This post is available as a PDF download here.

Summary­

• We typically discuss trend following in the context of risk management for investors looking to diversify their diversifiers.
• While we believe that trend following is most appropriate for investors concerned about sequence risk, levered trend following may have use for investors pursuing growth.
• In a simple back-test, a naïve levered trend following considerably increases annualized returns and reduces negative skew and kurtosis (“fat tails”).
• The introduced leverage, however, significantly increases annualized volatility, meaning that the strategy still exhibits significant and large drawdown profiles.
• Nevertheless, trend following may be a way to allow for the incorporation of leverage with reduced risk of permanent portfolio impairment that would otherwise occur from large drawdowns.

In an industry obsessed with alpha, our view here at Newfound has long been to take a risk-first approach to investing.  In light of this, when we discuss trend following techniques, it is often with an eye towards explicitly managing drawdowns.  Our aim is to help investors diversify their diversifiers and better manage the potentially devastation that sequence risk can wreak upon their portfolios.

Thus, we often discuss the application of trend following for soon-to-be and recent retirees who are in peak sequence risk years.

• Empirical evidence suggests that trend following can be a highly effective means of limiting exposure to significant and prolonged drawdowns.
• Trend following is complementary to other diversifiers like fixed income, which can theoretically increase the Sharpe ratio of the diversification bucket as a whole.
• Instead of acting as a static hedge, the dynamic approach of trend following can also help investors take advantage of market tailwinds. This may be particularly important if real interest rates remain low.
• The potential tax inefficiency of trend following is significantly lower when the alternative risk management technique is fixed income.

Despite our focus on using trend following to manage sequence risk, we often receive questions from investors still within their accumulation phase asking whether trend following can be appropriate for them as well.  Most frequently, the question is, “If trend following can manage downside risk, can I use a levered approach to trend following in hopes of boosting returns?”

This commentary explores that idea, specifically in the context of available levered ETFs.

Does Naïve Levered Trend Following Work?

In an effort to avoid overfitting our results to any one particular model or parameterization of trend following, we have constructed our signals employing a model-of-models approach [1] Specifically, we use four different definitions of trend for a given N-period lookback:

• Price-Minus-Moving-Average: When price is above its N-period simple moving average, invest.Otherwise, divest.
• EWMA Cross-Over: When the (N/4)-length exponentially-weighted moving average is above the (N/2)-length exponentially-weighted moving average, invest.Otherwise, divest.
• EWMA Slope: When the (N/2)-length exponentially-weighted moving average is positively sloped, invest. Otherwise, divest.
• Percentile Channel: When price crossed above the trailing 75^thpercentile over the prior N-periods, invest. Stay invested until it crosses below its trailing 25^thpercentile over the prior N-periods.  Stay divested until it crosses back above the 75^th

For each of these four models, we also run a number of parameterizations covering 6-to-18-month lookbacks.  In grand total, there are 4 models with 5 parameterizations each, giving us 30 variations of trend signals.

Using these signals, we construct three models. In the first model, we simply invest in U.S. equities in proportion to the number of signals that are positive. For example, if 75% of the trend following signals are positive, the portfolio is 75% invested in U.S. equities and 25% in the risk-free asset.

For our leveraged model, we simply multiply the percentage of signals by 2x and invest that proportion of our portfolio in U.S. equities and the remainder in the risk-free asset.  In those cases where the amount invested in U.S. equities exceeds 100% of the portfolio, we assume a negative allocation to the risk-free asset (e.g. if we invest 150% of our assets in U.S. equities, we assume a -50% allocation to the risk-free asset).

With the benefit of hindsight, we should not be surprised at the results.  If we know that trend following is effective at limiting severe and prolonged drawdowns (the kryptonite to levered investors), then it should come as no surprise that a levered trend following strategy does quite well.

It is well worth pointing out, however, that a highly levered strategy can be quickly wiped out by a sudden and immediate drawdown that trend following is unable to sidestep.  Assuming a 2x levered position, our portfolio would be quickly wiped out by a sharp 50% correction.  While such an event did not happen during the 1900s for U.S. equities, that does not mean it cannot happen in the future.  Caveat emptor.

Logarithmically-plotted equity curves can be deceiving, so it is important that we also compare the annual return characteristics.

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.

While we can see that a simple trend following approach effectively “clips” the tails of the underlying distribution – giving up both the best and the worst annual returns – the levered strategy still has significant mass in both directions.  Evaluating the first several moments of the distributions, however, we see that both simple and levered trend following significantly reduce the skew and kurtosis of the return distribution.

Nevertheless, the standard deviation of the levered trend following strategy exceeds even that of the underlying asset, a potential indication that expectations for the approach may be less about, “Can I avoid large drawdowns?” and more about, “Can I use leverage for growth and still avoid catastrophe?”  We can see this by plotting the joint annual log-return distributions.

We can see that for U.S. equity returns between 0% and -20%, the Levered Trend Following strategy can exhibit returns between -20% and -40%.  About 11% of the observations fall into this category, making it an occurrence that a levered trend follower should expect to experience multiple times in their investment lifecycle.  We can even see one year where U.S. equities are slightly positive and the levered model exhibits a near -30% return.  It is in the most extreme U.S. equity years – those exceeding -20% – that the trend following aspect appears to come into play.

We must also ask the question, “can this idea survive associated fees?”  If investors are looking to apply this approach using levered ETFs, they must consider the expense ratios of the ETFs themselves, transaction costs, and bid/ask spreads.  Here we will use the ProShares Ultra S&P 500 ETF (“SSO”) as a data proxy.  The expense ratio is 0.90% and the average bid/ask spread is 0.03%.  Since transactions costs vary, we will assume an added annual 0.20% fee for asset-based pricing.

In comparison, for the naïve model, we will use the SPDR S&P 500 ETF (“SPY”) as the data proxy and assume an expense ratio of 0.09% and an average bid/ask spread 0.004%.  Since most platforms have a vanilla S&P 500 ETF on their no-transaction fee list, we will not add any explicit transaction costs.

We plot the strategy equity curves below net of these assumed fees.

The annualized return for the Levered Trend Following strategy declines from 15.9% to 14.5%, while the unlevered version only falls from 10.1% to 10.0%.  While the overall return of the levered version declines by 140 basis points per year, it still far exceeds the total return performance of the unlevered version. 

Conclusion

Based upon this initial analysis, it would appear that a simple, levered trend following approach may be worth further consideration for investors in the accumulation phase of their investment lifecycle.

Do-it-yourself investors may have no problem implementing this idea on their own using levered ETFs, but other investors may prefer a simple, packaged approach.  Unfortunately, as far as we are aware, no such packaged product exists in the marketplace today.

However, one workaround may be to utilize levered ETFs to “make room” for an unlevered trend following strategy.  For example, if a growth-oriented investor currently holdings an 80/20 stock/bond mix and wanted to introduce a 20% allocation to trend following, they could re-orient their portfolio to be 60% stocks, 10% 2x levered stocks, 10% 2x levered bonds, and 20% trend following.  This would have the effect of being an 80/20 stock/bond portfolio with 20% leverage applied to introduce the trend following strategy.  While there are the nuances of daily reset to consider in the levered ETF solutions, this approach may allow for the modest introduction of levered trend following into the portfolio.

It is worth noting that while we employed up to 2x leverage in this commentary, there is no reason investors could not apply a lower amount, either by mixing levered and unlevered ETFs, or by using a solution like the new Portfolio+ line-up from Direxion, which applies 1.25x leverage to underlying indices.

As we like to say here at Newfound, “risk cannot be destroyed, only transformed.” While this commentary explored levered trend following in comparison to unlevered exposure, a more apt comparison might simply be to levered market exposure.  We suspect that the trend following overlay creates the same transformation: a reduction of the best and worst years at the cost of whipsaw. However, the introduction of leverage further heightens the risk of sudden and immediate drawdowns: the exact loss profile trend following is ill-suited to avoid.

[1] Nothing in this commentary reflects an actual investment strategy or model managed by Newfound and any investment strategies or investment approaches reflected herein are constructed solely for purposes of analyzing and evaluating the topics herein.

Corey Hoffstein

Corey is co-founder and Chief Investment Officer of Newfound Research. 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.