August 2015 -- sponsored by Thomson Reuters
Definitions and Organization
An insight one can derive from the responses to questions posed for this special report on risk data aggregation is that much of how well firms do with pulling essential data together for better risk management depends on how they define the data, divide or group that data, and examine the data.
Deloitte's Dilip Krishna, in our Q&A, says that raw data input must have "high fidelity" to produce high levels of risk data quality. He notes that risk data usually comes from other parts of a firm, in the form of booked trades or loans being originated or serviced, and therefore ends up getting enriched with risk metrics. Those metrics include client, facility and collateral data, often from a historical record of five years or more.
In our Virtual Roundtable, data management executive Rick Aiere stresses the importance of a common vocabulary for understanding data coming from all different units of an organization - as Krishna describes. Firms have to organize themselves internally so their units collaborate to create and maintain the necessary data dictionary.
Thomson Reuters' Kate Toumazi suggests that it may be possible to choose a "best in breed" dictionary, if multiple dictionaries are already in use throughout a firm. Ideally, this would make it possible to harmonize a broader array of data into a single, more scalable model, she says.
With enterprise-wide data, firms must break it down, scrutinize it and then reorganize it to address risk management, adds Aiere. A single repository, however, is not necessarily the only way to go, says Toumazi. A federated model can be just as effective at preserving an enterprise-wide view. Everyone has a role to play in defining and managing risk data.