Are You Trying to Time Your Factor ETFs?
The ability to time or rotate between a set of investment strategies can help investment managers add value for clients seeking actively managed solutions. Traditionally, many investors have accomplished this through an investment process focused on rotating between sectors, countries, and broad investment styles.
Today, as more and more investors have continued to adopt smart beta ETFs that emphasize a given set of equity risk factors, the number of investors claiming to have the magic formula to rotate effectively between these strategies has increased. Many view valuations as the foundation of what investors should consider when determining which factor ETFs to invest in. They are important, but for most investors, valuations are just one piece of a bigger puzzle to factor investing.
Factor ETFs Have a Wide Range of Cycle Lengths
Factors move in and out of favor just like any investment strategy over the course of a full market cycle. The past few years have shown that quality, momentum, and minimum volatility have all provided some opportunity to outperform a representative equity benchmark, while value and size have both lagged.
Click to enlarge:
Source: MSCI. Rolling 12-month excess returns for MSCI ACWI Factor Indices vs. MSCI All Country World Index. As of 8/31/2016.
So, how can one measure the typical length of a cycle for a factor? Given the slow-reverting nature of factors in general, they can often be analyzed with clever statistical techniques normally used to better understand properties of signals and wave lengths. The emphasis of this analysis is to find the lowest frequency that has the greatest magnitude in explaining the peak of a factor cycle. For our purposes, we define this as the fewest months where the magnitude is greater than 3% (or as close to it as possible).
Click to enlarge:
Sources: MSCI, CLS Investments. Spectral analysis done using rolling 12-month excess returns for MSCI ACWI Factor Indexes relative to MSCI All Country World Index.
The number of months it takes a factor versus the market to come full circle in performance can vary quite considerably. Value and size take a very long time to come full circle on average. On the other hand, minimum volatility can cycle much quicker, while momentum, quality, and dividend yield have more medium-term cycles.
But, what good is simply knowing the length of each factor’s cycle when it comes to timing? That information can be valuable when combining it with what’s currently known about the factor’s persistence in performance, which can determine the likelihood of the factor’s trend reversing or effectively reverting.