Low Volatility Anomaly Lacks Robustness?

/Low Volatility Anomaly Lacks Robustness?

Low Volatility Anomaly Lacks Robustness?

By | 2017-08-18T17:01:17+00:00 March 10th, 2014|Low Volatility Investing|18 Comments
Print Friendly, PDF & Email
(Last Updated On: August 18, 2017)

Question: How many ETF companies are hawking “Smart” beta products that offer low volatility or low beta portfolios (we could probably throw minimum volatility in this basket as well)

Answer: All of them

Another Question: How many ETF companies are concerned about the robustness of the products they are hawking?

Answer: Not many.

Here is a simple study on the low volatility anomaly using out of sample data on the Indian stock market.

The evidence suggests the low volatility anomaly is not a panacea:

Theory suggests a direct relationship between risk and return. But several empirical studies find that portfolio of low volatility stocks outperforms portfolio of high volatility stocks. This is termed as low risk anomaly. Our objective is to study whether low risk anomaly exists in India. Using data for the sample period running from January 1994 to June 2010 we find that high volatility quintile yields high return in India.

Source: http://www.iracst.org/ijcbm/papers/vol3no12014/14vol3no1.pdf

Nothing works all the time and every time, but robust alpha drivers should show some semblance of robustness in out of sample tests.


  • The views and opinions expressed herein are those of the author and do not necessarily reflect the views of Alpha Architect, its affiliates or its employees. Our full disclosures are available here. Definitions of common statistics used in our analysis are available here (towards the bottom).
  • Join thousands of other readers and subscribe to our blog.
  • This site provides NO information on our value ETFs or our momentum ETFs. Please refer to this site.

About the Author:

After serving as a Captain in the United States Marine Corps, Dr. Gray earned a PhD, and worked as a finance professor at Drexel University. Dr. Gray’s interest in bridging the research gap between academia and industry led him to found Alpha Architect, an asset management that delivers affordable active exposures for tax-sensitive investors. Dr. Gray has published four books and a number of academic articles. Wes is a regular contributor to multiple industry outlets, to include the following: Wall Street Journal, Forbes, ETF.com, and the CFA Institute. Dr. Gray earned an MBA and a PhD in finance from the University of Chicago and graduated magna cum laude with a BS from The Wharton School of the University of Pennsylvania.
  • Pingback: Monday links: further exuberance | Abnormal Returns()

  • Bid Shader

    So you ignore the global evidence published in the FAJ, Journal of Finance, JPM, etc, and look at one paper in an obscure journal on the Indian stock market over a fairly short sample, and then conclude that all the other papers are “bogus”? 🙂

  • Great point, however, whenever an academic paper gets published with a “non-story” it is interesting to me, because the incentives for data-mining and “making the data sing” are high in the top-tier journals. Regardless, outstanding point.
    I assume you are talking about the following FAJ: http://pages.stern.nyu.edu/~jwurgler/papers/faj-benchmarks.pdf
    Want to share with readers.

  • Eric Falkenstein

    Could it be they took the current BSE-500? That’s a very big bias. If you take the current S&P500, you’ll geta big positive relationship between vol and returns over the past N years. But even if not, they used Bloomberg for data. Bloomberg does not have a good historical database, it drops dead firms that aren’t highly conspicuous (eg, Enron is there), or at least makes it impossible to find them (you have to know their identifiers).

  • There could definitely be data issues. But even if the data were perfect (it never is), it is unclear why they find a result that is OPPOSITE of prior findings? I’m not suggesting the low volatility anomaly has no empirical basis, I’m just highlighting that the strategy–like ALL other strategies–is not a holy grail.

  • Eric Falkenstein

    It’s very simple. If you take the current S&P500, and look back at the volatility and average the future returns into deciles, you necessarily get a positive relation. It’s not subtle.

    And noting that a strategy is not perfect (aka panacea/holy grail) should go without saying, or rather, if one were found one wouldn’t write about it.

  • Steve

    You deserve to be taken to task for this one Wes! (Nicely though)
    🙂

    I’ve seen some discussion on the low volatilty not working post-discovery, but too many others to dismiss it. Also, it’s not a new discovery. I’d have to dig it up, but the low beta/volatility conundrum (low volatility beating high volatility) has been spoken about decades ago.

    Bob Haugen’s paper is one of my favourite (though it spans only a 20 year test).

    Intuitively (and yes, I know that’s very dangerous) it reminds me of the so called ‘quality’ (I prefer, ‘profitability’) effect. i.e. It’s what I would expect (as a contrarian): humans will over bet the long shots (poor quality) and underbet the favourites (quality stocks).
    It happens at the race track too (the world over)…long shots are actually OVER (not under) bet (in aggregate / quantitatively – not talking about individual situations), in that horses that should be paying 1000/1 or 10,000/1 (never going to win a race before becoming dog food) will only be paying 100/1 or 50/1…a massive difference in percentage terms.

    Anyway – bit long winded, sorry.
    All that to say, it wouldn’t surprise me if the same thing happens with volatility.

    However, I personally do not put volatility on my top shelf. It sits with liquidity, size and ‘quality’ on the second shelf.
    Only value and momentum sit on my top shelf, for good reason, and it would probably take decades before I’d elevate the others to sit alongside V & M.

    Now, re: your last comment; 2 things.
    1) You were *not* as you said, “just highlighting that the strategy…is not a holy grail.” If you had said that, that would be okay. In the article, you actually said, “The evidence suggests the low volatility anomaly is bogus.” Which is simply a silly conclusion, as others have pointed out.

    2) Secondly, where you expressed surprise at the guys finding an “OPPOSITE” of prior findings – I’d just like to posit the possibly totally wrong analogy of momentum in Japan (the thorn in the side of all the momentum studies until Asness came along and cleared things up by showing us that Japan was simply, ‘the exception that proves the rule’ – for readers who are unaware, value outperformed exceptionally well in Japan, ‘making up’ for momentum’s failure).
    Maybe India is the volatility anomaly’s, ‘exception that proves the rule?’

    (Please take my post in the right way – love your work, it’s some of my favourite!)

  • Bid Shader

    I submitted another comment that maybe went to spam, but I think saying it’s bogus is a lot different than saying it’s not a holy grail. I agree with you on the second point, but I think this study you’ve cited is not enough for me to reject the 10 or so other papers I’ve seen that show support for the effect in most markets. I mean, do we reject Fama and French based on one paper covering 10-15 years in the Indian market?

  • Great set of comments. Agree that value and momentum are “top-shelf” anomalies–love the classification system.
    Appreciate all the comments. Changed wording to “not a panacea.”
    Clearly, there are some folks who love the low-volatility anomaly!
    I’m still a bit curious if the anomaly isn’t a noisy proxy for the value-anomaly–seen any definitive evidence suggesting otherwise?

  • Bid Shader

    In response to your Q on whether it’s value in disguise – there’s always some co-mingling of return sources. You can adjust for Fama French factors and still have positive alpha from Low Vol portfolios, but work I’ve done suggests that the value (and other factor) exposures of Low Vol are highly time varying. Sometimes it has value characteristics, sometimes not. Interestingly the most powerful factor that I’ve seen in explaining Low Vol alpha is momentum, not value or size.

  • http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1949853
    See footnote 1. Any responses to these literature veins that find potential faults in the low vol anomaly?

  • sounds a lot like noise to me…

  • Bid Shader

    I think we’re talking past each other. Having time varying FF coefficients is not a bad thing, and indicates that it’s not just “value in disguise” or something. The fact is that Low Vol portfolios are sometimes cheap, sometimes expensive, sometimes high momentum, sometimes low momentum, etc. If that WEREN’T the case, then you could argue that it’s just repackaging exposure to well known factors, but it’s not.

    I haven’t read the paper you linked to. I’ll check it out. Thanks.

  • That does make some sense, didn’t mean to talk past you or over you. Just getting back from a wedding and I was firing off quick comments. I’d just be curious if the time varying aspect is predictable–that would be impressive and I’d love to learn more!

  • Bid Shader

    No worries – it’s been an interesting chat.

    I think that sometimes it’s counterproductive to think of Minimum Variance as a “Strategy”, necessarily. In truth, it can be considered more of a weighting scheme. This isn’t to say it’s unimportant or not valuable, but it may not be a strategy in the traditional sense.

    Instead, you can say that a strategy could be Value, or Growth, or Quality, or Momentum, or Low Volatility, and then you could look at weighting schemes like Market Cap, Equal Weighted,Variance Minimization, Risk Parity, Maximum Diversification, etc. So then you have 5 strategies and 5 weighting schemes (not counting multi-factor approaches), and each of the 5×5=25 portfolios will have certain characteristics and performance results that may or may not be favorable.

    So I think that whether or not Low Vol/Min Var is useful or “good” depends on what you expect from your portfolio and what are your goals. If you’re optimizing alpha relative to a market cap benchmark, then ignore low vol. If you’re maximizing wealth and minimizing wealth volatility, then Low Vol can be useful to you.

  • Andrew M.

    Wes – This link might help with your search for truth: https://www.brighttalk.com/webcast/2163/112745
    It discusses the apparent return driver behind both the value premium and the low vol premium (rebalance effects).

  • thx for sharing. Will check out!