I recently wrote an article explaining how it is possible for momentum traders to profitably implement a “best of” Forex momentum trading strategy, which included a back test conducted over a very recent 6 year period. There are a few loose ends in that article that are worth some more detail, so in this second part I want to make a stronger and more detailed case as to why standalone / time series momentum tends to be a better kind of momentum strategy overall, and clear up some concerns that might have arisen from my use of a 3 month look-back period in determining the best and worst performing currency pairs.
Why “Best of” Momentum Works
Academic studies have found that the most profitable trading strategy that can possibly be constructed based upon historical price data alone, is a time series momentum-based trading strategy. This can be implemented by momentum traders simply by selecting a diversified universe of tradable instruments and buying the ones going up and selling the ones going down. This is actually a method that tends to produce greater profits overall than adding a “best of” filter, but the draw-downs are larger and so it usually makes more sense to add a filter such as “best of” although there is no reason why fundamental analysis or other filters could not be used profitably instead.
There has been much academic speculation as to why momentum “works” and there is no consensus on this question. My own opinion is simply that for something to get from 100 to 200 in its price, it has to go up, and human nature is such that crowds tend to pile into moves at tipping points, making the momentum even stronger.
Now let’s turn to any concerns that might have been raised over my choice of 3 months as a look-back period for determining which pairs to trade.
Look-Back Period for Selecting Currency Pairs
I used a 3 month look-back period in my previous article simply because it produced the best overall result of all possible look-back periods. If you are a momentum trader concerned that the concept does not look very robust until some other look-back periods have been measured, you are absolutely right! In order to address this I am reproducing below the results for every look-back period at 2 weekly intervals from 2 weeks to 24 weeks (equating to 6 months), followed by another graph showing the average of all the samples.
Of the 12 samples, only 1 of them completes the test with a positive return, compared to 11 with a negative return. Therefore the 3 month look-back generation of a positive result might be a statistical fluke. You might say that since May 2012 the strategy overall has been slightly profitable, but not by much. Let’s look at the average performance of all of the 12 samples now:
The average performance is quite strongly negative, albeit marginally positive since May 2012.
Time Series Momentum
If these results make you feel nervous about using a “best of” Forex momentum strategy, you could instead consider using a simple time series momentum strategy. Here, momentum traders just select some Forex pairs, and for the purposes of our back test go long each week the price is higher than its own price of X time ago (X representing the look-back period), or short if the price is lower than its own price of X time ago.
The obvious question we run into first when trying to follow this kind of strategy is which Forex pairs to use? Do we want to be trading all the Forex pairs all time, without discriminating between them?
It makes sense to start by looking at the 4 major pairs: EUR/USD, GBP/USD, USD/CHF and USDJPY. Below are the results of a back test over a very long period of time – from January 2002 until early 2015, which represents more than 13 years. This test has some different parameters: the trades are taken only at the beginning of calendar months, trades are held for 1 month, and the look-back periods are previous calendar months. The look-back periods used were 1, 3, 6, and 12 months:
This is surprising, as all the look-back periods used were profitable. An average of all 4 strategies would have produced a return in excess of 100%, and currently it is only the 1 year period that is within a serious draw-down.
Trade USD and EURO Currency Pairs
The two biggest global currencies are the USD and the EUR. They are most prone to trending steadily and this is one of the reasons why time series momentum with the 4 major pairs has worked well: they are all USD currency pairs. Why should momentum traders be especially interested in these currencies?
Simply because they are the two largest currencies by volume and importance. It takes time to turn around a big ship.
Let’s conclude with some data showing how the USD and the EUR love to trend. Over a period of 6 years – from April 2009 to April 2015 – if you looked at the 28 most important currency pairs and went long or short of each every week depending upon its look-back periods of 13 or 26 weeks, the only currencies producing positive results were the EUR and the USD. Both currencies would have produced a return of 110% each based upon the 26 week look-back period (corresponding to 6 months). Using the look-back period of 13 weeks (corresponding to 3 months) produced a positive result of 161% for the USD and 82% for the EUR. Using this kind of momentum trading strategy could be a good way to turn $10,000 into $1 million.