This post is available as a PDF download here.

Summary­

  • While investors are often concerned about catastrophic risks, failing to allocate enough to risky assets can lead investors to “fail slowly” by not maintaining pace with inflation or supporting withdrawal rates.
  • Historically, bonds have acted as the primary means of managing risk.However, historical evidence suggests that investors may carry around a significant allocation to fixed income only to offset the tail risks of a few bad years in equities.
  • Going forward, maintaining a large, static allocation to fixed income may represent a significant opportunity cost for investors.
  • Trend following strategies have historically demonstrated the ability to significantly reduce downside risk, though often give up exposure to the best performing years as well.
  • Despite reducing upside capture, trend following strategies may represent a beneficial diversifier for conservative portfolios going forward, potentially allowing investors to more fully participate with equity market growth without necessarily fully exposing themselves to equity market risk.

In our recent commentary Failing Slow, Failing Fast, and Failing Very Fast, we re-introduced the idea of “risk ignition,” a phrase we first read in Aaron Brown’s book Red Blooded Risk.  To quote the book on the core concept of the idea,

Taking less risk than is optimal is not safer; it just locks in a worse outcome. Taking more risk than is optimal also results in a worse outcome, and often leads to complete disaster.

Risk ignition is about taking sufficient risk to promote growth, but not so much risk as to create a high probability of catastrophe.

Traditionally, financial planners have tried to find the balance of risk in the intersection of an investor’s tolerance for risk and their capacity to bear it.  The former addresses the investor’s personal preferences while the latter addresses their financial requirements.

What capacity fails to capture, in our opinion, is an investor’s need to take risk.  It would be difficult to make the argument that a recent retiree with $1,000,000 saved and a planned 4% inflation-adjusted withdrawal rate should ever be allocated to 100% fixed income in the current interest rate environment, no matter what his risk tolerance is.  Bearing too little risk is precisely how investors end up failing slowly.

The simple fact is that earning a return above the risk-free rate requires bearing risk.  It is why, after all, the excess annualized return that equities earn is known as the “equity risk premium.”  Emphasis on the “risk premium” part.

As more and more Baby Boomers retire, prevailing low interest rates mean that traditionally allocated conservative portfolios may no longer offer enough upside to address longevity risk. However, blindly moving these investors into riskier profiles (which may very well be above their risk tolerance anyway) may be equally imprudent, as higher portfolio volatility increases sensitivity to sequence risk when an investor begins taking distributions.

This is where we believe that tactical strategies can play an important role.

Holding Bonds for Insurance

In the simplest asset allocation framework, investors balance their desire to pursue growth with their tolerance (and even capacity) for risk by blending stocks and bonds.  More conservative investors tend to hold a larger proportion of fixed income instruments, preferring their defined cash flows and maturity dates, while growth investors tilt more heavily towards equities.  Stocks fight the risk of lost purchasing power (i.e. inflation) while bonds fight the risk of capital loss.

The blend between equities and bonds will ultimately be determined by balancing exposure to these two risks.

But why not simply hold just stocks?  A trivial question, but one worth acknowledging.  The answer is found in the graph below, where we plot the distribution fitting the annual returns of a broad U.S. equity index from 1962 to 2017.  What we see is a large negative skew, which implies that the left tail of the distribution is much larger than the right.  In plain English: every once in a while, stocks crash. Hard.

Source: Kenneth French Data Library.  Calculations by Newfound Research.  Returns are gross of all fees, including transaction fees, taxes, and any management fees.  Returns assume the reinvestment of all distributions.  Past performance is not a guarantee of future results.

The large left tail implies a drawdown risk that investors with short time horizons, or who are currently taking distributions from their portfolios, may not be able to bear.  This is evident by plotting the realized excess return of different stock / bond[1] mixes versus their respective realized volatility profiles.  We can see that volatility is largely driven by the equity allocation in the portfolio.

This left tail, and long-term equity realized equity volatility in general, is driven by just a few outlier events.  To demonstrate, we will remove the worst performing years for U.S. equities from the dataset.  For the sake of fairness, we’ll also drop an equal number of best years (acknowledging that the best years often follow the worse, and vice versa). Despite losing the best years, the worst years are so bad that we still see a tremendous shift up-and-to-the-left in the realized frontier, indicating higher realized returns with less risk.

Consider that the Sharpe optimal portfolio moves from the 50% stocks / 50% bonds mixture when the full data set is used to an 80% stock / 20% bond split when the best and worst three years are dropped.

Source: Kenneth French Data Library.  Calculations by Newfound Research.  Returns are gross of all fees, including transaction fees, taxes, and any management fees.  Returns assume the reinvestment of all distributions.  Past performance is not a guarantee of future results.

Note that in the full-sample frontier, achieving a long-term annualized volatility of 10% requires holding somewhere between 40-50% of our portfolio in 10-year U.S. Treasuries.  When we drop the best and worst 3 years of equity returns, the same risk level can be achieved with just a 20-30% allocation to bonds.

If we go so far as to drop the best and worst five years?  We would only need 10% of our portfolio in bonds to hit that long-term volatility target.

One interpretation of this data is that investors carry a very significant allocation to bonds in their portfolio simply in effort to hedge the left-tail risks of equities.  For a “balanced” investor (i.e. one around the 10% volatility level of a 60/40 portfolio), the worst three years of equity returns increases the recommended allocation to bonds by 20-30%!

Why is this important?  Consider that forward bond forecasts heavily rely on current interest rates.  Despite the recent increase in the short-end of the U.S. Treasury yield curve, intermediate term rates remain well-below long-term averages.  This has two major implications:

  • If a bear market were to emerge, bonds may not provide the same protection they did in prior bear environments. (See our commentary Bond Returns: Don’t Be Jealous, Be Worried)
  • The opportunity cost for holding bonds versus equities may be quite elevated (if the term premium has eroded while the equity risk premium has remained constant).

Enter trend following.

Cutting the Tails with Trend Following

At its simplest, trend following says to remain invested while an investment is still appreciating in value and divest (or, potentially, even short) when an investment begins to depreciate.

(Since we’ve written at length about trend following in the past, we’ll spare the details in this commentary.  For those keen on learning more about the history and theory of trend following, we would recommend our commentaries Two Centuries of Momentum and Protect and Participate: Managing Drawdowns with Trend Following.)

How, exactly, trend is measured is part of the art. The science, however, largely remains the same: trend following has a long, documented trail of empirical evidence suggesting that it may be an effective means of reducing drawdown risk in a variety of asset classes around the globe.

We can see in the example below that trend following applied to U.S. equities over the last 50+ years is no exception.

(In this example, we have applied a simple price-minus-moving-average trend following strategy.  When price is above the 200-day moving average, we invest in broad U.S. equities.  When price falls below the 200-day moving average, we divest into the risk-free asset. The model is evaluated daily after market close and trades are assumed to be executed at the close of the following day.)

 

Source: Kenneth French Data Library and Federal Reserve of St. Louis. Calculations by Newfound Research. Returns are gross of all fees, including transaction fees, taxes, and any management fees.  Returns assume the reinvestment of all distributions.  Past performance is not a guarantee of future results. 

While the long-term equity curve tells part of the story – nearly matching long-term returns while avoiding many of the deepest – we believe that a more nuanced conversation can be had by looking at the joint distribution of annual returns between U.S. equities and the trend following strategy.

Source: Kenneth French Data Library.  Calculations by Newfound Research.  Scatter plot shows the joint distribution of annual returns from 1962 to 2017 for a broad U.S. equity index and a trend following strategy.  Returns are gross of all fees, including transaction fees, taxes, and any management fees.  Returns assume the reinvestment of all distributions.  Past performance is not a guarantee of future results.

We can see that when U.S. equity returns are positive, the trend following strategy tends to have positive returns as well (albeit slightly lower ones).  When returns are near zero, the trend following strategy has slightly negative returns.  And when U.S. equity returns are highly negative, the trend following strategy significantly limits these returns.

In many ways, one might argue that the return profile of a trend following strategy mirrors that of a long call option (or, alternatively, index plus a long put option).  The strategy has historically offered protection against large drawdowns, but there is a “premium” that is paid in the form of whipsaw.

We can also see this by plotting the annual return distribution of U.S. equities with the distribution of the trend strategy superimposed on top.

Source: Kenneth French Data Library.  Calculations by Newfound Research.  Returns are gross of all fees, including transaction fees, taxes, and any management fees.  Returns assume the reinvestment of all distributions.  Past performance is not a guarantee of future results.

The trend strategy exhibits significantly less skew than U.S. equities, but loses exposure in both tails.  This means that while trend following has historically been able to reduce exposure to significant losses, it has also meant giving up the significant gains.  This makes sense, as many of the market’s best years come off the heels of the worst, when trend following may be slower to reinvest.

In fact, we can see that as we cut off the best and worst years, the distribution of equity returns converges upon the distribution of the trend following strategy.

Source: Kenneth French Data Library.  Calculations by Newfound Research.  Returns are gross of all fees, including transaction fees, taxes, and any management fees.  Returns assume the reinvestment of all distributions.  Past performance is not a guarantee of future results.

Our earlier analysis of changes to the realized efficient frontier when the best and worst years are dropped indicates that the return profile of trend following may be of significant benefit to investors.  Specifically, conservative investors may be able to hold a larger allocation to trend following than naked equities.  This allows them to tilt their exposure towards equities in positive trending periods without necessarily invoking a greater level of portfolio volatility and drawdown due to the negative skew equities exhibit.

In the table below, we find the optimal mix of stocks, bonds, and the trend strategy that would have maximized excess annualized return for the same level of volatility of a given stock/bond blend.

 TargetU.S. Equities10-Year Treasury IndexTrend Strategy
0/1007.4%34.7%58.0%
10/909.7%48.4%41.9%
20/8011.5%59.5%29.0%
30/7010.9%56.4%32.7%
40/608.9%43.8%47.3%
50/506.6%29.9%63.6%
60/4037.2%25.0%37.8%
70/3045.4%14.0%40.7%
80/2053.9%3.1%43.1%
90/1075.9%0.0%24.1%
100/0100.0%0.0%0.0%

We can see that across the board, the optimal portfolio would have had a significant allocation to the trend following strategy. Below, we plot excess annualized return versus volatility for each of these portfolios (in orange) as well as the target mixes (in blue).

In all but the most aggressive cases (where trend following simply was not volatile enough to match the required volatility of the benchmark allocation), trend following creates a lift in excess annualized return.  This is because trend following has historically allowed investors to simultaneously decrease overall portfolio risk in negative trending environments and increaseexposure to equities in positive trending ones.

Consider, for example, the optimal mixture that targets the same risk profile of the 30/70 stock/bond blend.  The portfolio holds 9.7% in stocks, 48.4% in bonds and 41.9% in the trend strategy.  This means that in years where stocks are exhibiting a positive trend, the portfolio is a near 50/50 stock/bond split.  In years where stocks are exhibiting a negative trend, the portfolio tilts towards a 10/90 split.  Trend following allows the portfolio to both be far more aggressive as well as far more defensive than the static benchmark.

Used in this manner, even if the trend following strategy underperforms stocks in positive trending years, so long as it outperforms bonds, it can add value in the context of the overall portfolio! While bonds have, historically, acted as a static insurance policy, trend following acts in a far more dynamic capacity, allowing investors to try to maximize their exposure to the equity risk premium.

Source: Kenneth French Data Library and Federal Reserve of St. Louis. Calculations by Newfound Research. Returns are gross of all fees, including transaction fees, taxes, and any management fees.  Returns assume the reinvestment of all distributions.  Past performance is not a guarantee of future results.

Conclusion

Historically, stocks and bonds have acted as the building blocks of asset allocation.  Investors pursuing a growth mandate have tilted towards stocks, while those focused on capital preservation have tilted more heavily towards bonds.

For conservative investors, the need to employ a large bond position is mainly driven by the negative skew exhibited by equity returns.  However, this means that investors are significantly under-allocated to equities, and therefore sacrifice significant growth potential, during non-volatile years.

With low forecasted returns in fixed income, the significant allocation to bonds carried around by most conservative investors may represent a significant opportunity cost, heightening the risk offailing slow.

Trend following strategies, however, offer a simple alternative.  The return profile of these strategies has historically mimicked that of a call option: meaningful upside participation with limited downside exposure.  While not contractually guaranteed, this dynamic exposure may offer investors a way to reduce their allocation to fixed income without necessarily increasing their exposure to left-tail equity risk.

 


 

[1]  We use a constant maturity 10-year U.S. Treasury index for bonds.

Corey is co-founder and Chief Investment Officer of Newfound Research, a quantitative asset manager offering a suite of separately managed accounts and mutual funds. At Newfound, Corey is responsible for portfolio management, investment research, strategy development, and communication of the firm's views to clients.

Prior to offering asset management services, Newfound licensed research from the quantitative investment models developed by Corey. At peak, this research helped steer the tactical allocation decisions for upwards of $10bn.

Corey is a frequent speaker on industry panels and contributes to ETF.com, ETF Trends, and Forbes.com’s Great Speculations blog. He was named a 2014 ETF All Star by ETF.com.

Corey holds a Master of Science in Computational Finance from Carnegie Mellon University and a Bachelor of Science in Computer Science, cum laude, from Cornell University.

You can connect with Corey on LinkedIn or Twitter.

Or schedule a time to connect.