Seasonal Dollar Patterns and Why the Statistics Stay Thin

[Shadow Trader]

Charts of average monthly dollar performance circulate every year, often showing that some months have tended to favour the currency and others to weigh on it. The patterns look tidy, but their statistical foundation is much weaker than the visuals imply.

The first issue is sample size. A floating dollar has only a few decades of history, so each calendar month has a few dozen observations at most. Each observation is also dominated by whatever macro regime prevailed that year, such as a tightening cycle, a crisis or a policy shift. One strong episode can drag an entire month's average.

The second issue is multiple comparisons. Twelve months, several pairs and many time windows produce a large number of possible patterns, and some will look striking by chance alone. Once found, a pattern also tends to erode as participants notice it, if it was ever real.

A reasonable use of seasonal charts is as a hypothesis generator rather than a signal. A claim becomes more credible when it has an economic mechanism, such as a known flow or a reporting date, and when it survives across pairs and periods. Otherwise, it is more likely a coincidence than an edge.