A while back Kathy Lien wrote an article on forex seasonality. Like much of the analysis that tends to float around the markets, however, it features a major flaw.
Kathy’s analysis of forex seasonality comprises examinations of 10 years worth of monthly returns. It looks very impressive to say that 80% of the time a given pair rises in a certain month. As any statistician will tell you, though, 10 observations simply isn’t enough data to draw a reasonable conclusion.
On top of that, just because the market rises or falls more often during a certain period, that doesn’t mean the pattern is a reliable one. We also have to look at how much the market actually moves.
The big factor in all this is statistical significance. To reference Investopedia once more, “Statistically significant is the likelihood that a relationship between two or more variables is caused by something other than random chance.”
One of the major issues we have with the stuff we hear about forex calendar patterns is that we don’t know if they are anything other than randomness at work. Sometimes they just aren’t using enough data to derive the pattern – as in Kathy’s examples. Other times there’s simply too much volatility in the data to draw a proper conclusion.
Here’s an example.
There are two months during the year where the EUR, broadly speaking, has been up nearly 70% of the time since it’s launch in 1999. That seems like a strong pattern, doesn’t it? When we look at the average performance during those months, though, it’s not so great. Statistically, the average return is not far enough away from 0% to be significant. This could be because the returns are just too small. It might also be because the volatility of those returns is too high. Maybe there’s just not enough data to say with a high level of confidence.
A market analyst actually fell into this very trap here. He reports on a high frequency pattern, but it’s one for which the market does not statistically show anything but a random return pattern. That’s not saying the information isn’t useful. It can be, for sure. Just that you must be aware of the limitations. This is why I include deeper statistical analysis in my own work.
The point is don’t just look at some impressive charts and conclude there is a strong pattern. It could lead to major disappointment.