What is the best “Risk-Off” Asset for Trend-Followers?

/What is the best “Risk-Off” Asset for Trend-Followers?

What is the best “Risk-Off” Asset for Trend-Followers?

By | 2017-08-18T16:52:27+00:00 October 13th, 2016|Guest Posts, Tactical Asset Allocation Research|14 Comments

So you’re a trend-follower. Great.

But here is a question:

What do you invest in when your rules suggest “risk off?”

Many investors suggest low duration cash or t-bills. Seems reasonable.

But is it optimal?

Perhaps we should invest in longer duration risk-off assets like 10-yr bonds? We investigate these questions and come to the conclusion that keeping it simple is probably the best solution — dump “risk-off” assets into truly low risk assets like cash or t-bills.

A quick introduction to trend-following

Volatility, as measured in academia using standard deviation, is often synonymous with risk. Yet while volatility is an easily quantifiable number, it does have a number of drawbacks as a risk measure. One of the key assumptions in using volatility to define risk is that investors actually care about volatility. I take the view that this is only partially true. Volatility is a two-headed beast: Investors love volatile markets that move upwards but despise the value destruction of drawdowns. This duality presents the motivation for a tactical risk management overlay, which I’ll summarize into two key points:

  1. Capture the upside
  2. Limit the downside (reduce losses)

I say “reduce losses” recognizing that no strategy is perfect. But over the long run, some strategies can be effective at managing this dynamic. For example, the ten month simple moving average (SMA) strategy proposed by Faber1 and a twelve month time-series momentum strategy from Antonacci2 are two related methodologies that have been successful in the past. In both instances the authors have demonstrated the ability of these strategies to generate similar returns to that of a buy-and-hold strategy, but with substantially lower drawdowns. A prior post on this site combines both of these rules and argues that this approach creates an even more robust tail risk management approach. Nonetheless, when it comes to “market timing” constructs, nothing will ever work all the time and things that have worked in the past may never work in the future.

Although neither method is particularly complicated, the more straightforward of the two is the simple moving average strategy from Faber. It consists of the following two simple rules:

  1. Buy risk assets (stocks) when the monthly closing price is greater than the ten month simple moving average of their price.
  2. Sell stocks (and buy Treasury Bills) when the monthly closing price is below the ten month simple moving average of their price.

While the details of time-series momentum are slightly different, both of these methodologies are effectively absolute momentum strategies. They both seek to identify buy and sell signals based on recent past performance. A buy signal prompts an investor to purchase risk assets while a sell signal prescribes a move into a “risk-free” asset—typically Treasury Bills.

For the analysis below, I use a twelve month simple moving average (SMA) strategy based on the same two trading rules from Faber mentioned earlier. US Large Company Stocks represent the risk asset and 30 day Treasuries are the risk-free alternative. The below chart details the returns associated with Faber’s SMA strategy, as compared with a buy-and-hold strategy from 1927 through 2012:

US Stocks 12 Month Simple Moving Average

Source: Ibbotson3 The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index. Additional information regarding the construction of these results is available upon request.

The results show that a simple moving average strategy, gross of transaction fees and taxes, actually outperformed the simple buy-and-hold investor over the 86 year period. How much did T-Bills contribute to the overall return versus the SMA strategy alone (i.e., using cash as the alternative asset)? One way to assess the impact is to rerun these numbers with a hypothetical zero return asset.

US Stocks SMA with Zero Return

The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index. Additional information regarding the construction of these results is available upon request.

The simple moving average strategy on its own reduced the rate of return from 9.83% to 9.11%. While the results look less attractive, they don’t exactly disqualify the use of a simple moving average strategy. They do, however, emphasize the importance of a risk-free alternative that provides some return. Adding Treasury Bills as the risk free alternative provided a slight boost of just over 1% (compared to the scenario with the zero-return asset). In other words the simple moving average strategy avoided large drawdowns in stocks, and simultaneously generated some additional return through exposure to Treasury Bills.

What Happens if we Change the “Risk-Off” Asset?

But let’s consider what happens when we move along the duration spectrum. What if Treasuries with a longer maturity were used as the alternative asset instead of Treasury Bills? A quick substitution for 5 year Treasury Notes and 20 year Treasury Bonds produced the following results:

US Stocks SMA with Various Treasuries

The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index. Additional information regarding the construction of these results is available upon request.

On the surface, this would appear to be a spectacular finding—simply increasing the maturity of Treasuries would have generated an even higher return and maintained lower drawdowns!

But I’m a little skeptical.

I can’t help but think there’s more to this story than meets the eye. Consider first the 1981 to 2012 period—more commonly referred to as the bond bull market.

US Stocks SMA (Jan 1981 - Dec 2012)

The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index. Additional information regarding the construction of these results is available upon request.

The theme of higher returns from increasing maturity looks even stronger, and this shouldn’t come as a surprise. During a falling interest rate environment yields will decrease and bond prices will rise. The resulting capital appreciation provided increasing rates of return to fixed income investors.

But what’s more interesting is what happens during a rising rate environment when yields increase and bond prices fall. Here we look at the results from 1965 to 1980.

US Stocks SMA (Jan 1965 - Dec 1980)

The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index. Additional information regarding the construction of these results is available upon request.

We see that increasing the maturity of the alternative asset improved returns, but only to a certain point. The simple moving average strategy utilizing Treasury Bonds actually ended up performing worse than Notes, and with a substantially higher max drawdown compared to both Notes and Bills. Not scared yet? Consider the drawdowns that occurred on Treasuries:

Treasury Drawdowns (Jan 1965 - Dec 2012)

The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged, do not reflect management or trading fees, and one cannot invest directly in an index. Additional information regarding the construction of these results is available upon request.

The results above were filtered to show only the drawdowns on Treasuries when the SMA strategy dictated that Treasuries be held. The effect of stock drawdowns was eliminated to provide an apples-to-apples comparison. The tried and true short-term Treasury Bill held firm over the entire period while drawdowns on longer duration Notes and Bonds were significant.

Concluding points

This gets back to the idea of crisis alpha. It is often taken for granted that during times of distress, Treasuries will hold their value or even appreciate in price—everybody knows that Treasury bonds are “safe.” Bonds are an excellent diversifier, but history has shown that their perceived safety doesn’t always manifest, and in some cases the expected safety of Treasury Bonds may turn out to be dangerous. They can, and have, declined in price when panic sets in, especially during rising rate environments. With longer duration bonds, investors have greater exposure to changes in interest rates.

The conclusion here is fairly simple. An absolute momentum overlay can help reduce drawdowns. That’s what it’s intended to do. The alternative risk free asset may help provide a small amount of additional return gross of taxes and fees. Chasing higher rates of return by taking on interest rate risk may expose those using these strategies to further losses. The alternative risk free asset should be truly risk free, and that means sticking with the tried and true Treasury Bill.

References

  1. Faber, Mebane T. A Quantitative Approach to Tactical Asset Allocation. February 2013. http://papers.ssrn.com/sol3/papers.cfm?abstract_id=962461.
  2. Antonacci, Gary. Absolute Momentum: A Simple Rule-Based Strategy and Universal Trend-Following Overlay. April 2013. http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2244633.
  3. 2013 Ibbotson SBBI Classic Yearbook. Morningstar Inc. Chicago, IL. 2013.

  • The views and opinions expressed herein are those of the author and do not necessarily reflect the views of Alpha Architect, its affiliates or its employees. Our full disclosures are available here. Definitions of common statistics used in our analysis are available here (towards the bottom).
  • Join thousands of other readers and subscribe to our blog.
  • This site provides NO information on our value ETFs or our momentum ETFs. Please refer to this site.

Print Friendly, PDF & Email

About the Author:

Daniel Sotiroff
Daniel currently holds a B.S. in mechanical engineering and an M.S. in applied mechanics. He firmly believes in using an evidence/quantitative based approach to investment management. He currently is employed full time as an engineer in the manufacturing industry and manages his own money. More of his work related to investment management can be found at thepfengineer.com
  • Govind

    I don’t think your data supports your conclusion. It looks like 5 year Treasuries give a nice boost to the return and Sharpe ratio without much increase in downside exposure compared to the use of Treasury bills. If one is willing to assume a little credit risk by using 5 year corporate bonds or an aggregate bond portfolio that typically has an average duration of about 5 years, the results should be even better than with 5 year Treasuries.

  • Daniel Sotiroff

    Good eye, and I don’t disagree with you. The historical data does show that the 5 year Treasury has indeed boosted returns–both absolute and risk adjusted–over longer periods of time

    Time horizon plays a role as well as personal preference. As you indicated it’s about how much interest rate risk you’re willing to take on. My perspective was that the additional interest rate exposure creates the potential for higher drawdowns when the risk-off asset is held. Should the SMA signal a move to risk-on during one of these drawdowns there’s the potential for taking a haircut in order to buy back into the risky asset (a short-term loss). Here’s a few examples of when that occurred and the associated drawdown (forgive my formatting)

    30 DAY DD 5 YR DD 20 YR DD
    August 1994 0.0% -4.53% -8.86%
    January 1995 0.0% -4.72% -7.63%
    July 2009 0.0% -3.76% -13.48%

    The original tables suggest that over a long enough period of time you still benefit from the additional interest rate exposure. My personal preference is to err on the conservative side

  • Daniel Sotiroff
  • Jason Mann

    I’d add a simple rule. own 20yr treasuries *only if* they themselves are above their 12-month SMA, otherwise hold T-bills. I suspect you’ll get best of both worlds…

  • What happens if you expand the data set to include other countries (Japan, Germany, UK, etc)?

    Nick de Peyster
    http://undervaluestocks.info

  • Daniel Sotiroff

    That’s another option. In a similar fashion the IVY5/RAA portfolio does what you suggest with the 7-10 Gov’t Bond Index:
    https://alphaarchitect.com/2014/12/02/the-robust-asset-allocation-raa-solution/#gs.KDdo4yo

  • Peter Wang

    Gary Antonacci runs to aggregate bonds in his proprietary E-GEM model, which AlphaArchitect is now offering. Is that a bad idea?

  • GM

    Great post, thank you!
    Do you (or does anyone) know of any low cost t-bill etfs one could utilize?
    Kind regards,
    G

  • Matt Barlow

    Great post. Whilst it is not a risk-free asset, what happens when Gold/Gold ETFs are used as the safe-haven in the SMA strategy?

  • By deploying an alternating “risk-off” asset, like BIL / IEF or SHY / IEF, you let the model select the safety asset appropriate for the prevailing yield environment. This approach was demonstrated for Global Protective Momentum (GPM): http://seekingalpha.com/article/3985525-generalized-protective-momentum.

    The alternation between SHY (rising yields) and IEF (falling yields) for 1971-1981 is shown in GPM’s “Manhattan Allocation Diagram”: http://www.screencast.com/t/glJD9ta2WU

    The Profit Contribution diagram shows the importance of the short-term treasury fund SHY during these 11 years with overall rising rates: http://www.screencast.com/t/ZEpHcZJV6Zv

    The equity comparison chart paints the performance of GPM (with SHY/IEF in blue, SHY in green, IEF in red) against SPY (in black) for 1971-1981: http://www.screencast.com/t/wRiMIuZGlP5

  • Govind

    I understand how personal preferences play into your decision. But the drawdowns when holding risk-on assets are considerably higher than the -4.72% and -4.53% drawdowns when holding 5 yr Treasuries. Because of that, I would opt for the 1.45% higher overall CAGR from using 5 yr Treasuries rather than 30 day Bills.

  • Daniel Sotiroff

    I’m not familiar with the E-GEM model, so I won’t comment on it specifically. Generally speaking I don’t think there are “good” or “bad” ideas. What I think is really important with any strategy is to understand the risks that you are exposed to and how comfortable you are with those risks. These sorts of decisions are subjective and depend on personal preferences

  • Daniel Sotiroff

    I didn’t run the numbers, but I would expect similar results. If you look at Faber’s paper referenced above he includes an in depth discussion about applying the 10 month SMA to various asset classes as part of a global tactical asset allocation portfolio, which includes foreign stocks, commodities, and REITs

  • Daniel Sotiroff

    I can’t really make any recommendations, but I was somewhat surprised to find that there is indeed a T-Bill ETF offered by SSGA: https://www.spdrs.com/product/fund.seam?ticker=BIL

    That being said, the fees are a little steep for what you’re likely to get in return with our current interest rate environment. The 3 and 5 year returns are slightly negative–not really surprising

    Again, not a recommendation, but what I’ve done in the past is cut out the middle man and purchase Bills directly from the US Treasury. They’re auctioned weekly and can be purchased through Treasury Direct or an online broker. My broker doesn’t charge for auction purchases and there’s a minor fee for secondary market transactions, so it’s been a better option than an ETF from a transaction cost perspective
    https://www.treasurydirect.gov/indiv/products/prod_tbills_glance.htm