Sell in May? A Brief Sanity Check
May 1, 2013
by Doug Short
On this first day of May, in homage to the Sell in May soothsayers (and in the spirit of market trivia), let's survey the behavior of the months in the S&P Composite since 1928. The Composite is a spliced index of the earlier S&P 90 with the S&P 500. Because of its greater breadth, I prefer it to the Dow for historic research of this sort.
Market lore is full of monthly associations: The January Effect, Sell in May and Go Away, Summer Rallies, the September Slump, Manic-Depressive October, December Rallies, etc.
The first chart shows the average monthly gains/losses, excluding dividends, since 1928 for all twelve months. May is one of the three months with a negative average. Incidentally, the monthly average of all months lumped together is 0.59%. So May has underperformed the mean by 0.73%.
Dividing the Timeline
Has May consistently been a poor performer? The next three charts divvy up our 85-year period into three parts: 1928-1949, 1950-1981, and 1982-present. The rationale is that the first chart includes the Crash of 1929, Great Depression, WWII, and ends around the time of the secular market bottom in 1949. The second chart covers the cycle from the beginnings of the post-war rally through the Decade of Stagflation and market bottom in 1982. The third chart begins with the great Boomer market that followed and runs to the present.
Of course we could slice and dice the decades any number of ways, but an overview and three subsets (a four-chart total) about maxes out my energy and interest in this topic.
May has been a performance laggard in two of the three timeframes and the worst performer in one of the three (1950-1981).
Without further ado, I'll let the next three charts speak for themselves.
Lest the charts above give the false impression that May is a consistently poor performer, let's close with a distribution of performance over the past 85 years.
Across the entire 85-year timeframe, May has an average of -0.14%. But if we exclude the three negative outliers, the average jumps to 0.59%, which is spot on the overall monthly mean. Pretty amazing!
Let's hope May 2013 behaves more like it did in 1933 and not like one of those naughty negative outliers (or any of the red markers, for that matter).