A new study jointly published by Scientific Beta and Scientific Analytics entitled “Inconsistent Factor Indices: What are the Risks of Index Changes?” highlights the dangers of making frequent changes to index methodologies.
The paper
- provides concrete examples from industry practices and analyses the possible implications of such inconsistencies, namely changes in factor definitions, factor selection and portfolio construction principles
- shows that methodological changes are quite common in the industry and sometimes happen across multiple dimensions at the same time, which can lead to striking differences in the performance of multi-factor indices.
The main problems with inconsistencies across time are data-mining risks and therefore the absence of any robustness and reliability of the performance presented. This research provides illustrations based on the practices of popular index providers such as FTSE Russell, MSCI and RAFI.
Commenting on the study, Professor Nol Amenc, CEO of Scientific Beta, said, “Providers can put stringent requirements on index changes by remaining consistent with their investment principles. Indeed, if the urge to innovate means deviating from investment principles that were not really justified by serious research and corresponded more to approaches driven by in-sample back-testing, there is a risk that index investors will be disappointed with results. Maintaining investment discipline by adhering to a set of long-term principles may be the best safeguard against negative surprises with factor indices.”
The white paper can be accessed through the link below:
