Kevin Sexton, vice president and division manager within the Key Technology Services division of KeyBank, an $84 billion bank based in Cleveland thats been working to ready the banks data collection systems for Basel II compliance since the late 90s, said theres still "a significant amount of work" to do in terms of pulling data together. "Were well on the way and further on the way than many of our peers and competitors," he said, with significant work remaining to be done in terms of data analytics and the creation of risk models. Regulators have told KeyBank to prepare to have from five to seven years worth of data history in each of the banks portfolios, according to Bob Kula, director of economic capital. "In some cases, if you only have three years of data, you need to continue to expand, and not throw historical information away," he said. "Over time, by the time [Basel II] takes effect, you can get to the historical data requirements Basel is expecting." Work on Basel II at KeyBank has included building an enterprise data warehouse with a downstream-dependent data martin other words, the data warehouse passes information on to the data mart, as opposed to the data mart collecting its own information, applying its own business rules and populating the data warehouse. KeyBank uses an IBM DB2 database on an OS/390 platform for the data warehouse. The bank has various data marts: some on the same OS/390 DB2 setup, while some are Oracle Corp. data marts. The bank uses analytics tools from SAS Institute Inc., Business Objects S.A. and MicroStrategy Inc.Working on Basel II has not been without benefit. One example of data-improvement opportunities that have popped up at KeyBank is a single view of customers, Sexton said. "Before we began all this, there were multiple, authoritative views of the customer, quote unquote," he said. As a matter of fact, KeyBank isnt even technically required to comply, since Basel II is mandatory only for large banks that have assets with international presence. The bank finds the benefits to be compelling enough to comply, however. "To not be a part of it would be a mistake," Kula said. "Wed be at a disadvantage [in terms of competition] not to be moving along this path."
The downstream data mart is significant in that it allows for consistency of information, Kula said. "We dont all walk into a room with 10 reports that should say the same thing and dont," he said.