The process of attribution can be broken down into a number of questions that need to be solved; with which approach to best serve the business, what will be the data requirements and costs, will I deliver full implementation, and are the answers accurate and insightful?
Most traditional attribution engines seek to attribute changes in yield for each security, but we know through market experience that this is not the case.
With this model, the provision of analytics by a software service provider, as well as portfolio data provision by the client, is required, creating restrictions around the software service provider’s prescriptive performance model.
This results in a range of issues, including:
- Assumption of singular price
- Analytics not as expected by client
- Prescriptive performance model
- Accuracy impact – major differences against official return (residuals)
- Limited to buy and hold
- Unable to extend to accommodate greater complexity (derivatives)
- Limited Middle Office use.