Last Updated on May 21, 2022 by Quantified Trading
Trend following has many followers, for some investors and traders trend following is like a religion. But does trend following work? And if so, why does it work?
Yes, trend following does work. If we look at history, even simple trend strategies work remarkably well. In this article, we look at the MLM Index and show that very simple (some would say naive) trend following strategies historically have produced better risk-adjusted returns than stocks.
This article looks at the performance of trend following, we explain the MLM Index, and we end the article by looking at trend following in stocks.
What is trend following?
Before you continue reading you might want to get a better understanding of what trend following really is. We have covered this in an earlier article, and we recommend you read about trend following strategies.
Does trend following work?
The risk premium earned by trend followers is hard to quantify. However, Mt. Lucas Management, a money management manager, has since the late 1980s made an index that we believe does a pretty good job in measuring the performance of trend following.
Their index, called Mt. Lucas Management Index (the MLM Index), is a very simplistic trend-following formula, some would even say naive. Nevertheless, it is widely used as a benchmark for many trend followers and the MLM index is a widely used benchmark for CTAs.
The index assumes no special knowledge of forecasting, math, algorithms – no nothing. The point Mt. Lucas is trying to make is the power of even simplistic trend following strategies. The idea is that commodity futures markets contain systematic movements that can be exploited even by simple methods.
What is the MLM Index?
MLM uses one indicator: The 200-day moving average! They use a moving average crossover for 25 commodity markets (equally weighted). When the price at the end of the month is above the 200-day moving average, a long position is held for the next month. If the price is below the moving average, a short position is held. Can it get any simpler?
There are 25 futures contracts in seven major categories in the MLM Index:
- Grains (corn, soybeans, soybean meal, soybean oil, and wheat),
- Livestock (live cattle),
- Energy (heating oil, crude oil, natural gas, and unleaded gasoline),
- Metals (gold, silver, and copper),
- Food and fiber (coffee, cotton, and sugar),
- Financials (5-year Treasury notes, 10-year Treasury notes, and Treasury bonds),
- Currencies (Australian dollars, British pounds, Canadian dollars, German Marks, Swiss francs, and Japanese yen)
Note that there are no stock market indices included (read more about trend following in stock indices below).
How has this very simple method performed?
The MLM index is calculated back to the 1960s. David Aronson, in his book called Evidence-Based Technical Analysis (a very good book), made some graphics of the performance up until 2005:
As you can see, the annual returns are more or less in line with stocks, despite having less “risk” (in terms of volatility, and we can argue how relevant that is of a risk parameter). The second graph shows how the MLM index has beaten stocks in terms of risk-adjusted returns (please read our article about trading strategy and system performance metrics to get a better understanding of risk).
Aronson’s results are confirmed by Mulvey, Kaul, and Simsek in Evaluating a Trend-Following
Commodity Index for Multi-Period Asset Allocation. Here are their findings:
(On a side note: The MLM index is sought replicated in an ETF with ticker code KMLM. By mid-February 2022 the ETF is up over 7% while the stock market is down double digits.)
Why trend following could mitigate risk:
Jensen et al (2002) examined the MLM index and concluded that trend following, and the MLM index in particular, performs better when the US monetary policy is contractionary.
In contrast, equities generally have a higher return when monetary policy is expansionary. Thus, from a portfolio viewpoint, trend following can offer both better returns and risk mitigation.
In a portfolio, you would want to have little correlation between your strategies. If we look at the correlation matrix from Mulvey et al. we see that the MLM index has zero correlation to the S&P 500 and also low correlation internally among the different futures contracts:
As evidence, just look at the performance of trend following since 2010 compared to stocks: stocks have been firing on all cylinders, while trend following has been left in the dust. But as of writing, this might be changing. For example, so far in 2022 trend following funds have performed fantastically over the last months. This is exactly what makes trend following so hard: very long periods with drawdowns or sideways performance with a low win rate.
Why does trend following work?
Trend following provides frequently large payoffs during periods of high volatility. The win rate is low, but the few rare outliers frequently produce big paydays.
Curtis Faith has written a very good book about trading – mainly about trend following. In chapter eleven he mentions something he calls the trader effect. The effect indicates most anomalies get arbed away because traders are seeking to exploit the same inefficiencies in the markets.
But trend following is a bit different. Why is that?
It’s because trend following often has long periods where you are slowly “bleeding” and getting whipsawed. As mentioned, the win ratio is low and it takes a lot of conviction to trade such strategies. Drawdowns are frequent and sometimes very big: 30-50% is not uncommon.
Most humans want to win frequently, even if it has a negative expected return. Most prefer many small winners instead of occasionally “lump-sum” big gains.
Does trend following work on stocks?
Yes, to a certain extent. David Aronson argues it does not work as well as compared to the commodity futures market, and it somewhat makes sense because there is no “risk transfer” in stocks.
The following chart shows the Sharpe Ratio for commodity futures vs. stocks in Aronson’s book:
But Wilcox and Crittenden (2005) argue differently. They looked at trend following in stocks (long only – no shorts) in a paper called Does Trend Following Work On Stocks? This is from the conclusion:
The evidence suggests that trend following can work well on stocks. Buying stocks at new all time highs and exiting them after they’ve fallen below a 10 ATR trailing stop would have yielded a significant return on average. The evidence also suggests that such trading would not have resulted in significant tax burdens relative to buy & hold investing. Test results show the potential for diversification exceeding that of the typical mutual fund. The trade results distribution shows significant right skew, indicating that large outlier trades would have been concentrated among winning trades rather than losing trades. At this stage we are comfortable answering the question “Does trend following work on stocks?” The evidence strongly suggests that it does.
In other blogposts we did some simple trend following tests as well by using the S&P 500 as a proxy:
You might also be interested in what we tested in gold and Bitcoin:
Evidence points out that the 200-day moving average is good at playing defense.
Does Trend Following Work? Conclusion
Yes, trend following does work. However, it works best in the commodity futures markets and is less effective in the stock market. Trend following works because they are less likely to get “arbed” away.
But trend following is not for the faint at heart as it takes courage and conviction to ride through the whipsaws and drawdowns. But the low correlation to stocks might make trend following a good supplement to a portfolio of stocks.