August was once one of the best month for the inventory market. Then it grew to become the worst.

In the years since then, in distinction, August has been the worst month for the inventory market, on common, lagging the opposite months’ common by 1.7 proportion factors. Since 1986, in reality, August has been a worse month for the inventory market than even September, whose fame for inventory market losses is broadly recognized.

If the 36 years since 1986 had been all that statisticians needed to go on, they’d conclude that August’s underperformance was important on the 95% confidence stage — simply the other of the conclusion that emerges from the 90 years prior. But when analyzing the Dow’s total historical past since 1896, August’s efficiency isn’t any higher or worse than common.

This August, with a view to use historical past as a foundation for investing, you’d first must provide you with a believable clarification of what modified within the Eighties that precipitated August to swing from greatest to worst.

Though I’m not conscious of any such clarification, it’s all the time attainable that one exists. To seek for it, I analyzed month-to-month values again to 1900 for the Economic Policy Uncertainty (EPU) index that was created by Scott Baker of Northwestern University, Nicholas Bloom of Stanford University, and Steven Davis of the University of Chicago. We know from Finance 101 that the inventory market responds to modifications in financial uncertainty, so we’d be onto a attainable clarification of August’s seasonal tendencies if the EPU underwent some elementary change in 1986.

But no such change reveals up within the knowledge. August’s common EPU stage isn’t any totally different than for any of the opposite months of the calendar, both earlier than or after 1986.

Another attainable clarification would possibly hint to investor sentiment. To examine that risk, I analyzed inventory market timers’ common advisable fairness publicity ranges, as measured by the Hulbert Stock Newsletter Sentiment Index (HSNSI). I used to be trying to see if, after 1986, the HSNSI was considerably totally different firstly of August than in different months, on common. The reply is “no.”

A believable clarification would possibly nonetheless exist for August’s change of fortune starting within the mid-Eighties, however my incapacity to search out one. But absent such an evidence, the probably clarification is that it’s a random fluke.

It would hardly be a shock if randomness is the wrongdoer. Most of the patterns that seize Wall Street’s consideration are in reality nothing greater than statistical noise. The cause we however insist that important patterns exist is as a result of — as quite a few psychological research have proven — we’re hardwired to search out patterns even in randomness.

That’s why your default response to all alleged patterns, not simply these involving August, ought to be skepticism. The odds are overwhelming that they aren’t real. Only if these patterns can survive the scrutiny of a skeptical statistician do you have to even start to have an interest.

Mark Hulbert is an everyday contributor to MarketWatch. His Hulbert Ratings tracks funding newsletters that pay a flat price to be audited. He could be reached at mark@hulbertratings.com

More: Puzzled by the stock-market surge? Overshoots are the brand new regular, Bank of America strategist says

Plus: Here’s how lengthy the inventory market rally might final