The Reality Behind Stocks’ ‘Worst’ Month

After starting on a heck of a down note, August also brought a sharp recovery from the midsummer pullback. The S&P 500 finished up 2.4% on the month and a whisker away from July 16’s prior high.^[i] Yet some fear the real pain lies ahead, as we now move into September—historically the “worst” month for stocks. Don’t let September’s purported poor past haunt your portfolio—it is just a month, not inherently better or worse for stocks.

Market history buffs may know September has the dubious honor of being the only month whose average total return is negative (-0.8%).^[ii] It has also been on a losing streak recently—down in the past four years (including September 2022’s -9.2%).^[iii] But take a step back. Over the past 98 years, September is still positive 51% of the time—slightly more often than not.^[iv]

Moreover, some huge outliers skew the monthly average.^[v] The worst monthly return of all time was September 1931 (-29.6%), during the Great Depression. Another deep negative September (-8.9%) was in 2008, the month the US government intentionally dismantled Wall Street as we knew it. It sowed chaos by forcing Lehman Brothers to fail after having midwifed a Bear Stearns rescue six months prior and nationalizing Fannie Mae and Freddie Mac days before … then nationalizing AIG days after Lehman. The worst month of 2022’s sentiment-driven bear market was September (-9.4%)—and a bull market began a couple weeks later. The myopic focus on some poor examples also overlooks some excellent Septembers, including 1939’s (16.9%) and 2010’s (8.9%) or 1954’s (8.7%). The upshot: The median return, which removes outliers both ways, is 0.0% (flat).

Look, the numbers are the numbers, and September’s haven’t been great relative to other months. But correlation isn’t causation. Lehman Brothers didn’t collapse because the calendar flipped to September. It is also happenstance that stocks took their standard late-bear market pounding in September 2022. Other bear markets ended—and had their worst runs—at other points on the calendar.

One could look at any calendar month and find bad outliers. March had two big double-digit declines in 1938 and 1939, respectively (-24.5% and -13.2%).^[vi] November has one of the highest average monthly returns (1.4%); that didn’t prevent 1948’s -10.3% drop.^[vii] December—one of the strongest months historically—fell -9.0% in 2018.^[viii] But these months don’t have the “worst month” reputation, so no one goes back to parse bad Marches, Novembers and Decembers. Humans unconsciously seek things that support pre-existing narratives—classic confirmation bias, like all those shark sightings after Jaws came out.

Exhibit 1 provides another way to look at this, displaying each month’s average return and its tally of big up and big down moves. As you will see, September isn’t a huge outlier in the big down department. Its 15 tallies of returns worse than -5.0% is the highest, but barely. August has 14, while May has 12.

September’s real weakness is that it doesn’t have as many big positive returns as the other months to offset the bad times. September topped 5% just 8 times since 1926, with May next closest (with 12 instances). Amusingly, the Santa Claus Rally is the inverse of this. There were 15 Decembers greater than 5.0%, which is middle-of-the-road—but just five below -5.0%. So the bullish December phenomenon is more like a lack of Santa Crashes (not as snappy of a ring, we guess).

Exhibit 1: Seasonality, Dissected

Source: Global Financial Data, Inc., as of 8/30/2024.

Some argue election uncertainty may rear its head this September, making one of those bad Septembers even likelier. Sounds logical. But does it hold true? Actually, in presidential election years, September has been positive 62.5% of the time (15 of the 24 instances since 1928)—better than the month’s overall frequency and less than 1 percentage point below all months’ frequency of positive returns.^[ix] Its average return is still negative, at -0.1%, but this is due primarily to the aforementioned 2008 debacle.^[x] The month’s negativity far predated John McCain’s shock decision to leave the campaign trail to join Senate deliberations on the crisis. Similarly, in 2000, the presidential campaign didn’t pop the Tech bubble and cause the -5.3% September return.^[xi] The median return, less skewed by outliers, is 0.3%.^[xii]

But here is the really critical thing: September is only a month, which likely isn’t very relevant to most investors’ actual time horizons. Shed the myopia and think bigger picture. Even with the history of some awful Septembers, stocks’ long-term average annualized return is 10.2%.^[xiii] If you try to invest around the possibility of a bad September, you could miss one of the good Septembers that contributes to this return—to say nothing of knowing when to get back in. After all, people see October as an extension of September’s “Financial Hurricane Season.”

So rather than get hung up on the calendar, we suggest refocusing on your long-term goals and the path likeliest to get you there. If that path includes stock exposure, then entering the month of September isn’t solid reason to veer from that.