Challenge: Normalization of ratings data across vendor API’s
To satisfy the need for wider rating coverage and more frequent updates, our team was asked to replace an existing rating component with a stand alone tool.
After initial investigation, it turned out that each rating agency exposes data via a different protocol, and in a different format. Additionally they differ in the frequency of data releases, and their handling of the data updates. Some publish full data universe on a monthly basis with daily updates, and others publish full data universe on each major update. To cover all corner cases, we familiarized ourselves with data dictionaries and API specifications provided by each agency.
The lack of a standardized API shared between the data providers required a very flexible metadata driven tool, which can be easily tuned per provider, and can ensure that all the data will be stored in a consistent way.
Our tool allows users to easily configure the connection protocol, data normalization algorithm, and the way of recognizing unrated securities. To address the inconsistencies in the data release frequency, we implemented various data update strategies (e.g. daily patch or monthly full universe update), that could be easily combined via the configuration file.
The single standalone tool let the client set up a robust ratings retrieval process that handles all the rating data in a consistent way, and audits it in order to comply with government regulations, regardless of the rating provider.