Scientific Beta

This paper analyses a set of characteristics-based indices that have recently been launched on the US market said to outperform standard market cap-weighted indices over particular backtest samples, and shows that the outperformance over value-weighted indices may be negative over long time periods and that characteristics-based indices do not significantly outperform simple equal-weighted indices.


This paper analyses a set of characteristics-based indices that have recently been launched on the US market and have been said to outperform standard market cap-weighted indices over particular backtest samples. It analyses the performance of an exhaustive list of such indices and shows that:  

  • The outperformance over value-weighted indices may be negative over long time periods and that
  • Characteristics-based indices do not significantly outperform simple equal-weighted indices.

Furthermore, an analysis of both the style exposures and the sector exposures of characteristics-based indices reveals a significant value tilt. When properly adjusting for this tilt, these indices do not show any abnormal performance.

Therefore, the paper argues that the main value added of these indices may be to provide investors with a liquid, systematic, and relatively cheap alternative to other value-tilted strategies. However, it should also be noted that if one recognises the potential to tilt exposures to sector or style factors, it is straightforward to construct factor portfolios that beat the characteristics-based indices in the sense of mean-variance efficiency.

A revisited version of this paper was published in the March 2009 issue of European Financial Management.