Scientific Beta

The February 2016 issue of the P&I EDHEC-Risk Institute Research for Institutional Money Management supplement contains several articles on the subject of smart factor indexing, looking at the long-term performance and live track records of smart factor indexes, the performance of different approaches to Value factor indexes, examining whether smart beta strategies are vulnerable to "crowding", and addressing the question of proprietary equity risk factor definitions and their deviation from the academic consensus on factor definitions.

Several articles in this supplement examine the area of smart factor indexing. We first look at the performance of smart factor indexes that are constructed based on combining a stock selection that targets a factor tilt with a diversified weighting scheme known as Diversified Multi-Strategy. We focus on assessments which take into account long-term evidence and on the performance observed after the commercial launch of the index. Every smart factor index outperforms the corresponding concentrated cap-weighted factor index over both the long term and the live period, providing very strong evidence of the robust benefits of diversification.

While there is a consensus on the existence of the value factor and the fact that it is rewarded over the long term, the implementation of value indexes, notably in the long-only universe, is not subject to the same consensus. Index construction mechanisms and various proprietary variable definitions and algorithms affect the return and risk properties of the resulting indexes and are different from provider to provider. Focusing solely on maximizing the value exposure may lead to concentration, which will result in greater unrewarded risk and wrong tilts to other rewarded risk factors, thus compromising the overall performance of the indexes. Investors should therefore not only prioritize selection of the right factor tilt but should also perform due diligence in comparing the different index providers and their offerings for the desired factor tilt in order to obtain the right factor tilt in an efficient way with robust performance.

Some argue that smart beta strategies are vulnerable to “crowding,” with increasing popularity posing a risk of overpricing and lower future returns. We find no evidence of this, but if one is concerned about potential crowding, the immediate concern should be to 1) hold well-diversified rather than concentrated strategies, and 2) spread out over many different strategies. Such an approach of avoiding concentration and diversifying across strategies is easy to implement with smart beta indexes.

We address the question of proprietary equity risk factor definitions and their deviation from the academic consensus on factor definitions. While providers refer to indexes resulting from such proprietary factor definitions as “enhanced” or “prime” factor indexes, one wonders whether such indexes might not also lead to an increased risk of data-snooping. Such data-snooping risk could mean that “enhanced” back-tested performance may not be repeatable out-of-sample.