Can we sidestep the 'irrational exuberance' in generic low volatility?

Low volatility strategies have been growing in popularity, as investors are attrated to the potential of generating equity-like returns with better capital preservation and lower levels of drawdown. But in the context of potential crowding and overvaluation, generic approaches may no longer offer a safe harbour. We examine the reasons that a differentiated low volatility approach may become important for investors, illustrating the significant difference that portfolio construction choices can make to an investment strategy. We argue that investors should opt for those low volatility approaches which are designed to steadily exploit the underlying anomaly, while side-stepping risks associated with crowdedness and overvaluation.



Low Volatility strategies have been among the strongest asset gatherers in the recent past. As an example, the shares outstanding of the US Minimum Volatility iShares ETF (USMV) has more than doubled in the past 12 months (Chart 1). Questions have been raised in the popular press about such crowding as well1. Meanwhile, low volatility is becoming troublingly more expensive: the Price-to-Earnings ratio of the lowest 1/5th of stocks sorted by Barra GEM2 Global Beta is higher than that of the market and rising, while that of the highest 1/5th is lower than that of the market and falling (Chart 2).

Chart 1. Shares outstanding of US Minimum Volatility iShares ETF (October 19, 2011 – August 17, 2016)

Source: Bloomberg and iShares.

Investors are attracted to low volatility’s potential to help preserve capital with lower drawdowns, while still seeking equity-market-like returns. With issues such as potential crowding and overvaluation though, generic low volatility may no longer seem like a safe harbour in a potential storm. What does this mean for a low volatility investor: Is it possible to capture the low volatility anomaly2 and avoid the crowded trades and holding a portfolio that is more expensive than the market?

Chart 2. Price-to-Earnings ratio of low- and high-beta quintile in the MSCI ACWI universe (December 2002 – June 2016)

Source: Numeric Investors LLC and MSCI.

Characteristics of generic low volatility - Should we worry?

The question of crowdedness cannot be answered definitively. The growth in ETF shares outstanding (seen in Chart 1) may be large, but it is off a low base. When we look at the active holdings in the low volatility tails of beta, leverage, or yield, we do not see historically high levels of crowding (not shown here). Intuitively it seems as if more money is allocated to investment ideas such as Value than to low volatility. It is, therefore, possible that the crowdedness of popular versions of low volatility has not yet reached an alarming level. The rapid rise in investor interest in any investing approach should certainly give one pause: is this yet another example of performance chasing that may be destined to disappoint?

The expensiveness of low volatility is also an open question. It is not easy to disentangle whether low volatility has become expensive or whether value has become volatile; these appear to be two sides of the same coin. No investor can be faulted for not wanting to toss this coin – is it possible instead to have a low volatility approach that is either neutral or even positively oriented towards value?

The momentum exposure of low volatility – while it doesn’t quite have the same fundamental connotations as value exposure – is also reaching interesting levels (Chart 3). Given its recent success, it is no surprise that low beta has a large positive exposure to momentum. The gap between the momentum exposures of low- and high-beta tails, however, is increasing. For how long can this be sustained?

Chart 3. Barra GEM2 Momentum of low- and high-beta quintiles in the MSCI ACWI universe (December 2002 – June 2016)

Source: Numeric Investors LLC and MSCI Barra.

Of course, there is no conclusive evidence that the story has to end badly for generic low volatility. Risk estimates change over time – it is possible that stocks that are currently low risk will evolve into high risk as they beat a rising market, causing a generic low volatility strategy to rotate out of those expensive names. It is also possible that valuation imbalances are cured not by price corrections but by improving fundamentals. Nonetheless, we believe a cautious investor would be well served in looking for a low volatility strategy that is insulated from these topical concerns.

Is it possible to differ from generic low volatility?

Low Volatility itself is a loose term, broadly applied to approaches ranging from high dividend yield, to high quality, to minimum variance. In this paper, we limit ourselves to the possibilities within the minimum variance framework – which is the closest to some of the most popular ETFs and to low beta (whose value exposure was discussed earlier). As it turns out, even within this minimum variance framework, the portfolio construction process involves a number of choices, presenting an opportunity for a thoughtful investor to choose wisely and build a portfolio that is sufficiently different from the crowd and yet positioned to potentially harvest the low volatility anomaly.

Among the most important of these choices are:

  • What does one optimize (lowest total risk, weigh factor and specific risk differentially, involve transaction costs in the optimization, bring in any consideration of forecast returns, etc)
  • How many and which risk model(s)
  • Currency of risk measurement (common currency or local)
  • Turnover constraints and frequency of portfolio rebalance
  • Sector/country/other constraints on the portfolio and any other fundamental controls on risk

An example of a differentiated low volatility approach

To illustrate how much of a difference the portfolio construction choices can make, we created a simulated low volatility portfolio. Just like the MSCI Minimum Volatility Index, it too has at its heart the goal of minimum variance. The implementation differs considerably: two risk models are used (Barra GEM2 and an in-house Statistical Factor Risk Model); risk is measured in local currency; transaction costs are incorporated in the optimization; a stock-selection model is used to bring in steady exposures to attractive characteristics (value, quality, information from markets other than the equity market, etc.). Each of the portfolio construction choices was done for a reason – to discuss them all would be vastly beyond the scope of this document. What interests us here is whether these choices produce a portfolio that maintains a low volatility stance, while sidestepping the crowdedness of generic low volatility and while remaining desirably cheap with respect to the market.

We find that the simulation produces a portfolio that has a low overlap with the MSCI Minimum Volatility Index and is cheaper than the cap-weighted market index (Chart 4). In ensuring that we produce a low volatility portfolio that is generally cheaper than the market, there is a sacrifice of about 0.1 units of ex-ante Barra GEM2 Beta3. We believe such a compromise is worthwhile – it is better to have a low volatility portfolio that has known, steady, and desirable exposures rather than to have one with transient exposures to different characteristics. This also helps reduce the potential of pursuing a low volatility strategy that is expensive.

Chart 4. Characteristics of MSCI ACWI Index, a Simulated Numeric Low Volatility portfolio, and MSCI Minimum Volatility Index. As of June 30, 2016
  Gem2 Value GEM2 Mo P/E Number of holdings Overlap with MSCI MinVol Overlap with MSCI ACWI Barra gem2 gbeta Yield
MSCI ACWI Index 0.02 0.03 14.8   24.8% 100% 0.92 2.7%
Simulated Low Volatility 0.17 0.44 13.6 107 15.4% 8.4% 0.67 2.5%
MSCI ACWI Min Volatility Index -0.21 0.32 18.6 359 100% 24.8% 0.56 2.7%

Sources: Numeric Investors LLC and MSCI.

Chart 5. Barra GEM2 Momentum of low- and high-beta quintiles in the MSCI ACWI universe (January 2003 – June 2016)
2003 Jan – 2015 June MSCI ACWI Simulated Low Volatility
(net of 0.60% assumed fee)
Return 9.3% 13.0% 11.1%
Volatility 15.6% 11.2% 10.7%
Realized beta 1.00 0.65 0.63
2015 July – 2016 June      
Return -3.2% 1.4% 12.3%
Volatility 15.5% 12.0% 11.5%
Realized beta 1.00 0.74 0.65

Source: Numeric Investors LLC, MSCI Barra and Bloomberg.

While this simulated approach proved superior to the MSCI Minimum Volatility Index in the long term (Chart 5), in the recent two years the simulated approach has lagged. We feel this recent weakness validates our approach of not being in the seemingly crowded trade.

An aside on benchmarking

The recent performance differential also brings up questions of benchmarking: should a manager’s low volatility strategy be deemed insufficient if it lags a popular signpost in the landscape such as the MSCI Minimum Volatility Index? What other benchmarks are possible (we think there are at least two) that will help us understand whether the portfolio at hand is continuing to behave as is expected of a low volatility portfolio? While beyond the scope of this paper, these are questions that asset owners and managers must contemplate at a time when generic low volatility indices seem to be in frothy territory.


The potential crowdedness and overvaluation of generic low volatility brings up legitimate concerns about the solid footing that many investors expect from low volatility in its potential to provide low drawdowns and to preserve capital. We believe investors should look for, and choose, those low volatility strategies that are designed to exploit the underlying anomaly while side-stepping risks associated with crowdedness and expensive valuation, and while providing known, desirable, and steady portfolio characteristics.


1. “Have Low Volatility ETFs Gotten Too Crowded” August 19, 2016. Barron’s; “’Low Vol’ Funds Attract More Than $10bn of Inflows This Year” May 30, 2016.; “For Low Volatility ETFs, More Money Could Mean More Problems” June 2, 2016.

2. The low volatility anomaly can be described as follows: Fully invested low volatility portfolios have been empirically observed to produce returns comparable to that of the market, but at a lower volatility.

3. The MSCIs Minimum Volatility Index is built by using the Barra risk model, whereas Numeric’s simulation uses an additional in-house risk model. As such, it is natural that the former appears as lower volatility when viewed through Barra’s own risk measure.

Download Full Article

Important information

Opinions expressed are those of the author and may not be shared by all personnel of Man Group plc (‘Man’). These opinions are subject to change without notice, are for information purposes only and do not constitute an offer or invitation to make an investment in any financial instrument or in any product to which the Company and/or its affiliates provides investment advisory or any other financial services. Any organisations, financial instrument or products described in this material are mentioned for reference purposes only which should not be considered a recommendation for their purchase or sale. Neither the Company nor the authors shall be liable to any person for any action taken on the basis of the information provided. Some statements contained in this material concerning goals, strategies, outlook or other non-historical matters may be forward-looking statements and are based on current indicators and expectations. These forward-looking statements speak only as of the date on which they are made, and the Company undertakes no obligation to update or revise any forward-looking statements. These forward-looking statements are subject to risks and uncertainties that may cause actual results to differ materially from those contained in the statements. The Company and/or its affiliates may or may not have a position in any financial instrument mentioned and may or may not be actively trading in any such securities. This material is proprietary information of the Company and its affiliates and may not be reproduced or otherwise disseminated in whole or in part without prior written consent from the Company. The Company believes the content to be accurate. However accuracy is not warranted or guaranteed. The Company does not assume any liability in the case of incorrectly reported or incomplete information. Unless stated otherwise all information is provided by the Company. Past performance is not indicative of future results.


Please update your browser

Unfortunately we no longer support Internet Explorer 8, 7 and older for security reasons.

Please update your browser to a later version and try to access our site again.

Many thanks.