Market Data Repository
An international bank commissioned Sinara to conduct a review into its market data usage centred around an internal market data database and its related processes. In particular, we needed to look at the feasibility of extending the use of this database to become the bank’s central market data repository. In parallel, the bank was also in the process of changing their main market data vendor.
The market data database had been developed internally to receive and store data from various vendors. Automatic calculations were run periodically, often via spreadsheets, and the original and derived market data distributed to internal systems, most notably for pricing structured trades.
Sinara carried out an analysis of the existing market data database showing that although it satisfied its trading related objectives very well, it was overly complex. As a consequence it lacked resilience, provided limited monitoring facilities, and made it difficult to evaluate data quality.
Sinara’s analysis also identified a number of market data pathways within the bank which by-passed the repository completely – often involving desktop spreadsheets. Centralised knowledge of what market data was flowing, where and for what purpose, was not readily available to management. Technical knowledge was also dispersed, making changes to the system problematic.
One clear outcome from Sinara’s analysis was that the bank needed clearer governance over their market data usage, so as to define ownership, eliminate duplication and better assess data quality.
Sinara’s review also made a number of recommendations for the future development of the repository:
- the existing database, despite its restrictions, should form the basis of a more cohesive system so as not to lose the benefits of its complex pricing functions
- various technical changes would be needed to improve security, resilience and allow better operational monitoring
- how additional market data could be incorporated into the repository
- plans for how a more flexible service-oriented approach could be used for accessing data.