3 Numbers To Watch: EURUSD, EURGDP, USDJPY levels
Monday is a slow news day for economic releases, which provides an opportunity for a quick review of the momentum factor in three currency pairs: EURUSD, EURGDP, USDJPY. A few weeks ago I profiled this trio and their respective 20-day return histories through time. It is always helpful to periodically revisit a given analytical methodology and how it compares after the fact on the assumption that progress in estimating future values for currencies (or any other data series) is a learning process rather than a reliable point forecast.
EURUSD: On December 16, I wrote that EURUSD "still has room to run higher," based on the historical perspective for a 20-day trading horizon. The reasoning was that most (97%) of EURUSD's rolling 20-day return histories for the previous 10 years are contained within a +/-6 percent performance band. True, but I should have also mentioned that a +/-4 percent performance band captured almost as much (86%) of the fluctuations over the past decade. As it turns out, the +/-4 percent band was a short-term ceiling. EURUSD's 20-day return topped out around a 3 percent gain recently and has been sliding in the new year.
Deciding where and when a trading position has run its course is never easy, but we can benefit from a number of empirical facts in forex, starting with the persistence of momentum. Relying on one metric alone for assessing momentum is problematic, but studying historical return patterns is usually a good start. Assuming a 20-day trading horizon, for example, implies that we should be increasingly cautious whenever returns move into the outer range of historical experience. But it is also prudent to look for confirmation, or rejection, from several methodologies for a richer read on estimating expected return. Analysing the trend via several technical analysis filters is one possibility. That includes reviewing the +/- percentage deviation for the latest price versus a 20-day moving average of those prices. Looking at EURUSD through this analytical lens implies that the outlook is relatively neutral compared with the last several weeks.
EURGBP: My previous comments on EURGBP also failed to mention that a tighter performance band of +/-2 percent cannot be ignored, given its history of representing the upper and lower range of 20-day returns for 80 percent of these rolling periods over the past decade. This turned out to be a hard mark for EURGBP recently, which has hit some headwinds lately. The 20-day return has fallen to virtually nil as of Friday. The percentage deviation in the EURGBP's current price relative to its 20-day average is also hovering around zero. The message here too is that the momentum signals look considerably closer to neutral compared with late-December.
USDJPY: It is another story for USDJPY, which has been in a powerful rally lately. Last month I noted that USDJPY was closing in on its historical upper range for 20-day returns, which suggested caution. The observation looked relevant for a few days after my December 16 post, but then events intervened with the late-December installation of a new prime minister in Japan. As we all know, Shinzo Abe is leading the charge for a substantially more aggressive round of monetary stimulus, which has weakened the yen lately.
The bottom line: USDJPY has continued to rally, with the 20-day return pushing higher into rarefied performance territory for this time horizon. As of January 4, USDJPY is up 7.5 percent over the past 20 days. Since the end of 2002, 7 percent-plus returns have occurred only 0.6 percent of the time for 20-day rolling periods. The +/- percentage deviation in the current price versus its 20-day average also confirms that recent performance is in the outer range for positive returns for the past decade.
USDJPY could run higher still, of course. So, what is the value of quantitatively profiling momentum and other aspects of trading histories? Perspective. If USDJPY's track record reveals a slim 0.6 percent frequency of 20-day returns north of 7 percent, the implication is that betting on returns over that level will be a losing proposition 99.4 percent of the time. That is not necessarily going to help you on the next trade, of course. On the other hand, that is a considerably more informative piece of information if you are designing a trading system that has a decent chance of surviving the test of time.