Finding Robust Smart Beta Solutions

February 25, 2015

[The following "ETF Issuer Perspective" is sponsored by ETF Securities]


Smart Beta: Factors & Weights

Prior to 2008, numerous hedge funds affirmed that their strategies were supposedly uncorrelated to the performance of the S&P 500. The financial crisis that followed brought a different story. The current evolution of smart-beta indexes seems to be a repeat of this period, where any nonmarket-cap-weighted strategy is called a "smart beta" index, which makes it quite challenging for investors to select a particular model.


Smart-beta indexes should directly address the historical limits of market-cap indexes. Based on this premise, investors should then be able to review any "smart beta" model by focusing on two questions:

  • What potential issues linked to market-cap-weighted indexes does the model try to address?
  • Are the solutions effectively designed to address those issues?


Distinguishing the potential issues with market-cap-weighted indexes

There are two very distinct criticisms of market-cap-weighted indexes:

  1. Factor tilt: Market-cap indexes naturally tilt exposure toward (i) growth stocks; and (ii) the largest market-cap stocks. However, in 1993, Fama and French showed that a portfolio tilted toward small-cap and value stocks would have had outperformed the market.
  2. Concentration: It is generally expected that most indexes seek to provide well-diversified exposure. Interestingly though, market-cap-weighted indexes can be considered as not properly designed to do so. For instance, the top 25 percent of holdings in the S&P 500 have generally represented more than 68 percent of the index performance. Academic research (Haugen and Baker, 1991; Cochrane 2001; and Goltz 2010) has shown that this potential overconcentration can lead to a suboptimal risk/return profile.


Two distinct issues and two distinct solutions

Addressing the factor tilt issue

Factors are determined by their ability to provide risk-adjusted outperformance. To determine how robust a factor model is, one should focus on the following two aspects:

  • Empirical evidence: Ideally, the performance analysis for a given factor should be carried over an extensive period of time
  • Economic rationale: A factor model should not be a data-mining exercise, and the selected factor should have a strong economic basis


Based on the current status of academic research, there is overall agreement on four key factors.



One should be conscious that each factor can be quite cyclical. Because of this, combining factors should help improve long-term risk-adjusted returns.


Find your next ETF

Reset All