No, It's Not A Minimum Volatility Bubble

Includes: EUSA, USMV
by: Dan Morillo

On the back of a market rally and increased investor interest in minimum volatility strategies, I have been getting a lot of questions on whether those strategies now appear expensive compared with a standard capitalization-weighted portfolio. In short, clients are asking if minimum volatility has been “overbought” and is not likely to be effective going forward. My response is that while at first glance the answer would seem to be “yes,” a deeper analysis shows the answer is actually “no.” But more importantly than that, it’s important for clients to realize that minimum volatility should not be thought of as a “market timing” strategy. Let me explain.

As of May 28th the valuation [1] for the MSCI USA Minimum Volatility Index was 8.62x while the MSCI US Index – the comparable cap-weighted index – was 8.16. Based on this metric alone, investors may conclude that exposure to the minimum volatility strategy is “expensive” compared with a typical capitalization-weighted investment and thus potentially not attractive.

This is, however, not the whole story. Valuations will vary across investments when the underlying fundamentals justify reasonable valuation differences. In other words, differences in valuation could be justified by differences in fundamentals of the underlying companies, such as profitability or balance sheet strength. These distinctions are particularly important at the aggregate level where diversified strategies or baskets will “wash out” individual company-level differences and will tend to reflect broad fundamental characteristics of the underlying investments.

In the case of minimum volatility, profitability is one of those differences. To illustrate, the chart below shows the profitability (measured as Return-on-Assets) as well as the valuations (measured as Price-to-EBITDA) across a range of developed countries [2] as of the end of April. Close to 75% of the variation in valuation can be explained just by looking at the aggregate profitability level, with every percentage point in additional profitability explaining about 2 points worth of additional valuation. The US cap-weighted benchmark is on the “expensive” side with a valuation ration of 8.16x but this is partly explained by profitability levels higher than that of the average developed country.

The chart also shows that the minimum volatility index is indeed more expensive than the benchmark, but the aggregate profitability of the portfolio of companies in the minimum volatility index is also higher than the benchmark. Indeed, the minimum volatility exposure seems better valued in the sense that its higher profitability would justify higher valuations given the observed relationship between valuations and profitability across developed countries.

The bottom line is that assessing whether or not a particular exposure or strategy is attractive should include consideration of the many factors that drive valuations in the first place. Most importantly, however, investors should keep in mind that the case for including a minimum volatility strategy in a portfolio is not based on whether its valuation is attractive or not at any given moment. Instead, the key idea behind these strategies is to reflect market behavior that is additive to a portfolio and diversifying to a “traditional” investment style.

I have argued on this blog that behavioral biases as well as institutional constraints are a key driver of the “low beta effect” that is behind the better return per unit of risk that minimum volatility strategies have shown in the past. In fact, the MSCI USA Minimum Volatility Index has been more “expensive” than the standard benchmark every month since the end of 2011, but it has nevertheless delivered better returns per unit of risk as compared to the standard cap-weighted benchmark [3].

Disclaimer: There is no guarantee that minimum volatility funds will attain a more conservative level of risk, especially during periods of extreme market conditions.

[1] Valuation ratios as computed as Price divided by EBITDA. Index-level valuation data is from Bloomberg for the standard MSCI country indices. EBITDA is a standard choice for cross-sectional analysis as it is generally less prone to large observations as simple bottom-line earnings.

[2] The chart shows all developed countries where their standard MSCI index contains 20 or more names. Index-level data for ROA is from Bloomberg as of May 28. ROA at the index level is computed aggregating Earnings and Assets separately and then dividing the sums.

[3] Data is from Bloomberg. Annualized returns from monthly data for the MSCI US index since end-December 2011 until end-April 2013 are 16.3% with annualized risk of 9.9% for a Sharpe ratio of 1.64. The equivalent numbers for the Minimum Volatility index are 17% annualized returns, 7.6% annualized risk for a Sharpe Ratio of 2.23.

Original post