The past decade – a ‘free money’ era – can be earmarked as an exceptional period for markets with an ultra-low or negative interest rate setting encouraging exuberant behavior. But this has now given way to normalization: interest rates are on the rise, volatility has once more become a recurring market feature, risk is relevant again, fundamentals are top of mind for investors, and exuberance is dissipating.
Against this backdrop, guarding against downside risk becomes an increasingly important consideration in the context of an overall portfolio. In a recent article,1 Robeco researchers showcased how a low volatility approach typically preserves capital in down markets and participates meaningfully in up markets over a 155-year sample period.
In a new research paper,2 we took a different perspective when assessing the defensive characteristics of low volatility portfolios, by evaluating how they are affected by a wide range of macroeconomic risk factors. While these portfolios have lower overall volatility and beta than the market, their concentration in certain segments could lead them to be similarly or even more exposed to some specific sources of systematic risk.
Low Volatility portfolios dampen exposure to risk factors
To investigate this, we compared the S&P 500 Index against two generic low volatility strategies – the MSCI USA Minimum Low Volatility Index and the S&P 500 Low Volatility Index – by assessing their exposure to various risk factors. Our sample spanned the period from January 1991 to December 2021.
In our findings, we saw that the low volatility strategies were less affected by bond yield changes, indicating that they were less sensitive to moves in interest rates. Rising equity and bond implied volatilities also tended to coincide with far larger negative returns for the market than for the low volatility strategies. Similarly, the latter were also less exposed to liquidity risk than the former.
Looking at commodities, we found that the low volatility indices were only half as sensitive to oil price changes versus the market, which we attribute to their lower exposure to fossil fuel stocks. Meanwhile, the betas towards gold price movements were all insignificant, implying that these are not a relevant systematic risk factor.
In line with the previous findings, we observed that low volatility portfolios were consistently less exposed than the market to shifts in consumer sentiment, investor sentiment, the economic policy uncertainty index or the ISM purchasing managers’ index. Again, we saw the same trend when analyzing changes in traditional macroeconomic indicators such as CPI, GDP, industrial production and unemployment.
Lastly, we established that the market was more sensitive to changes in carbon emissions one quarter ahead compared to the low volatility strategies. This serves as another indication that the latter are less exposed to climate risk than the former.
The exception that proves the rule
Low volatility portfolios are not entirely foolproof and can be caught off guard. A concrete example of this oddity is illustrated by their uncharacteristic performance during the Covid-induced sell-off. In this exceptional period, these strategies failed to provide downside protection as they experienced similar losses as the market. Moreover, they lagged the market in the subsequent recovery.
We argue that this event was very different from the wide variety of macroeconomic risks with which low volatility portfolios typically cope well, as Covid completely surprised investors. Online stocks that used to be rather speculative suddenly became defensive holdings when the world went into lockdown, while traditionally safe offline stocks such as commercial real estate suddenly became high risk.
Historical stock prices do not properly reflect certain risk factors if investors previously dismissed them as irrelevant or were simply unaware of them. So when a novel risk factor quickly becomes the dominant market theme, data-driven methods understandably need time to adjust to the environment, leading to atypical outcomes like in the Covid scenario.
But generally speaking, stock price movements are structurally affected by bond markets, commodity markets, macroeconomic indicators and sentiment. As a result, low volatility portfolios are able to continuously adapt to this information, enabling them to offer typically consistent and robust risk reduction over the long run.
Based on our findings, low volatility portfolios do not merely reduce risk in general, but also consistently lower the exposure to the various drivers of systematic risk. That said, the lack of downside protection provided by low volatility strategies during the Covid pandemic episode shows that data-driven investment approaches can be temporarily challenged by rapidly unfolding novel events.
1 See: Baltussen, G., Van Vliet, B. P., and Van Vliet, P., May 2022, “Conservative investing stands the test of time”, Robeco article.
2 Blitz, D., September 2022, “Macro risk of low volatility portfolios“, SSRN working paper.