Should We Trust Seasonality?

If you pay attention to financial media, you’re likely to hear quite a bit about September being the worst month for investors.  While this is indeed borne out by the data – September is indeed the month with the lowest mean return for key indexes – seasonality is a fickle guide.  Long-term averages are one thing; performance in any given month is another.

Looking back at the data from the start of this century, only four calendar months show average negative returns for the S&P 500 (SPX): January (-0.08%), February (-0.50%), June (-0.05%), and September (-1.31%).  In the case of the Nasdaq 100 (NDX), there are only three: February (-0.46%), September (-1.84%), and December (-0.32%).  When we shorten the timeframe to the past 10 years, the same months show negatives for NDX but the values differ: February (-0.49%), September (-1.68%), and December (-0.11%).  For SPX, the list of down months shrinks to two: February (-0.54%) and September (-1.34%).

Bottom line, it is much better, on average, to remain in the market.  Most months show positive average returns, and the majority of individual months over each period show positive returns as well. 

On the other hand, it is quite glaring when one specific month is glaringly worse than the others.  Note that September consistently shows average losses of more than 1% when the few other months with negative average returns barely register in the minus column.  On that basis, it is understandable why it looms large in investors’ mindsets.  September is an obvious outlier.

Even so, averages hardly tell the complete story.  SPX closed lower in just 5 of the 10 past Septembers, while NDX had 6 lower closes in the prior 10.  That is about as close to random as we can get.  Furthermore, while both indexes had ugly four-year stretches from 2020-2023 (-3.92%, -4.76%, -9.34%, and -4.87% for SPX; and -5.72%, -5.73%, -10.60%, -5.07% for NDX), the last two Septembers were solidly positive for both (+2.02% and +3.53% for SPX; and +2.48% and +5.40% for NDX). 

Although this September is not getting off to a positive start, this morning’s dip is relatively modest.  Besides, a few hours of trading hardly sets the tone for an entire month.  Yields continue to rise, with a move toward 5% in 10-year Treasuries seeming increasingly likely – particularly as the situation in the Persian Gulf gets no better (we received news of new tanker attacks overnight and new US strikes on IRGC targets as I typed this piece).  Volumes remain relatively suppressed ahead of Labor Day, and many traders may be more likely to focus on risk aversion ahead of the August jobs report on Friday morning.

If you want to play the odds regarding seasonality, decide which odds you’re playing.  Is it the historical evidence that September is consistently the worst month of the year for index investors?  Or is it the relatively random nature of up-versus-down Septembers in any given year?  Or is it the two-year September winning streak that comprised the most recent performance?  Maybe we should focus more on fundamentals and key externalities rather than strictly on seasonality, no?