This is the first post in a series discussing some of the problems associated with investing in “smart beta” strategies.
Everyone’s seen the quilt chart used to show the return dispersion between asset classes, but it works equally well for factors.
Factors have just as much dispersion as asset classes, and can have wildly different performances year-to-year – in 2009, there was an over 30% difference between the MSCI World momentum factor and the small cap factor.
Research from S&P Dow Jones Indices has also shown how their single-factor indices have had very different returns since their inception:
Source: S&P Dow Jones Indices
Someone tracking S&P’s low volatility index would’ve had markedly different returns to someone tracking their quality index, who in turn would’ve had very different returns to someone tracking their momentum index – particularly in 2000.
Further evidence from Research Affiliates has found that not only do factors provide very different returns, but a number of factors have a tendency to suffer from large crashes. They noted that “The two factors with the most negative skewness [a measure of the frequency/magnitude of crashes] are momentum and illiquidity. These results are not surprising: momentum is known to be prone to crashes, and the prices of illiquid companies tend to plummet when liquidity dries up”.
To make life even more difficult for single factor investors, performance varies not only between factors, but between geographies too. The chart from Credit Suisse below shows how factors have performed in the US vs the in the UK since the financial crisis. It’s interesting to see how there hasn’t been a single year where the ranked factor performances have been the same between countries:
Source: Credit Suisse
The problem with investing in a single factor is that it will underperform. Sometimes by a lot, and sometimes for a long time. Not only will the factor underperform the other factors the investor could have chosen, it’ll also underperform the market.
If we take the value factor in the US as an example, value has had two periods where it’s underperformed the S&P 500 over multiple decades:
Source: Alpha Architect
The problem with prolonged underperformance is that the longer a fund underperforms, the more likely it becomes that it’ll be sold, and investors will move on to something else. Research from State Street Global Advisors has found that only 1% of smart beta managers and 11% of active managers would tolerate 3 years of underperformance before seeking a replacement:
Source: State Street Global Advisors
Ditching underperforming funds sounds intuitive and sensible, but investors often sell underperforming funds at exactly the worst time, and re-allocate that cash to funds that’ve had strong recent outperformance. This is known as performance chasing.
I’ll be dedicating posts in the future on how performance chasing destroys returns (spoiler: it’s because of mean reversion), but one study specifically related to performance chasing in factors found that “fund flows have been driven by factor funds earning high past returns and not by the funds providing factor exposures” and that “we [the authors] do not observe a positive relationship between fund flows and future performance”.
I.e. investors are showing signs of performance chasing within factor funds, and are hurting their returns as a result.
Because factors behave differently to each other, one of the problems with investing in smart beta funds is that there will be (sometimes long) periods where the fund underperforms. Despite most people investing with the best intentions, sticking with an underperforming fund is incredibly difficult – whether you’re a retail investor or a seasoned asset allocator. The underperformance will likely lead to performance chasing, and will ultimately lead to lower returns.
One way in which investors can try to combat the effects of performance chasing within factors is by allocating to several different factors simultaneously through a multifactor strategy. Given the varying performances of different factors, it would seem to make sense not only from a performance standpoint, but also from a behavioural one, to diversify both between factors and between geographies. I’ve written about the benefits of multifactor strategies here, and a few of the drawbacks here.
The next post in the series focuses on why it’s so difficult to define factors, and how that affects the performance of smart beta strategies.