EconomPic has outlined the seasonal performance of the equity market (with a "secret sauce" twist) a number of times (most recently here).
Here are the specifics of seasonality: Imagine we start with two $10,000 accounts, and use them to make investments in an S&P 500 Index fund. One account invests in one 6-month period, the other invests in the remaining 6-month period. Account A is invested from November 1st through April 30th each year, while Account B is invested from May 1st through October 31st.
Here are the numbers:
• Account A portfolio grew from $10,000 to over $438,967. That is a 42-fold increase.
• Account B portfolio barely doubled to $22,659.
A chart outlining the above phenomenon going back to 1959 can be found here, but I thought I'd take some alternative looks.... one that goes back further in time (all the way to 1871) to see when this seasonality started and one taking a look at the real return (i.e. after inflation) of each leg since 1959.
140 Year Rewind
Using S&P data from Professor Shiller's Irrational Exuberance site, I constructed the below chart going all the way back to 1871. The outperformance of the November - April time frame since 1959 can be seen, but interestingly enough before that date both periods had almost the exact same performance. Begging the question... what changed around 1959?
![](https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgkkXfU9qf97JXSKgp67yyDJj3UxlpSgyZTwj0GXDew33Lp4fX-CXSfhQbs0LD3fDHin8qGD6FQUFiAfyYJmpqPUQh-SB52X2ZhMZNr6Ux6UK-kcfRAwav9CVxia9XedGNW-UmNtPGCtg/s800/seaonsal.png)
Real Seasonality (1959 - 2011)
You thought the original May - October figure looked bad in nominal terms? After inflation, total returns for that six month period over 52 years (312 months) were negative, while the November - April time frame posted annualized real returns ~10%.
![](https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEh34541FO13UpvFvW_LKrJ_iAy35sHiCwIkxo3zeonVLJheTTp9eKRMWbIJWMiBASJlGFsseLGNcb5UNC_1lnkbx32mKw0Z2Va8BjoosONr6S3EM_8KYZFYEddQxI9-199JhHM_Vq7XXQ/s800/seas.png)
Source: Irrational Exuberance
Amazing. Thanks Jake.
ReplyDeleteSuggest the analysis is redone by shifting the positions of the 6 month periods to see if the results can be improved and then from there break the year up into 3 periods and do the same. This might narrow down the specific features causing the result even more.
ReplyDeleteAnother look at this is at Does the S&P 500 exhibit seasonality throught the year? My conclusion was that the phenomenon probably is real, but it seems to be fading.
ReplyDelete