Thursday, August 29, 2024

Decoding S&P 500 Seasonality | Martin Biber

The following chart shows three different calculations of the S&P 500's seasonality [...] one commonality is the V-shape into early October and out of it. That is the reason why everyone is talking about the weakness in September.


The following chart shows the seasonality of all the years since 1950. The seasonality only touches the low of the chart once, which means that over all those years, the market has always had one weakness at the same time. The low happens in late September, followed by a substantial rally until January. The seasonality shows some weakness into March, followed by a reaction that ends at the beginning of May. After that, the seasonality, calculated without a trend component in this chart, goes down until late September. This is why we often say, “Sell in May and go away.”

 
The red line on the following chart also shows seasonality, but only the last ten years were used to calculate the results. The behavior around the seasonal low in October looks similar to the seasonality in all years. I have, however, a second low in the middle of March and a clear top in mid-July. This top is only a few days after the one we experienced this year. Interestingly, the saying “Sell in May and go away” wasn’t good advice over the last ten years, as we would have missed the rallies between March and mid-July.

 
The following chart compares the blue seasonality with the seasonality of the years ending with 4 (2014, 2004, etc.). The low occurs in early October, compared to the two previous seasonalities, which have the seasonal low at the end of September. The rally only lasts until early November. But there is a second rally immediately afterward, which lasts until late January.

 
Summary:
While all seasonalities point to a weak September and a crucial low late in September, the seasonality of the years ending in 4 lets us assume that the seasonal low could occur one or two weeks later than usual. Anyway, no other date in a year is statistically as significant for the S&P 500 as everyone knows and is waiting for the late September or early October low.