Fixed Income Analytics Remediation

The Problem:

A large bond asset manager had assembled a piecemeal fixed income analytic system over the course of several years.  The lack of an overarching strategy, together with multiple acquisitions over the several years, left the client with multiple fixed income applications in a suboptimal state.  The client was suffering from three specific problems that were materially affecting multiple business units within the firm:

• EXCESSIVE COST – By paying licensing fees to access Yield Book, POINT and BondEdge/CMS, the firm was incurring expenses for redundant functionality.
• MISSING DATA  – Reports provided to trading, portfolio management and research contained a large amount of irrelevant data but lacked critical data needed on a daily basis.
• INADEQUATE SUPPORT PROTOCOL – When processes failed, resulting in missing or inadequate analytics, there was no defined protocol for manually correcting the data.  Historical time series analysis was often flawed and traders needed to make on-the-fly estimates on their portfolio duration levels to compensate for the erroneous data.

What We Did:

Adeptyx inventoried the firm’s analytic processes running on Yield Book, POINT and CMS.  After a thorough analysis, Adeptyx worked with key business users to document the specific analytic and risk calculations needed. Adeptyx created a roadmap to a final solution that improved the data being delivered to the business while improving efficiency by eliminating redundancy and consolidating processes.

 Adeptyx transferred existing CMS and Yield Book processes to POINT by building automated processes that would allow 100% of current functionality to be migrated onto POINT while offering additional features such as daily processing of analytics, highly customizable reporting templates, and reduced support costs.  To compensate for the certain universe of securities that POINT’s internal modeled group did not set up, Adeptyx leveraged the client’s existing security reference data for these bonds from internal systems (Bloomberg AIM) and utilized the automated variable-length User Defined Instrument (UDI) functionality currently available in production versions of POINT to model the securities on the fly.

 At a high level, the key project deliverables included:

• Development of an automated feed from client internal systems to POINT containing the securities needing to be processed for analytics and risk exposures
• Creation and scheduling a new overnight batch process on POINT to process, extract and calculate the analytics and scenarios (yield curve, prepayment and volatility) required by the business
• Facilitation of a one-time setup of existing missing securities on POINT
• Leveraging of security reference data from client internal systems to produce a UDI import file to automatically build any missing securities on POINT every evening
• Creation of multiple customized reporting templates on POINT to meet the requirements of each business constituency

The Results:

The benefits available to the client included:

• Substantial cost savings —  By consolidating the existing Yield Book and CMS processes into POINT, the client was able to reduce their Yield Book license count  and eliminate CMS entirely resulting in a total annual savings greater than the initial cost of the project. 
• Consistency of data — By updating the firm’s risk scenarios, exposures and reporting templates, Trading and Portfolio Management now received the same data used by Risk, Compliance, and several other key groups.  Reports contained consistent data and manually reconciling reports before they were presented to clients became a thing of the past.
• Improved decision support — With a new collection of support processes in place, critical areas of the firm were no longer stuck using old data when overnight processes encountered a problem.  By defining and implementing an “Early Action” group that began addressing problems as soon as they were systematically identified, the client was rarely in a position where essential data was unavailable for the business to take action.