A seasonal sector-switching strategy has beaten the S&P YTD

The Study

I ran an analysis on monthly stock-market returns over the last 20 years or so, to determine which sector delivered the best median dividend-adjusted returns in each month. December 2020 was the last month included in the study.

The sector funds included in the analysis were all equal-weighted, although they don't all use identical methodologies or have identical expense ratios, so keep that in mind. Also, for some funds, there was more data than for others. Some have been around since 2005, some 2007, some 2012. Here are the funds I included: EWRE, RCD, RGI, RTM, RYE, RYF, RHS, RYH, RYU, XAR, XBI, XHB, XME, XTL, XTN, XSW, XSD.

The limitations of the data mean that the results are probably pretty noisy. I've got no real way to determine the statistical significance of these results, because there's just not enough data. A lot of this will be "noise," but probably there's some "signal" here too.

Basically I compared these sector funds' median return in each month and determined which fund gave the best median return. (The median should be a more robust statistical summary than the mean, because the mean will be affected by outliers like meltups and crashes.)

The Results

Here are the best-performing sectors by month:

January: Biotech
February: Aerospace
March: Real Estate
April: Energy
May: Semiconductors
June: Real Estate
July: Semiconductors
August: Semiconductors
September: Materials
October: Transportation
November: Transportation
December: Metals

The Backtest

Since I didn't use data from 2021 to generate these results, we can backtest a sector-switching strategy on 2021. What if, in each month, we switched to the sector with the best median return for that month?

The answer is that the equal-weighted index returned about 18% YTD, whereas the sector-switching strategy returned about 21%. So there's a slight edge here, but not a large one. If you're trading in a tax-deferred IRA, you'd have come out ahead by using the sector-switching strategy. But if your account isn't tax-deferred, then this strategy will have cost you more in capital gains taxes than you're making in excess returns vs. simply holding the index.

The Code

I've put the R code for this analysis on Github, should anyone wish to check my work: github.com/chriscarrollsmith/wallstreetpettingzoo/blob/main/sector-seasonality.R
activebacktestmonthspassiveSeasonalitysectorstaxes

Disclaimer