This is Part 6 in a series about smart beta ETFs. See below for links to other posts in the series. In this installment, we look at the low volatility anomaly: the surprising idea that stocks with lower risk have tended to outperform.
In our introduction to smart beta a couple of weeks ago, we discussed the capital asset pricing model (CAPM), which is based on the idea that you should expect higher returns from stocks with higher risk. That’s an idea many investors now take for granted, since being rewarded for additional risk makes intuitive sense.
But what if it’s not true?
As early as 1972, Robert Haugen and James Heins found the exact opposite in the data on US stocks between 1926 and 1971. The researchers uncovered a negative relationship between risk and return: high volatility stocks actually tended to deliver lower returns, while low-vol stocks outperformed. In the decades since, many researchers have demonstrated that this low volatility anomaly exists in stock markets around the world, and even in many bond markets.
As with all of the smart beta factors, it’s important to understand how this one is defined. One methodology focuses on the individual stocks. These low volatility indexes start with a universe of stocks (say, all of the companies in the S&P/TSX Composite Index) and then select those that had the lowest volatility over some time period. Volatility can be measured in one of two ways: one is to look at standard deviation, or the amount by which a stock’s price movements vary around its average. But other low-volatility indexes choose their stocks according to beta, which has two components: the stock’s volatility and how closely it is correlated with the market as a whole.
Now consider a minimum volatility (or minimum variance) index, which is built without regard to how volatile the individual stocks are in isolation, or even relative to the market. What’s important is their correlation with each other: the goal here is to combine stocks in a way that results in a portfolio with the lowest possible volatility. It helps to think of this strategy like a recipe that combines items like baking powder and salt, which can be “volatile” on their own but are delicious when combined with the other ingredients.
It’s worth noting that these two different methodologies can produce very different portfolios:
- A stock that has a low standard deviation can have a high beta if it is highly correlated with the market.
. - A low-beta stock can be highly volatile (high standard deviation): its wild price swings would just be uncorrelated with the market.
. - A minimum volatility portfolio is specifically designed to have more or less the same sector breakdown as the broad market. However, portfolios based on standard deviation or beta are often skewed toward sectors that are less volatile than average, such as utilities, consumer staples and real estate.
Why have low-volatility stocks outperformed?
It’s hard to explain the outperformance of low volatility stocks by arguing that they reward investors for higher risk. After all, risk and volatility are often considered synonymous. As Nardin Baker and Robert Haugen describe in their 2012 paper: “The fact that low-risk stocks have higher expected returns is a remarkable anomaly … because it contradicts the very core of finance: that risk bearing can be expected to produce a reward.”
Some have offered behavioural reasons, which are similar to those used to explain the value premium. Investors may prefer the lottery-like quality of high-volatility stocks—which offer the possibility of outsized returns—and therefore will overpay for them.
Baker and Haugen (in the same paper linked above) present evidence that institutional investors may also have a preference for high-volatility stocks. If you’re an analyst recommending a stock to an investment committee, it’s easier to make your case when the company is in the news, and media coverage tends to increase volatility. This creates “demand by professional investors and their clients for highly volatile stocks. This demand overvalues the prices of volatile stocks and suppresses their future expected returns.”
What the critics say
Whenever a strategy focus on stocks with a particular characteristic, there is a danger that the portfolio can be too concentrated, which adds additional risks. Low-vol index providers often apply constraints—for example, limiting the amount of weight given to a specific stock or sector—but by doing so they tend to make the portfolio look more like a traditional cap-weighted index, which of course dilutes any advantage of a smart beta strategy.
Perhaps the biggest danger with low-volatility strategies is investor expectations. While it might be possible to select stocks with lower-than-average volatility without sacrificing expected returns, there will still be plenty of gut-wrenching drawdowns along the way. As The Wall Street Journal puts it in a recent article: “Expecting a low-volatility stock portfolio to eliminate most of the painful turbulence that stocks deliver isn’t realistic. If you want a serious cushion for volatility in your portfolio, you need some bonds.”
How to access the low volatility factor with ETFs
Investors looking for low-volatility ETFs can get by using only Canadian-listed funds, as several of the big providers offer them. However, these ETFs all use different methodologies, so pay close attention to what you’re buying.
BMO’s family of low volatility ETFs select large-cap stocks based on their beta over the last five years. The BMO Low Volatility Canadian Equity (ZLB) holds 40 companies, while the BMO Low Volatility US Equity (ZLU) includes 100. The BMO Low Volatility International Equity (ZLI) includes 100 large caps from outside the US and Canada, with the biggest holdings currently in Japan and France. In all cases, the methodology specifies that no individual stock can make up more than 10% of the portfolio, and no sector can comprise more than 35%. In the international ETF, individual countries are capped at 25%.
PowerShares offers four low-vol ETFs covering Canada, the US, international developed and emerging markets. These funds use standard deviation over the last 12 months to select the lowest volatility stocks in their parent S&P indexes. The PowerShares Low Volatility Portfolio (PLV) combines all four of these ETFs (as well as two US-listed funds and a 30% allocation to a bonds) in a globally balanced portfolio.
In contrast with both BMO and PowerShares, iShares uses the minimum variance strategy to build its family of low-volatility ETFs. Again, there are funds for Canada, the US, international and emerging markets, all built from MSCI indexes. There’s also a global version that simply holds the US-listed iShares Edge MSCI Min Vol Global ETF (ACWV). It’s roughly 57% US with the next largest holdings in Japan, China, Switzerland and Canada.
The Vanguard Global Minimum Volatility (VVO), part of Vanguard Canada’s new lineup of smart beta ETFs, is an actively managed fund that does not track an index or make its methodology public. In general, however, the fund uses a minimum variance approach rather than selecting individual stocks based on their volatility. This ETF holds about 50% of its stocks in the US, 6% in Canada and the remainder in overseas developed and emerging markets.
Other posts in this series:
Smart Beta ETFs: Your Complete Guide
Understanding the Value Factor
Understanding the Momentum Factor
All of these articles on these smart beta ETFs leave me eagerly anticipating your verdict on their expected out-performance vs. their additional cost. Am I correct in assuming that your advice will relate to the amount invested? I.e., large investors seeking gains on the margins will likely see more benefit in these smart beta ETFs?
@Raman: I’ll certainly share my thoughts in more detail later in the series, but I don’t think an investors portfolio size is an important variable. Overall it’s true that small portfolios should be kept simple, but it doesn’t follow that large ones need to be more complex. Remember that costs and taxes are also higher (in dollar terms, if not percentage terms) on larger portfolios as well.
I have also been very interestingly reading your smart beta ETF series of blogs. With so many anomalies that have been found to outperform the market, I’m almost questioning the CAPM and market efficiency. Which also leads me to question market weighted indexes. I will continue to read the series! Thanks,
@cpp
Could one just use a global low volatility etf (XMW) along with a canadian bond index etf (XQB) for a portfolio? Maybe add XMV if one wants to add more Canada exposure.
@jake that is exactly what I’ve done for our RESP, with the added bonus that XMW holdings are light on carbon, making it a semi-ethical ETF.
@Matt: We’ll definitely get to a discussion of these issues, so stay tuned. In the meantime, these might be of interest:
https://canadiancouchpotato.com/2013/07/15/does-smart-beta-really-beat-cap-weighting/
https://canadiancouchpotato.com/2013/07/18/why-your-problem-is-not-your-funds/
https://canadiancouchpotato.com/2016/02/17/how-long-will-you-wait-for-smart-beta-to-work/
@Jake: Could you use XMW as an all-in-one equity portfolio? Sure. The question is whether you would expect it to outperform a traditional all-world index fund, like XWD, for example.
“researchers have demonstrated that this low volatility anomaly exists in stock markets around the world, and even in many bond markets”
This raises an interesting question: is there smart beta in the bond market? If so, what does it look like?
To quote the above-linked 2013 article:
“The search for smart beta makes me think of an out-of-shape guy with a poor diet who’s convinced his biggest problem is he’s wearing the wrong brand of running shoes.”
For the record, that has to be one of the greatest analogies of all time.
@ Raman
*** spoiler alert ***
Smart beta is not new, and certainly not new to CCP. His recommended portfolios are very simple and made up of market weighted indices. CPP has said many times before that there are some merits to factor based investing but due to the additional costs, complexity, and additional willpower required to stick with them when they are not working, he will continue to recommend only market cap indices for average investors.
@Shaun: I expect that’ll be the gist of his recommendations, but he might still have a few ideas about ways to access part of the smart-beta premium through low-cost etfs, which is why I’ve been keeping a particular eye on this series of posts.
Don’t these ETF have like 10x the MER of index funds? Although 10x of like 0.05 is not much.
It’s just that from my one year experience of holding the BMO ones, they don’t follow the TSX. It’s a good thing but kinda hard to know when to put more in during the past few corrections since the radio doesn’t talk about these funds.
It also seems like I’m actually behind VUN for the past year for some reason. Could it be because ZLB is too big now?
And like others have mentioned, these are too different than indexing so it’s not gonna be that good if people ended up buying high and selling low. Because indexing had a benchmark that everybody understands, it’s easier to accept drops in price.
If someone asks me ZLB vs VUN, I’ll say VUN or XDV instead. At least they’ll be able to collect dividend during bad times. It’ll put them at ease somewhat.
Owell. Most will just end up chasing part performance anyways. And ZLB isn’t really outperforming this year. O.o
Branko, here’s a better analogy …. low vol smart beta is like the marathon runner who uses a different training method than all the mainstream runners and that the mainstream runners laugh at, saying “it cannot work, you are a fool”, yet the oddball runner wins the race. The other runners explain this by saying he must have cheated or used drugs or had better running shoes, everything but recognizing his different training method (not using market cap to select and weight stocks).
I get the sense that some investors are calling minimum volatility black magic or something. There’s no mystery to it that I can see. In Canada, anyway, they have basically removed volatile energy and materials stocks from the index.
In 2015, when energy and materials tanked (along with all of the broad-based ETFs in Canada), the minimum volatility ETFs won. In 2016, when oil surged, the minimum volatility ETFs lagged. Some years one will win and some years the other will win. The Canadian market is so thin, it’s hard for any ETF to not heavily overlap other ETFs, anyway.
Since I need reliable monthly income, I’ve decided to allocate 50% of my Canadian and US equities to minimum volatility.