This article was published in 2011 in response to the growing popularity of passive equity investing strategies claiming to offer higher risk-adjusted performance than traditional market-cap-weighted indices. Each strategy offered a different explanation for its ability to deliver better returns than the market, prompting the authors to replicate the strategies of the product providers in order to study why the products were able to deliver superior performance. They apply standard factor return decomposition found in the academic literature with the result that value and size are responsible for most of the outperformance. The article was written before the term “smart beta” was popularized. Thus, the authors use the term “alternative beta,” current usage at that time.
Chow et al. discover that the value and size loadings should not be surprising: the products use weights unrelated to prices, which results in value and small-cap exposures when compared to the capitalization-weighted market index. Because the driver of the superior returns is the same for this group of products, the expected return for all the products should be similar. Absent a difference in performance, investors should use variables related to execution cost—product fees, trading costs, capacity, and so forth—to select a strategy.
The methodology of the analysis involves generating a U.S. portfolio and a developed global portfolio for each strategy and backtesting them over the periods 1964–2009 and 1987–2009, respectively. Date ranges are determined by the availability of data for each portfolio.
Strategies examined are classified as one of two types of weighting schemes: 1) heuristic-based or 2) optimization-based. The heuristic-based weighting methodologies are equal weighting, risk-cluster equal weighting (RCEW), diversity weighting, and fundamental weighting. The last three strategies attempt to improve on a naïve equal weighting. The optimization-based weighting methodologies are minimum variance and two Sharpe ratio-maximizing strategies—maximum diversification and efficient indexing.
All of the strategies studied generate meaningfully higher returns than their cap-weighted benchmarks. The authors state this may be explained by selection bias as all of the strategies have achieved commercial or publication success. None of the strategies exhibits statistically significant Carhart (1997) four-factor alpha.1 Nevertheless, the alternative betas are valuable to investors by offering a cost-efficient, long-only means to capture the size and value premium.
The authors’ analysis shows that both heuristic-based and optimization-based strategies are naturally biased toward value and size. In terms of the heuristic-based betas, this observation is consistent with the findings of Arnott and Hsu (2008) and Arnott et al. (2010): any portfolio that rebalances on non-price-based weights incurs a positive value load. The diversity-weighted, RCEW, and risk-efficient strategies also show high small-cap loading because they are derived from equal weighting, which systematically overweights smaller stocks relative to a cap-weighting approach. The optimization-based strategies typically have a portfolio beta significantly below 1, favoring low-beta stocks, which often display both value and small-size tilts.
The alternative beta strategies improve on existing value and small-cap indices to the extent they are able to raise the investor’s Sharpe ratio and information ratio. Not only do traditional value and small-cap indices exhibit negative Fama–French alphas, indicating they may be a suboptimal means for achieving these tilts, but Fama–French factor portfolios require shorting, have high turnover associated with rebalancing, and hold many illiquid stocks (Hsu, Kalesnik, and Surti ). The authors also observe that optimized portfolios do not appear to produce a meaningful improvement in diversification over nonoptimized portfolios. Likewise, the investment insights associated with a stock’s expected return and volatility or downside semi-volatility, as provided by MVO, do not appear to generate a unique performance benefit compared to those provided by a simple heuristic-based approach such as equal weighting.
As a result of their analysis, the authors conclude that investors should be diligent in assessing the potential implementation costs of alternative beta strategies because much of the anticipated performance advantage can be dissipated through the implementation required for more complex indexing strategies. By combining less costly beta strategies, such as diversity weighting and fundamental weighting, both of which have two to three times the average market capitalization of the other smart beta strategies, investors can more efficiently gain the same factor exposures.
Summarized by Kay Jaitly, CFA.