Practical Trend-Following (Part 1)

There are lots of good papers on trend-following. My goal here isn’t to prove to you that trend-following works, but rather to show you how it can be used in practice.

Samuel Lee 12 March, 2015 | 9:00
Facebook Twitter LinkedIn

Soon after I joined the ETF team, my colleague Tim Strauts showed me some back-tests for a simple trend-following strategy. At the end of each month, check whether an asset’s price is above its moving average, calculated with the last 10 month-end prices. If it is, own the asset; if it isn’t, go to cash. I thought it was dumb, because I knew technical strategies and market-timing can’t work.

 

In order to prove that the results were a fluke, I tested the trend-following strategy on pretty much every return series I could get my hands on—gold, commodities, emerging-markets stocks, developed-markets stocks, U.S. stock sectors, currencies, high-yield bonds, investment-grade bonds, Treasuries, and so on. To my surprise, almost every return series showed improvement in risk-adjusted return, over long periods and across many different specifications (using six to 18 month returns, six to 18 month moving averages, and even some nonsensical variations). There’s nothing like trying your hardest to kill an idea and failing. I became a believer. The experience was one of many that taught me the value of independent thinking.

 

There are lots of good papers on trend-following. My goal here isn’t to prove to you that trend-following works, but rather to show you how it can be used in practice. Most investors are market-timers, and many will always be market-timers. If you’re going to do it, you might as well do it in a way that could work.

 

Exhibit 1 shows the growth of $1 in the U.S. stock market and in a simple trend-following strategy, adjusted for inflation. At the end of each month, check whether the current total return index is above the 12-month simple average; if yes, go long, otherwise, own Treasury bills. The y-axis uses a logarithmic scale, so each tick up represents the same percent change. I used U.S. stock market and Treasury bill returns from the French Data Library. The data cover June-end 1926 to December-end 2014, but the strategy return begins in May-end 1927 on account of the data required to compute the moving average. The strategy would have avoided the worst of the Great Depression, the recession of 1937–38, the crash of 1973–74, the bursting of the dot-com bubble, and, last but not least, the financial crisis. Absolute returns would have been a bit higher, too.

 

Setting aside taxes and transaction costs, there’s a small problem: You would underperform the market for decades. Trend-following is like disaster insurance. Most of the time, it’s a drag on your returns. Once a generation it pays off big.

 

Exhibit 2 plots the cumulative wealth of the trend-following strategy divided by the cumulative wealth of the market. When the line rises, the strategy is outperforming, and when it falls, it’s underperforming. The four big spikes you see correspond to the four big panics the U.S. market experienced. The dot-com bubble is unusual in that it ends with a downward grind, which shows up in this chart as an upward-sloping line. The line steadily slopes down from 1940 to 1972 and from 1975 to 2000. A manager using this strategy would’ve been fired long before it paid off. During the 1980s, market-timing funds sprouted up; none of them survived the great 80s and 90s bull market.

 

Trend-following has had a tough time since the financial crisis. Investors have been abandoning it in droves. This is short-sighted. History shows long periods of big underperformance are common in the years (or even decades) after a big market crash. The tracking error is more palatable when you think of trend-following as a method of risk control, not return enhancement. Framed this way, its logical competitor is a balanced stock-bond portfolio, not the stock market.

 

Exhibit 3 plots the wealth ratio of a trend-following strategy with a twist: rather than going into cash, it goes into intermediate Treasuries, represented by the Ibbotson Associates Intermediate-Term Government Bond Index. Its benchmark is a static 75–25 stock-bond portfolio, rebalanced monthly. Remarkably, its only episode of sustained underperformance is the great bull run in the 80s and 90s.

 

 

In part 2 of this article, we will try to see why trend-following strategy might have worked.

 

 

Samuel Lee is a strategist covering passive strategies on Morningstar’s manager research team and editor of Morningstar ETFInvestor, a monthly investment newsletter

Facebook Twitter LinkedIn

About Author

Samuel Lee  Samuel Lee is an ETF strategist with Morningstar and editor of Morningstar ETFInvestor

© Copyright 2024 Morningstar Asia Ltd. All rights reserved.

Terms of Use        Privacy Policy       Disclosures