The success of smart beta ETFs has raised concerns over a possible ‘overcrowding’ of factor strategies. But a recent analysis of US equity ETF factor exposures, by Robeco’s David Blitz, suggests this is far from being the case.
Exchange traded funds (ETFs) have become an increasingly popular way for investors to gain exposure to equity markets at limited cost. Among the most dynamic segments of the ETF industry, in recent years, are those specifically designed to capture specific factors such as value, momentum or low risk.
Factor investing is based on the existence of various premiums that can be systematically harvested in order to achieve consistent higher risk-adjusted returns and better diversification than traditional market cap-weighted indexes. In recent decades, many academic studies have advocated its use and investors can now choose from a wide variety of factor-based products, in particular ETFs. These solutions, often branded as ‘smart beta’, follow indices specifically designed to give exposure to some of the most commonly exploited factors.
The commercial success of smart beta ETFs, however, has raised serious concerns over a possible ‘overcrowding’ of factor strategies and the fact that the academically-documented premiums might disappear. But a recent academic paper by David Blitz, Robeco’s Head of Quantitative Equity Research, focusing specifically on the US equity market, suggests these fears are exaggerated.
Indeed, while some US equity ETFs do show significant exposure to factors that successfully deliver premiums, this is not the case at aggregate level. Detailed analysis shows great disparities in factor exposures across the US equity ETF universe. When all exposure is added together, all that actually remains is plain market exposure, or beta.
In his study, Blitz analyzed the excess returns of all the ETFs that are listed in the US and that invest in US equities and that had at least a 36-month return history at the end of 2015. This amounted to 415 individual funds, with combined assets under management of over USD 1.2 trillion. That’s about 5% of the value of the entire US stock market. Data on the risk-free return, market excess return and the size, value and momentum factor returns was obtained from Professor Kenneth French’s online data library.
These findings refute the idea that factors are being arbitraged away by ETF investors’
Based on this broad sample, Blitz regressed ETF returns on the returns of various well-established factors, over the 2011-2015 period. He found that many funds offer a large positive exposure to factors, such as size, value, momentum and low volatility. As such, they could be considered suitable instruments for investors seeking to systematically harvest these premiums, except perhaps in the case of momentum.
At the same time, however, many other US equity ETFs showed a large negative exposure towards these factors. And in fact, on aggregate, the overall exposures towards the size, value, momentum and low volatility factors turned out to be very close to zero.
The increase in the popularity of low volatility strategies is a good illustration of this. At first sight, investors looking only at the billions of dollars invested in ETFs specifically targeting the low-volatility anomaly may rightfully be concerned about possible overcrowding. However, a closer look shows the funds in question only represent a small fraction of the total ETF market.
Moreover, when looking at the other end of the spectrum, it turns out that there are a similar number of ETFs which provide exactly the opposite factor exposure, with a significant bias towards high-volatility stocks. These ETFs are typically sector-focused funds. They are obviously not labelled as ‘high-volatility funds’ but they do effectively neutralize the exposure of the low-volatility ETFs.
These findings refute the idea that factor premiums are rapidly being arbitraged away by ETF investors, and also the related concern that factor strategies could be turning into overcrowded trades.