S&P Capital IQ's Sam Stovall appeared on Nightly Business Report on Friday, June 29, 2012 to provide some statistics on how the stock market tends to perform in the second half of the year. Here are the statistics he provided going back to 1900 (without indicating which index he used to calculate the returns of "the market"):
- The 2nd quarter is usually down
- About 80%+ of the time the market posts its high for the year in the second half.
- About 70% of the time the market posts its high for the year in the fourth quarter.
- Whenever the market has risen from July 31st to October 31st, the incumbent has won the Presidential election 80% of the time.
- Whenever the market has declined from July 31st to October 31st, the incumbent has lost the Presidential election 88% of the time.
The statistics on pre-election performance really caught my attention because I know that incumbents rarely lose U.S. Presidential elections. This fact was part of my case last year arguing that insufficient data exist to usefully correlate the unemployment rate to the outcome of Presidential elections. So, I decided to take a closer look at Stovall's analysis on pre-election results.
My historical data for the S&P 500 (SPY) go back to 1950. Unlike Stovall, I think the stock market (and economy) is much too different a beast before 1950 to compare well to the present era. I took three different views for the S&P 500′s performance in the July 31st to October 31st time period which I will call the "pre-election period": every year, election years, and election years featuring an incumbent (I included Gerald Ford's campaign against Jimmy Carter in 1976). It turns out that since 1950 this period tends to perform very well in election years and even better when an incumbent is running for re-election.
First, since 1950, 30 of the 62 years have delivered negative performance. 2008 of course stands out as the worst of the entire series. The tendency of the market to deliver major sell-offs in September or October skews this period to some large downside performance, but I find it instructive that taken over time, this period tends to be much more benign than I expected. For example, while the S&P 500 has traded in a wide range since 1996, the pre-election period has delivered a strong bias to upside performance. Six of the eight upside years featured gains of at least 5%. Five of the eight negative years featured losses less than 3%.
Performance of the S&P 500 From July 31st to October 31st Every Year Since 1950
The performance of the pre-election period has been stellar in election years since 1952. Again, 2008 sticks out like a throbbing sore thumb. Up until 2008, the pre-election featured great trading odds. Only the first three elections of this period featured significant negative gains. Seven of the pre-election periods delivered positive performance and two of those clocked in at 10%. (The performance for 2000 was -0.10%).
Performance of the S&P 500 From July 31st to October 31st During Election Years
Finally, with the data filtered down to just election years featuring an incumbent fighting to keep his job, the performance of the pre-election period REALLY stands out - mercifully, we can drop 2008 from the series. In fact, these data demonstrate that betting positively on the pre-election period is a winning strategy. Note well that in 1956, Dwight D. Eisenhower won re-election despite an awful performance for the S&P 500 in the run-up to the election. In Stovall's dataset, this victory looks like an oddity. Limiting the view to the post-1950 period, the oddity is not Eisenhower's victory but instead the poor performance of the S&P 500.
Performance of the S&P 500 From July 31st to October 31st During Election Years With A Running Incumbent
Source for data: Yahoo!Finance
What I particularly like about these results is that they add one more support to my conclusion that traders should buy the dips going forward. The lesson from this analysis is that using datasets that are too large can obscure much more interesting results. In this case, it is much more useful to think of the typical performance of the pre-election period under different scenarios rather than to think of this period of trading as providing an indicator of the eventual outcome of the Presidential election.
Be careful out there!
Disclosure: I am long SDS.