The previous posts in this series (part 1 and part 2) used two different methods to analyze the monthly seasonality in the S&P 500 since 1950. Here we'll apply the second method — monthly averages of daily closes — to the seven two-decade periods since 1871.

Let's study a table showing the results of our calculations:

As we can readily see, there has been little in the way of long-term consistency in monthly seasonality based on monthly averages of daily closes. The notion of a summer rally, for example, had more substance before 1950 than after. The reverse is the case for end-of-year rallies. November and December have had a better record since mid century.

The two consistent months have been January and October. January alone has turned in above-average performance in all our two-decade snapshots. October has been the only perpetual underperformer. However, these 20-year timeframes obscure a key difference between the two. January has been more consistent in its positive monthly averages of daily closes. Since 1871 the first month of the year has registered in the green 69% of the time compared to October's 51%. However, when we compare the average of the negative months, the nasty side of October becomes apparent. The 43 negative Januaries have averaged -1.65%. The 68 negative Octobers have the stunningly bad average of -3.73%. The worst October on record was in 2008 with its -20.39% decline in its monthly average of daily closes. Since 1871 there have been two months with more dismal records: -23.97% in April 1932 and -26.47% in November 1929.

Next up in the 2008 calendar is September. How has it fared since 1871? It falls about midway between January and October with its 59% positive performance in our featured metric. However, like October, the negative years have been disproportionately painful with a -3.48% monthly average of daily closes.

Let's hope September and October show us their sunnier personalities in 2009.