So-called smart-beta equity indexes have proliferated in the last several years. Products now number in the hundreds, with most seeking to capture the excess returns of a handful of well-defined risk factors such as size, value, low volatility and dividend yield. The popularity of these new indexes is due in part to this relative simplicity.
But as more investors shift from fundamental active or traditional passive equity strategies to these alternative indexes, will the results match the label on the package? It turns out that smart-beta indexes often produce markedly different performance results, even among those targeting the same factor. High-dividend indexes, for example, can produce a huge range of risk-adjusted returns: Over the past ten years, the Dow Jones U.S. Select Dividend Index achieved a Sharpe ratio (return per unit of risk) of just 0.37 while the S&P 500 Dividend Aristocrats index yielded a Sharpe ratio of 0.67, a difference of 80 percent in risk-adjusted returns. Our research shows this is not an isolated phenomenon.
So what explains the disparity in performance? In our view, the success of a smart-beta index depends on its ability not only to capture exposure to targeted risk factors but also to minimize unintended, uncompensated factors such as industry concentration, currencies and leverage. These by-products of a risk factor investment strategy can contribute meaningfully to risk without producing excess returns.
To more accurately measure intended risk factor exposure, Northern Trust Asset Management created a metric called the factor efficiency ratio (FER). Conceptually very simple, FER is the ratio of an index’s intended factor exposures to its unintended exposures. To achieve a high FER, an index must have a strong tilt toward the compensated risk factor or factors (high numerator) while minimizing unintended factor exposure (low denominator). Factor exposures are calculated using established risk models, such as Barra or Axioma. Technical details on the process can be found on Northern Trust’s website or in Michael Hunstad and Jordan Dekhayser’s paper on the Social Science Research Network.
By applying risk models to index returns and composition, we determined the intended exposure and unintended exposures for some of the most widely used smart-beta indexes and calculated the FER for each of them. Our research demonstrates that indexes that are more efficient achieve more intended factor exposure per unit of unintended exposure. Further, by measuring the FER of indexes targeting small size, value, low volatility and dividend yield and relating these metrics to historical Sharpe ratios, we can show that high efficiency also leads to high risk-adjusted returns.
Unfortunately, what we also found was that FERs were strikingly low in most smart-beta indexes analyzed. In other words, existing indexes were generally unable to provide desired factor exposures without taking on substantial unintended exposures. We feel this problem stems from the relative simplicity of smart-beta index construction. For most indexes there exists no mechanism to control unintended bets. The result is at best an inefficient allocation of assets, and potentially a significant risk to investors.
One lesson to be drawn from our research is caveat emptor — make sure you understand the true risk profile of indexes, and don’t put too much credence on index names and descriptions. On the positive side, we believe that with proper design and implementation, risk factor investing can live up to the expectations of many investors and consistently deliver risk premiums while minimizing uncompensated factors. More on that in a future post.
Matthew Peron is managing director of global equity and Michael Hunstad is director of quantitative research at Northern Trust Asset Management in Chicago.
See Northern Trust Asset Management’s disclaimer.
Get more on indexing.
