SHOULD YOU CARE ABOUT VALUATIONS IN LOW VOLATILITY STRATEGIES? July 2017 UNCORRELATED ANSWERS TM Executive Summary Increasing popularity of low-volatility strategies has led to fear that low-volatility stocks have become overvalued, suggesting an increased likelihood of future underperformance. In reality, the relative performance of low-volatility strategies in volatile markets has much less to do with stocks valuations than the degree of risk reduction and the reliability of the alpha source (if any). Vassilios Papathanakos, PhD Executive Vice President Deputy Chief Investment Officer Richard Yasenchak, CFA Senior Managing Director, Head of Client Portfolio Management Kevin Armstrong, CFA Senior Product Specialist Attractive valuations do not provide optimal risk reduction in turbulent markets. By adapting the level of risk reduction to the market volatility, a portfolio benefits from varying its exposure to growth and value characteristics. Even in the presence of high valuations, low-volatility strategies should still outperform the market over a market cycle due to compounding effects. Historically, making a tactical allocation to a low-volatility strategy based on valuation signals has had mixed results. A long-term strategic allocation is warranted in order to fully realize the benefits of a low-volatility investment strategy.
Introduction Over the past six years, the number of products and assets invested in low-volatility strategies 1 has increased dramatically. 2 Some investors are concerned that low-volatility stocks have become a crowded trade, and that their valuations are too high, leading to a potential correction as capacity is used up. After all, history has many examples where substantial capital inflows into similar investment strategies has led to future underperformance. In the first part of this series 3, we contested the notion that all low-volatility strategies form a homogenous class of portfolios, whose holdings and trades must necessarily overlap, and are therefore all exposed equally to the risk of overcrowding. In this paper, we examine the relationship between valuations and low-volatility strategies specifically, and demonstrate that their long-term relative performance is more related to the compounding benefits associated with their deliberate downside protection than exposure to stocks with cheaper valuations. What is the relationship between risk and valuations? An observation that may surprise many investors is that attractive valuations do not generally protect against volatility. As seen in Figure 1, both value and growth stocks exhibit a wide range in their respective betas. For example, during the Global Financial Crisis, the drawdown period for the MSCI World Index lasted from November 2007 through March 2009, during which time the MSCI World Value Index declined 46.2% and the MSCI World Growth Index declined by 41.6%. The beta of the MSCI World Value Index relative to the cap-weighted index rose to almost 1.1 and, consequently, this index underperformed the MSCI World Growth Index by 4.7%. 4 For investors seeking out value stocks because of a presumed higher margin of safety, there would have been little downside protection given the higher beta of the value universe of stocks. 1 For convenience, we use the term low-volatility strategies to cover all approaches that attempt to significantly reduce the absolute risk of an equity portfolio, whether or not they actively seek to outperform a capitalization-weighted index via some alpha source. 2 From December 31, 2010 to December 31, 2016, institutional assets under management in the evestment All Low Volatility Equity universe have grown from $20B to $174B. 3 Richard Yasenchak, and Vassilios Papathanakos, How to Overcome Overcrowding in Low Volatility Investing, Intech, July 2016. 4 Furthermore, the drawdown period for the Russell 1000 Index lasted from August 2007 through March 2009, at which time the beta for the Russell 1000 Value Index exceeded 1.1 and underperformed the Russell 1000 Growth Index by 7.7%. The absolute return for the Russell 1000 Value Index was -32.6% and for the Russell 1000 Growth Index was -24.9%. 2
FIGURE 1 BARRA ESTIMATE OF THE BETA OF THE MSCI WORLD GROWTH AND MSCI WORLD VALUE INDICES Style Index Beta vs. MSCI World Index 1.15 1.10 1.05 1.00 0.95 0.90 0.85 12/2004 12/2005 12/2006 12/2007 12/2008 12/2009 12/2010 12/2011 12/2012 12/2013 12/2014 12/2015 12/2016 MSCI World Growth Index MSCI World Value Index This is also seen in Table 1, which shows that attractive valuations do not materially help reduce volatility in a variety of time periods and markets. In the table, we compare the MSCI World Value Index and the MSCI World Index. In general, we find minimal volatility reduction for the value index relative to the full index. The reason why value stocks do not generally provide a good defense against volatility is that their betas are quite variable. The drivers of changes in the beta include both stock-specific (such as the relation of debt, dividend yield, and quality to valuation) and market-wide considerations (such as fund TABLE 1 PERFORMANCE STATISTICS OF THE MSCI WORLD VALUE INDEX RELATIVE TO THE MSCI WORLD INDEX FOR VARIOUS PERIODS, USING MONTHLY RETURNS Relative Beta Volatility Period Return Reduction* 1975 2016 0.99% 0.97 0.8% (14.6% vs 14.7%) 1992 2016 0.63% 0.99-2.0% (14.8% vs 14.6%) 2007 2016-0.92% 1.02-3.0% (17.0% vs 16.5%) flows in and out of value stocks depending on regime shifts). What is clear is that a low valuation does not materially affect a stock s volatility, and a portfolio of stocks with lower valuations does not necessarily lead to a portfolio with a lower risk profile. In other words, valuations and risk do not have a direct relationship, and a valuation-unconstrained 5 low-volatility strategy may tend to favor alternately growth or value stocks, depending on the changing nature of the volatility environment of the market in order to achieve the goal of minimizing risk. Can low-volatility strategies still provide value in the presence of high valuations? Absolutely! Low-volatility strategies have demonstrated an ability to add value over time despite periods when they have faced heightened valuations. In fact, some of the periods when low-volatility strategies have yielded the most significant value have been in periods when they have looked more expensive than the market and their own historical trends. For example, in Figure 2 we plot the cumulative drawdown for two simulated valuation-unconstrained global low-volatility strategies and highlight periods when these strategies have looked expensive relative to the MSCI World Index. * Additionally, the maximum drawdown was only reduced by 2.6% over the entire period, from 56.3% to 53.7%. 5 This is a portfolio that is constructed without using valuation and growth factors as a constraint, or relying on them for outperformance. See Simulations Disclaimer at end of paper for additional information regarding Intech simulations 3
TABLE 2 SIMULATED INTECH GLOBAL LOW VOLATILITY AND GLOBAL ADAPTIVE VOLATILITY STRATEGIES VS. MSCI WORLD INDEX January 1, 1992 to December 31, 2016 Standard Volatility Max Sharpe Return Deviation Reduction Drawdown Ratio Simulated Intech Global Low Volatility 8.90% 9.28% 36.20% 27.30% 0.68 Simulated Intech Global Adaptive volatility 11.00% 10.95% 24.70% 33.50% 0.76 MSCI World Index 7.20% 14.55% 53.65% 0.32 At the onset of the Global Financial Crisis, both low-volatility strategies were expensive relative to the MSCI World Index. Yet, they both demonstrated substantial downside protection during this period. Additionally, we see that during the S&P U.S. debt downgrade in 2011, both low-volatility strategies again provide substantial downside protection while being deemed expensive relative to the broad market index. More generally, a portfolio gains more by moderating large negative returns than it does by boosting positive returns by the same amount. As the magnitude of the drawdown (and resulting volatility) increases, a portfolio s compound return decreases. If investors crowd into low-volatility strategies during crisis periods, low-volatility strategies are expected to still benefit as the prices of these names are driven up through the increased demand regardless of valuations. Yet, even during the recovery when they are likely to temporarily underperform as a result of changing risk appetites, they generally benefit from the more efficient compounding of returns associated with reduced drawdowns and shortened recovery periods (Table 2). FIGURE 2 CUMULATIVE DRAWDOWN OF THE SIMULATED INTECH GLOBAL LOW VOLATILITY STRATEGY VS. MSCI WORLD INDEX CUMULATIVE DRAWDOWN OF THE SIMULATED INTECH GLOBAL ADAPTIVE VOLATILITY STRATEGY VS. MSCI WORLD INDEX Highlighted region indicates periods when the P/E of the strategy relative to the MSCI World Index exceeds its historical average. 0% -10% 0% -10% -20% -30% -40% Drawdown -20% -30% -40% -50% -60% -50% -60% 12/2004 12/2005 12/2006 12/2007 Drawdown 12/2008 12/2009 12/2010 12/2011 12/2012 12/2013 12/2014 12/2015 12/2016 12/2004 12/2005 12/2006 12/2007 12/2008 12/2009 12/2010 12/2011 12/2012 12/2013 12/2014 12/2015 12/2016 Simulated Global Low Volatility MSCI World Index Simulated Global Adaptive Volatility MSCI World Index 4
Do valuations predict future performance of a low-volatility strategy? It is tempting to try to time the market by switching in and out of low-volatility strategies based on one s market view. However, empirically, it is well-known to be very challenging to anticipate the future direction of the market. Also, valuationbased models have demonstrated little ability to help predict future performance of low-volatility strategies given the lack of a clear relationship between valuations and risk as previously illustrated. For example, in Figure 3 we plot the relative P/E ratio and the subsequent 1- and 3-year excess return of two lowvolatility strategies versus the MSCI World Index. Each point on the horizontal axis represents how far from the historical average the P/E ratio of the strategy lies at that point in time, normalized into units of standard deviation, with negative values representing periods when the strategy is cheaper than historical average. On the vertical axis, each point represents the subsequent 1- or 3-year excess return of the strategy. An interesting observation from both figures is that there does not appear to be a strong relationship between historical valuation level and the future performance of either strategy as demonstrated by the low correlation coefficients. In other words, buying low-volatility strategies when they were relatively expensive did not fare worse than when they were relatively inexpensive over 1- and 3-year subsequent periods. FIGURE 3 RELATIVE PRICE-TO-EARNINGS RATIO VS. 1-YEAR, AND 3-YEAR SUBSEQUENT EXCESS RETURN OF SIMULATED INTECH GLOBAL LOW VOLATILITY STRATEGY AND GLOBAL ADAPTIVE VOLATILITY STRATEGY VS. MSCI WORLD INDEX December 31, 2003 to December 31, 2016 1-Year Future Excess Return vs. Normalized Relative P/E Ratio 40% 3-Year Future Excess Return vs. Normalized Relative P/E Ratio 20% Excess Return vs. MSCI World Index 30% 20% 10% 0-10% -20% -30% y = 0.0223x + 0.0186 R² = 0.0707 y = 0.0275x + 0.005 R² = 0.0635 Excess Return vs. MSCI World Index 15% 10% 5% 0-5% -10% -15% y = 0.0127x + 0.0221 R² = 0.0812 y = -0.0026x + 0.0145 R² = 0.0042-40% -3-2 -1 0 1 2 3 4-20% -3-2 -1 0 1 2 3 4 # of Standard Deviations from Historical Average P/E Ratio # of Standard Deviations from Historical Average P/E Ratio Simulated Intech Global Low Volatility Simulated Intech Global Adaptive Volatility Simulated Intech Global Low Volatility Simulated Intech Global Adaptive Volatility 5
Conclusion In light of the increased popularity of lowvolatility investing, there has been mounting concern that these strategies have become a crowded and expensive proposition, and are poised for a future correction. History has indeed shown past examples where heightened valuations have led to factor crashes and subsequent underperformance, but we believe that the threat of factor crashes from expensive valuations is based on an overgeneralization of the category. Our analysis demonstrates that valuations and risk do not have a clear relationship. Low-volatility strategies will tend to favor alternately growth or value stocks based on the changing nature of the volatility environment of the market in order to achieve the goal of minimizing risk. As a result, the relative performance of these strategies has much less to do with valuations of stocks than the degree of risk reduction, which depends on the implementation details of each strategy (primarily, the reliability of the covariance estimates and the targeted alpha source, if any). Overall, it is both simpler and more robust to use a low-volatility allocation as a strategic decision rather than base it on tactical considerations given that timing the market is fraught with peril. Since the benefit of lowvolatility strategies is substantially achieved by limiting the drawdown and shortening the recovery period, it is most appropriate to employ these strategies independently of whether the times appear to be good or bad. Even in the presence of high valuations, these strategies may yield a net performance benefit due to more efficient compounding of returns as drawdowns are reduced through time. 6
Simulations disclaimer All simulated performance results have been compiled solely by Intech and have not been independently verified. Simulations potentially allow investors to understand and evaluate our investment process by seeing how a strategy/ product would have performed hypothetically during certain time periods. This material is provided for illustrative purposes only and should not be construed as an offer to sell, or the solicitation of offers to buy, or a recommendation for any security. It has been prepared for, and authorized for internal use by, designated institutional and professional investors and their consultants or for such other use as may be authorized by Intech or its affiliates. This material and/or its contents are current at the time of writing and may not be reproduced or distributed in whole or in part, for any purpose, without the express written consent of Intech. Although the information contained herein has been obtained from sources believed to be reliable, its accuracy and completeness cannot be guaranteed. Simulated results are hypothetical, not real. They do not reflect the results or risks associated with actual trading or the actual performance of any account. Simulated performance results are prepared with the benefit of hindsight. As a result, the simulations may be theoretically changed from time to time to obtain results that are more favorable. Simulation results do not reflect material, economic, and market factors that may have impacted our trading or decision-making in the actual management of a client s account. Simulated results should not be considered indicative of the Intech mathematical process, as Intech may not have managed money during some of the periods shown or may not have managed money for the particular strategy/product shown. The Intech mathematical optimization process was applied to historical data to produce the simulations. Unlike traditional simulations that do involve fundamental estimates, ours do not. In addition, the proprietary mathematical investment process used by Intech may not achieve the desired results. Our simulated performance results have inherent limitations, including, among other things: 1) simulated performance results are prepared with the benefit of hindsight; 2) no price-based or volume-based deleted list; 3) no posted list; 4) index constituent changes done as a group at the beginning of the month (typically done once or twice a year based on the index changes); 5) simulated trades take place at the closing price (+80 bps for countries in the MSCI Emerging Markets Index and +40 bps for developed countries), while Intech actually trades intra-day (historically, our domestic trading costs have been below the 40 bps used in the simulations); and 6) six trading tranches are simulated with the average of the six tranches being reported as the result for the period. Past performance of simulated data is no guarantee of future results. Therefore, no current or prospective client should assume that future performance will be profitable, or equal to either the simulated performance results shown or any corresponding historical index. In particular, simulations do not reflect actual trading in an account, so there is no guarantee that an actual account would have achieved the results shown. In fact, there may be differences between simulated performance results and the actual results subsequently achieved. In no circumstances should simulated results be regarded as a representation, warranty, or prediction that investors will achieve or are likely to achieve the performance results displayed, or that investors will be able to avoid losses. Investing involves risk, including fluctuation in value, the possible loss of principal and total loss of investment. There are numerous other factors related to the markets in general or to the implementation of any specific trading strategy, which cannot be fully accounted for in the preparation of simulated performance results, all of which can adversely affect actual trading results. Any clients invested in the strategy/product may have experienced investment results during any relevant periods that were materially different from those portrayed in the simulations. The simulated results include the reinvestment of all dividends, interest, and capital gains, but do not reflect the deduction of investment advisory fees unless otherwise noted. Thus, simulated results will be reduced by advisory fees and any other expenses that may be incurred in the management of an account, which will materially lower results over time. An index is unmanaged, is not available for direct investment, and does not reflect the deduction of management fees or other expenses. MSCI makes no express or implied warranties or representations and shall have no liability whatsoever with respect to any MSCI data contained herein, if shown. The MSCI data may not be further redistributed or used as a basis for other indices or any securities or financial products. This report has not been approved, reviewed, or produced by MSCI. Data Source: The Center for Research in Security Prices ( CRSP ) Deciles are market value weighted benchmarks of common stock performance provided by the CRSP at the University of Chicago Booth School of Business. The CRSP universe includes common stocks listed on the NYSE, AMEX, and the NASDAQ National Market excluding the following: preferred stocks, unit investment trusts, closed-end funds, real estate investment trusts, Americus Trusts, foreign stocks and American Depositary Receipts. This material is has been prepared for designated institutional and professionals investors. The views expressed herein are subject to change based on market and other conditions. This material is for general information purposes only and is not intended as investment advice, as an offer or solicitation of an offer to sell or buy, or as an endorsement, recommendation, or sponsorship of any company, security, advisory service, or fund. The contents are current at the time of writing. Although the information contained herein has been obtained from sources believed to be reliable, its accuracy and completeness cannot be guaranteed. This information should not be used as the sole basis for investment decisions. Past performance is no guarantee of future results. 7
Intech is a specialized global asset management firm that harnesses stock price volatility as a source of excess return and a key to risk control. Founded in 1987 in Princeton, NJ by pioneering mathematician Dr. E. Robert Fernholz, Intech serves institutional investors across five continents, delivering relative return, low volatility, adaptive volatility and absolute return investment solutions. 525 Okeechobee Boulevard, Suite 1800 West Palm Beach, FL 33401 (561) 775-1100 intechinvestments.com C-0418-1945 03-31-19 FOR INSTITUTIONAL INVESTOR USE ONLY