Subsequent to a global financial crisis, trend-following performance tended to be weak for four years on average. This lack of time-series return predictability reduces the opportunity for trend-following to generate returns.
Comparing the performance of crisis and no-crisis periods, the average return (4.0 percent) in the 24 months following the start of a crisis was less than one-third the return (13.6 percent) earned in no-crisis periods. Performance in the 48 months following the start of a crisis (6.0 percent) was well under half the return (14.9 percent) in no-crisis periods.
Results were consistent across stocks, bonds and currencies. The exception was commodities, where returns were of similar magnitude in both pre-crisis and post-crisis periods.
A similar effect appeared when examining portfolios formed of local assets during regional financial crises.
The authors noted that behavioral models link momentum to investor overconfidence and decreasing risk aversion, with both leading to return predictability in asset prices. Under these models, overconfidence should fall and risk aversion should increase following market declines, so it seems logical that return predictability would drop after a financial crisis. It is also important to note, as the authors did, that “governments have an increased tendency to intervene in financial markets during crises, resulting in discontinuities in price patterns.” Such interventions can lead to sharp reversals, with negative consequences for trend-following strategies.
Hutchinson and O’Brien concluded that the performance of trend-following strategies is “much weaker in crisis periods, where performance can be as little as one-third of that in normal market conditions.” They continue, writing: “This result is supported by our evidence for regional crises, though the effect seems to be more short-lived. In our analysis of the underlying markets, our empirical evidence indicates a breakdown in the time series predictability, pervasive in normal market conditions, on which trend following relies.”
Summary
As an investment style, trend-following has existed for a long time. Data from the aforementioned research provides strong out-of-sample evidence beyond the substantial evidence that already existed in the literature. It also provides consistent, long-term evidence that trends have been pervasive features of global stock, bond, commodity and currency markets.
Regarding whether we should expect trends to continue, as AQR’s Hurst, Ooi and Pedersen concluded, “the most likely candidates to explain why markets have tended to trend more often than not include investors’ behavioral biases, market frictions, hedging demands, and market interventions by central banks and governments. Such market interventions and hedging programs are still prevalent, and investors are likely to continue to suffer from the same behavioral biases that have influenced price behavior over the past century, setting the stage for trend-following investing going forward.”
Implementing the Strategy
To gain access to the time-series momentum factor, our preferred vehicles are two funds managed by AQR Capital Management: the AQR Managed Futures Fund (AQMRX) and the AQR Managed Futures High Volatility Fund (QMHRX). As with the AQR Style Premia Alternative Fund, the managed futures funds provide exposure to time-series momentum across four asset classes: equity indices, bonds, commodities and currencies. These funds diversify risk through exposure to more than 100 liquid futures and forward contracts.
AQMRX has an expected return similar to U.S. equities and is managed to a 10 percent volatility level. QMHRX is managed to a 15 percent volatility level, with proportionally higher forward-looking return expectations. Their expense ratios are 1.15 percent and 1.58 percent, respectively.
Chapter 10:
How Much to Allocate to Alternatives?
Just as there is no right answer to how much an investor should “tilt” their portfolio to factors such as size and value, there is no right answer to how much an investor should allocate to alternatives. With that said, we believe there is a good way to think about the question. We begin with the basic assumption that, while markets are not perfectly efficient, they are highly efficient. In efficient markets, all risky assets should deliver similar risk-adjusted returns. Because this is the case, all else being equal, investors should not prefer one investment risk to another. (Note that all else is not always equal. For example, if your labor capital correlates highly with an investment risk, you should consider underweighting that risk in your portfolio.) That leads us to conclude a risk parity type of portfolio is a good starting point. Current stock valuations play an important role in determining future returns. Thus, they (not historical returns) should be used to help determine a portfolio’s expected return. The Shiller CAPE 10 is a useful metric to help estimate future returns.
Risk Parity
The risk parity approach to asset allocation focuses on the amount of risk in each portfolio component rather than the specific dollar amount invested in each portfolio component. In other words, it emphasizes not the allocation of capital (like traditional models), but on the allocation of risk. Consider the following example.
While the typical 60 percent stock/40 percent bond portfolio has 60 percent of its dollars allocated to equities, because stocks are so much more volatile than safe Treasury bonds, about 80 percent to 90 percent of the portfolio’s risk is equity risk. Broadly diversified equity portfolios historically have experienced volatility (as measured by the annual standard deviation of returns) of about 20 percent. This compares to volatility of about 5 percent for a high-quality intermediate-term bond portfolio with an average maturity of five years. Thus, if we consider how much risk (as measured by volatility) we have allocated to stocks and bonds, we see the following:
Equity Risk: 60 x 20 = 1,200
Bond Risk: 40 x 5 = 200
Total Risk: 1,200 + 200 = 1,400
Percentage Equity Risk: 1,200/1,400 = 86 percent
As you recall from our discussion of how to build more efficient portfolios, adding more exposure to the size and value factors allows investors to lower their overall allocation to stocks (and, thus, market beta) and increase their allocation to safer bonds. This is possible because the stocks they do own have higher expected returns than the market. The result is more of a risk parity portfolio. Let’s explore another example to illustrate this point.
Once again, Portfolio A is the typical 60 percent stock/40 percent bond portfolio. Portfolio B reduces its stock allocation to 40 percent, but substitutes the DFA U.S. Small Cap Value Portfolio (DFSVX) for the Vanguard Total (U.S.) Stock Market Index Fund (VTSMX). The increased exposure to the size and value factors, with their higher expected returns, allows us to increase its exposure to safe bonds from 40 percent to 60 percent. Our example covers the period from April 1993 (chosen because it is the inception date of DFA’s small-cap value fund) through December 2016.
Portfolio A: 60 percent Vanguard Total (U.S.) Stock Market Index Fund (VTSMX)/40 percent Vanguard Intermediate-Term Treasury Fund (VFITX)
Portfolio B: 40 percent DFA U.S. Small Cap Value Portfolio (DFSVX)/60 percent Vanguard Intermediate-Term Treasury Fund (VFITX)
The following table presents Portfolio A and Portfolio B’s exposure to the market beta, size, value, quality (discussed in Appendix B) and term factors. The figures inside the parentheses indicate the utilized mutual fund’s exposure to (or loading on) each factor. To calculate the portfolio’s factor exposure, the figures found to the left of the parentheses, multiply the fund’s loading by the allocation percentage. For example, over the sample period, DFSVX had a loading of 0.65 on the value factor. Portfolio B’s 40 percent allocation to DFSVX results in a value factor loading for the entire portfolio of 0.26 (40 percent x 0.65). Data comes from the regression tool available at the website www.portfoliovisualizer.com.
April 1993–December 2016
As you can see, while the two portfolios had relatively similar returns and volatility, Portfolio B was more efficient in both respects. In addition, Portfolio A had a 0.60 loading on marke
t beta, just a 0.17 loading on the term factor, and negligible exposure to the other factors. Portfolio B was far more diversified, with relatively more equal weightings on the other factors. Portfolios can also be structured to gain exposure to the momentum factor, though neither of these portfolios are. A portfolio with even greater risk parity can be constructed through the use of long-short funds.
The benefits of diversifying across factors and creating a portfolio with greater risk parity are apparent in the following table. The data for the market beta, size, value and momentum factors covers the period from 1927 through 2016. We have included two additional factors, profitability (data begins in 1964) and quality (data begins in 1958), both of which are discussed in Appendix B. The table presents each factor’s premium, volatility and Sharpe ratio. It also shows the same information for three naïve 1/N portfolios. Portfolio 1 (P1) is allocated 25 percent to each of four factors: market beta, size, value and momentum. Portfolio 2 (P2) is allocated 20 percent to each of the same four factors, but adds a 20 percent allocation to the profitability factor. Portfolio 3 (P3) is allocated the same way, but substitutes the quality factor for the profitability factor.
Factor Diversification 1927–2016
As you review the data in the preceding table, note that the three portfolios’ Sharpe ratios are dramatically higher than the Sharpe ratios of any individual factor. This is a direct result of each factor’s low correlation with the others, and demonstrates the benefits of diversifying across unique sources of risk. The table below shows the annual correlations of the market beta, size, value, momentum, profitability and quality factors over the period from 1964 through 2016.
Annual Correlations 1964–2016
Note that with the sole exception of the high correlation between the related profitability and quality factors (which we should expect because profitability is one of the traits of quality companies), correlations are low to negative. Notice in particular the negative correlations of the momentum premium to the beta, size and value premiums. This demonstrates the diversification benefit of adding momentum factor exposure to a portfolio that incorporates these other factors.
We can also see the benefits of diversifying across factors in the following table, which shows the odds of underperformance (that is, the odds of producing a negative return) over various time horizons.
Odds of Underperformance 1927–2016
In each case, the longer the horizon, the lower the odds of a negative result become. However, no matter how long the horizon, each of the individual factors still experienced some periods of underperformance, even at 20 years. The sole exception is momentum at the 20-year period. Even this outcome, though, does not guarantee future success for momentum at that horizon.
The key takeaway from the tables and data in this chapter is that diversification across unique sources of risk and return has resulted in more efficient portfolios. That, or course, brings us to the alternative investments we have discussed, and the impact of an allocation to them.
Adding Alternatives
Like with factors, adding alternatives that represent unique sources of risk and return and that offer equity-like expected returns should allow us to lower a portfolio’s allocation to market beta (the riskiest factor), thus creating portfolios with even greater risk parity and even more diversified sources of risk. Because each of the alternatives we have discussed shows low correlations to other portfolio assets, we should end up with a more efficient portfolio—one with similar returns but less risk.
How Much to Allocate to Alternatives?
We offer this suggestion: Investors should consider an allocation to the alternatives we recommend of at least a 10 percent and up to 30 percent. With that said, because most alternatives are not tax efficient, many investors may be constrained in their ability to include them in their portfolios due to limited capacity in tax-advantaged accounts.
You can use the following hypothetical, which is based on the concept of risk parity and the hypothesis that all risky assets should have similar risk-adjusted returns, as a starting point when contemplating a specific allocation to our recommended alternatives. Because all risky assets should earn similar risk-adjusted returns, investors should not prefer one alternative over another. As a result, assume you have decided to include an allocation to four of the alternatives we have discussed: alternative lending, reinsurance, the variance risk premium and the AQR Style Premia Alternative Fund. You might choose to commit one-fourth of your total alternatives allocation to each of the four strategies. If you would then also like to include an allocation to time-series momentum using AQR’s managed futures funds, consider this: Instead of thinking about it as a fifth alternative, consider it a fifth factor beyond the four to which the AQR Style Premia Alternative Fund already provides exposure. Thus, if you had decided that you wanted a 25 percent allocation to alternatives, you might allot 6 percent to each of the alternative lending, reinsurance and variance risk premium strategies. You would next include a 4.8 percent allocation (0.8 x 6 percent) to the AQR Style Premia Alternative Fund and a 1.2 percent allocation (0.2 x 6 percent) to a managed futures strategy.
From Where Should the Allocation to Alternatives Come?
Whenever we add an investment to a portfolio, it must reduce the allocation to other assets. The question that arises is whether an allocation to alternative strategies should come from stocks or bonds. Because the expected return to each of the alternatives we have considered is equity-like (about 7 percent), and given that their volatility is much lower (about one-quarter to one-half the 20 percent volatility of equities), investors with relatively high equity allocations (more than 60 percent) should consider taking an allocation to alternatives from their equity holdings. On the other hand, because the volatility of an equal-weighted portfolio of alternatives is expected to be about 6 percent (compared to about 5 percent for an intermediate-term, high-quality bond strategy), and given their equity-like expected return, investors with relatively low equity allocations (less than 40 percent) should consider taking an allocation to alternatives from their bond holdings. If you decide to do the latter, it is important to recognize that in return for significantly higher expected return, the left tail risk of your portfolio has now increased. Investors with equity allocations between 40 percent and 60 percent can consider taking the allocation pro rata from the two.
Conclusion
We hope you have found your journey with us both informative and of great value. After all, the holy grail of investing is the search for investment strategies that can deliver higher expected returns without increased risk, or the same expected return with reduced risk. We have attempted to give you the road map to it. Clearly, the way evolved since the publication of the original version of this book in 2014. New alternatives have become available, allowing us to create even more efficient strategies that improve the odds of achieving your financial goals while also greatly reducing the risk of outliving your assets. The road map comes with the following “directions.”
Current stock valuations play an important role in determining future returns. Thus, they (not historical returns) should be used to help determine a portfolio’s expected return. The Shiller CAPE 10 is a useful metric to help estimate future returns.
Consider the expected return only the mean of a wide dispersion of potential returns. Your plan should incorporate options (such as staying in the workforce longer, moving to a location with a lower cost of living, and so on) that you will adopt to minimize the risk of failure, regardless of which potential outcome becomes reality.
Think about diversification in terms of the factors (rather than asset classes) that explain returns, diversifying risk across those factors as well as the other unique sources of risk we have discussed.
To the extent you are willing to accept the risk of tracking error regret, concentrate the equity portion of your portfolio in the highest expected returning factors. All else equal (for example, expense ratios), use funds that have
the highest loadings on (exposure to) the factors in which you want to invest. That allows you to minimize your exposure to market beta, which is the biggest risk factor.
Diversify your portfolio across the globe. For example, the portfolio’s equity portion might have a 50 percent allocation to U.S. small value stocks, a 37.5 percent allocation to international developed markets small value stocks and a 12.5 percent allocation to emerging markets value stocks. These percentages reflect current geographical market weightings.
We also offer an important caution. Just as we can only estimate the stock market’s future expected return, we can only estimate the future return to small value stocks. History and current valuations provide a guide, helping us make estimates. However, all crystal balls are cloudy—there are no guarantees. What we do know is that a low-market-beta/high-tilt portfolio does reduce the risk of fat tails (both good and bad). We cannot guarantee, though, that it will produce the same return as a more market-like portfolio with a higher equity allocation.
Your journey with us is not quite over. The book’s appendices address the following six important topics:
The use of Monte Carlo simulations in determining your asset allocation.
Since Eugene Fama and Kenneth French’s work on their three-factor model, academics have “discovered” other factors that not only help explain the differences in returns of diversified portfolios, but also carry premiums. Among the additional factors we believe investors should consider are cross-sectional momentum, profitability, quality and carry (one of the four factors in the AQR Style Premia Alternative Fund).
Reducing the Risk of Black Swans Page 10