, in turn, link to the client database, Brierley says. However, the second phase of the project proved more difficult. "On the instrument side, we had a lot of problems in our overall flow, collection and cleansing of the data," she says.
Brierley notes that central management of the project - with 14 team members working to support the New York and London offices, and a few more in Asia for other global locations - has been both a blessing and curse. While the central team ensures a holistic view, localized business lines have expressed reluctance to give up ownership of their data, she says. "It can be hard when you have ... a local team that was using and operating all of their own data. We can support them in the way that they require, but it is a cultural switch."
Another hurdle Barclays faces, one which has been widely echoed throughout the industry, is unique instrument identification. If one instrument is identified in two different ways, its connection to its relatives may be missed.
While using one identification standard would be ideal, Brierley explains that, in order to get a complete credit view on the instrument side, it is necessary for the firm to take in as much information as possible, regardless of the identifier in which the information is made available. "Our life would be easier if every instrument had an ISIN [a local identifier with a prefixed country code] and every client had some unique identification," she adds.
While industry working groups have put significant manpower behind creating unique-instrument identification, legal-entity identification, a potential code to identify each unique legal entity of a corporation, has been more challenging, says Tim Lind, a senior analyst at TowerGroup. Bank Identification Codes (BIC), assigned by SWIFT (Society for Worldwide Interbank Financial Telecommunication) upon request, have been examined to solve this dilemma but, Lind says, "There is no one BIC code for every legal partnership." While the structure of the code may be adequate, he continues, the business rules are not.
"There need to be very clear rules when you assign unique identifiers, especially with mergers and acquisitions," he continues. "The entities that could tackle this [challenge] need to be driven by a defined business model, and there must be business incentive beyond making the world a better place. Someone needs to get paid before it will get solved."