Halloween marks the beginning of one of the most seasonally favorable periods of the stock market calendar.
That should come as welcome news indeed to beleaguered investors, since the S&P 500 SPX, -1.51% has shed nearly 7% so far for the month of October.
I’m referring, of course, to the U.S. stock market’s famous six-months-on, six-months-off seasonal pattern that goes variously by the names of “The Halloween Indicator” and “Sell in May and Go Away.” On average over the last century, almost all of the stock market’s gains have been produced in the Halloween-through-May-Day period — the so-called “winter” months. In contrast, the market has been little better than flat, on average, during the “summer” months.
To be sure, recent research has found that the market’s superior gain during the winter months comes from just one year of the four-year presidential term. That superior year is the third year, which is now just beginning.
So investors who follow this updated version of the Halloween Indicator are finally, after several years of patiently waiting, on the verge of the six-month period for which optimism appears to be based on more than mere hope.
How confident should we be that the pattern will hold up from this coming Halloween to next year’s May Day? Quite confident, it would seem. Since the Dow Jones Industrial Average DJIA, -0.92% was created in 1896, the “winter” months of the third years of the presidential term have experienced an average gain of 12.0%. That compares to an average gain of 2.9% in the winter months of years one-, two-, and four, and an average gain of just 1.9% during the summer months. These differences are significant at the 95% confidence level that statisticians often use to determine if a pattern is genuine.
Furthermore, the stock market’s gain during the winter months of presidential-term third years has been impressively consistent. The worst six-month winter return in all third years since 1940 was a Dow loss of just 1.7%, and even this loss turns into a gain if you include dividends in the calculation.
To be sure, as I wrote recently, you can also find reasons for doubt. You always can, of course.
For example, it’s not clear that there is a strong theoretical basis for expecting the third year of President Trump’s term to live up to the historical pattern. The only plausible theory of the Halloween Indicator that has emerged from academic research traces to pre-midterm-election uncertainty (which causes the summer months to be below average). Once that uncertainty gets resolved on Election Day, the market can bounce back.
The reason to question this theory’s application this time around is that, at least according to the Economic Policy Uncertainty index, the past six months have experienced below-average levels of uncertainty. And, to that extent, according to the theory, the market will have less upside potential once the midterms are behind us.
Devotees of the Halloween Indicator can nevertheless point to the remarkable consistency of the six-month winter period of presidential term third years. As you can see from the accompanying chart, the Dow’s average return over this period has been virtually the same regardless of whether the market was strong or weak over the preceding six-month “summer” period.
In any case, it’s interesting to note that the stock market’s decline so far this October has brought its gain for the “summer” period that began this past May 1 back to almost precisely its historical average. With just four more trading days until the full “Sell in May and Go Away” is complete, the Dow Jones Industrial Average is ahead just 1.7% from where it stood on May Day — versus an historical average of 1.9%.