August the best month for average stock market performance? Or is it the worst?
The answer depends on the period of stock-market history you examine. Over the 90 years from the Dow Jones Industrial Average’s DJIA, +0.50% inception in 1896 until 1986, August on average was far ahead of the other months — more than four times larger, as you can see from the table below. August outperformed the other months’ average by 1.4 percentage points. This difference is significant at the 95% confidence level that statisticians often use when determining if a pattern is genuine.
In the years since then, in contrast, August has been the worst month for the stock market, on average, lagging the other months’ average by 1.7 percentage points. Since 1986, in fact, August has been a worse month for the stock market than even September, whose reputation for stock market losses is widely known.
|August’s average DJIA return||Average return of all other months||August’s rank among all 12 months|
|1896 to 1986||+1.8%||+0.4%||1st|
If the 36 years since 1986 were all that statisticians had to go on, they would conclude that August’s underperformance was significant at the 95% confidence level — just the opposite of the conclusion that emerges from the 90 years prior. But when analyzing the Dow’s entire history since 1896, August’s performance is no better or worse than average.
This August, in order to use history as a basis for investing, you’d first need to come up with a plausible explanation of what changed in the 1980s that caused August to swing from best to worst.
Though I’m not aware of any such explanation, it’s always possible that one exists. To search for it, I analyzed monthly values back 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 stock market responds to changes in economic uncertainty, so we’d be onto a possible explanation of August’s seasonal tendencies if the EPU underwent some fundamental change in 1986.
But no such change shows up in the data. August’s average EPU level is no different than for any of the other months of the calendar, either before or after 1986.
Another possible explanation might trace to investor sentiment. To investigate that possibility, I analyzed stock market timers’ average recommended equity exposure levels, as measured by the Hulbert Stock Newsletter Sentiment Index (HSNSI). I was looking to see if, after 1986, the HSNSI was significantly different at the beginning of August than in other months, on average. The answer is “no.”
A plausible explanation might still exist for August’s change of fortune beginning in the mid-1980s, notwithstanding my inability to find one. But absent such an explanation, the most likely explanation is that it’s a random fluke.
It would hardly be a surprise if randomness is the culprit. Most of the patterns that capture Wall Street’s attention are in fact nothing more than statistical noise. The reason we nevertheless insist that significant patterns exist is because — as numerous psychological studies have shown — we’re hardwired to find patterns even in randomness.
That’s why your default reaction to all alleged patterns, not just those involving August, should be skepticism. The odds are overwhelming that they aren’t genuine. Only if those patterns can survive the scrutiny of a skeptical statistician should you even begin to be interested.
Mark Hulbert is a regular contributor to MarketWatch. His Hulbert Ratings tracks investment newsletters that pay a flat fee to be audited. He can be reached at email@example.com