Hedging Factors with ETFs
As we've discussed in previous blogs, there are various reasons why portfolio managers need to monitor and adjust the factor risk of their portfolio. Much of the time, it's prudently done to avoid unknown or unintended exposures. Other times, the emergence of a new factor risk in a portfolio can be evidence of style drift that a PM doesn't want. Sometimes, a manager has a macro view that leads to the conclusion that certain factors like Value or Momentum are being misvalued, suggesting that reweighting the portfolio’s factor exposures will lead to additional alpha.
We've developed a technical indicator that will not only alert managers concerned about factor risk, but also assist those who see the management of factor exposure as a way to actively increase alpha rather than mitigate unwanted risks.
To find a predictable pattern in factor movements, we smooth the data and then Z-score the factors, looking at how many standard deviations they are from their recent mean. Factors that are over one standard deviation above or below the mean are labeled as overbought or oversold, respectively. When they are two standard deviations above the mean, they are labeled as extremely overbought/oversold. We’ve found that over 70% of the time, factors in "extreme" territory revert to the mean within the next quarter.
With the knowledge that a factor may revert towards the mean, Omega Point already provides many different work flows that allow users to increase or decrease their portfolio’s exposure to stocks with factor exposures. However, sometimes it is hard to adjust the size of an equity position. There's a reason that equity positions are initially sized as they are, and the appetite for taking on additional idiosyncratic company risk to adjust factor exposure is often minimal. That’s why we also provide a list of ETFs correlated with the factor returns, daily over the past year, alongside the factor signal.
Let’s look at the Size factor as an example. At the beginning of May, it was slightly oversold (at trough it was at -1.47 standard deviations) and has since reverted back to neutral territory as large caps outperformed small caps.
For ETFs that are positively correlated to Size, we don’t see anything over 35%. When looking at the most negatively correlated ETFs, however, we find many more options that move in sync with Size factor, albeit inversely correlated.
When seeking to hedge out factor risk with ETFs, it's best to take ETFs that make intuitive sense. If an ETF has high correlation with a factor but it’s unclear why, then it's safer to avoid using it. If two ETFs have factor exposure, but for different reasons, then it may be a good idea to use a little bit of each rather than only picking one. It’s likely that the ETFs are correlated to factors for different reasons.
Let’s use the Dividend Factor as an example, which has seen a lot of movement as of late and is currently oversold (-1.19 standard deviations).
The ETFs most correlated with the Dividend Yield factor own treasury bonds. When interest rates go up, the high yielding dividend stocks are worth a bit less, and when interest rates go down, alternative sources become more valuable so this makes sense. Still, the correlations are below 0.5, because there are many risks that impact treasuries and stocks differently.
When we look at the ETFs that are negatively correlated to Dividend Yield, we see that they hold mostly Biotech stocks. These stocks are burning through cash and debt and are not going to provide their investors with any yield anytime soon, so it makes sense to see them negatively correlated with the dividend yield factor. We also see an equal weight index of the Nasdaq, which gives more weight to tech stocks that aren’t returning capital to shareholders. A proper ETF hedge of the dividend factor would likely include long positions in bond ETFs and short position in biotechnology and smaller sized technology ETFs. This covers multiple different drivers behind the dividend factor while at the same time avoiding the introduction of any new significant idiosyncratic company risk.
We hope that you find this tool helpful. Please reach out to us with any questions or comments that you might have.
Thank you and good luck!