Todd's Take

Donchian Rules

Todd Hultman , DTN Grains Analyst
This chart shows hypothetical results of three trend-following rules, based on Richard Donchian's simple method. For corn, going long on a new 30-day high and short on a new 30-day low has been surprisingly profitable and far better than buy and hold results for December corn (Source: DTN Analyst Todd Hultman).

For those who spend a lot of effort trying to understand the grain markets and sometimes feel overwhelmed by the complexity of it all, you may want to sit down for what I am about to show. It is one of those strange things in life that is hard to explain, but can't be denied.

In 2016 alone, we have seen enough events for Billy Joel to add another verse to his 1989 montage, "We Didn't Start The Fire." I admit the lyrics aren't as catchy, but this year's market-threatening topics already include record harvests, rising dollar, zika virus, Janet Yellen, Dilma Rousseff, wildfires, La Nina, hot June, wet July, more wheat, Chinese floods, building walls, and Brexit.

As DTN Senior Analyst Darin Newsom wrote in Friday's Newsom on the Market column "The Haystack," much of this year's grain moves can be explained by shifts in noncommercial positions which don't always seem to make sense fundamentally. One of the biggest surprises this year was the rally in soybeans after noncommercials turned net long in mid-March.

At the time, it was widely expected Brazil was harvesting a record soybean crop and would soon be taking export business away from the U.S. Spot soybean prices, however, were not bothered by the lack of fundamental explanation and proceeded to climb from roughly $9.00 in mid-March to $12.00 by early June. It turned out to be a significant move that many grain-watchers missed.

So what is the strange thing I promised? It comes from Richard Donchian, born in 1905 in Connecticut to Armenian immigrants. As Wikipedia tells it, Donchian began his Wall Street career in the 1930s, worked as a securities analyst, wrote newsletters, and later started one of the first publicly-held commodity funds.*

Donchian tinkered with various trend-following methods and it is not clear when he came up with the four-week trading rule for which he is known, but it's easy enough for anyone to understand. The rule states that one should go long when the price of any commodity exceeds its high of the past 20 trading days (four weeks) and reverse short when the price falls below its low of the past 20 trading days.

The advantage of following Donchian's rule is that one is always positioned with the trend whether we understand why prices are going a certain direction or not. The disadvantages are numerous and start with the fact that one is always in the market, right or wrong. When prices chop up and down as they sometimes do, this method loses money.

According to Investopedia, "Test results for this (Donchian four-week rule) system were published as early as 1970, and it was found to be the most profitable system then known."** I decided it was time to update those test results and enlisted the help of DTN's ProphetX software to put together over 20 years of back-adjusted price data for December corn. To keep things simple, I only used daily closing prices.

Applying Donchian's four-week rule to December corn price data from 1995 up to July 14 of this year showed a hypothetical trading profit of $5.53 a bushel or $27,650 per contract, not including slippage or commission. While that is somewhat successful, the profit per trade only came to $151.92 or slightly over 3 cents a bushel, which does not leave much room for error.

Because I have always suspected that 20 trading days was too short of a time frame, I decided to re-test Donchian's method with other time periods. Not surprisingly, a 10-day rule generated too many trades and was slightly unprofitable. A 50-day rule did better, producing a hypothetical profit of $45,450 per contract on less than half of the trades required by the 20-day rule.

The star performer of the past 21 years however, was the 30-day rule. Going long Dec corn on a new closing high of the past 30 trading days and reversing short when prices closed at a new 30-day low produced $66,525 of hypothetical profit on 102 trades. This worked out to roughly five trades a year of $652.21 each, a winning track record by any standard.

In fact, the 20-day, 30-day, 40-day, and 50-day versions of Donchian's method all easily outperformed simply buying and holding one contract of December corn throughout the 21 1/2-year period tested.***

With that kind of trading math at work, why doesn't everyone use a 30-day version of Donchian's rule for their trading and/or risk management decisions? I suspect it's because it runs against the grain of human nature. As profitable as the history of this method has been, most understand past success does not guarantee a profitable future and the uncertainty of not knowing can be stressful.

A closer look showed that following Donchian's rules did not always produce rosy results. After the 30-day rule amassed big gains by November, 2010, the hypothetical trading account lost 31% of its equity in the following year and a half. Many would not be stoic enough to endure that much loss for that long of a time and still stay true to the system.

Most likely the toughest difficulty to overcome is the reluctance to give up old habits. Most of us have learned to study up on the news, weather, USDA reports, etc. We also watch out for numerous what-if scenarios that can make our head spin and keep us up at night.

Trend-followers, on the other hand, have to learn to disregard everything except for price and that is not necessarily easy. The traditional approach makes trades based on known factors, but trend-following forces one to also consider factors at work which aren't known. Factors like the ones that propelled this spring's soybean rally.

Every trade is not a winner and Donchian's method is not for everyone, but it deserves a look and should give us pause to consider the pros and cons of how we look at markets. As a market analyst, I have to tip my hat in amazement to a man who devised a trading method decades ago which still outperforms most efforts of traditional analysis, all without the benefit of any supply or demand information. In my book, Donchian rules.

* https://en.wikipedia.org/…

** "Four-Week Rule Boosts Winning Trades," by Michael Carr, August 5, 2014 at:

http://www.investopedia.com/…

*** Of course, one contract of December corn can't be held for 21 1/2 years, so the data assumes that the position is rolled forward on the final day of November each year.

Todd Hultman can be reached at todd.hultman@dtn.com

Follow Todd on Twitter @ToddHultman1

(CZ)

Todd Hultman