Dylan B. Minor PhD, MS, CFP, ChFC, CLU, CIMA
“How come stocks always fall in the Fall?” Caroline Minor
My wife, Caroline, recently asked me the following question: “How come stocks always seem to fall in the Fall, and, it seems, in October in particular?” Indeed, the Great Crash of 1929 began in October. Black Monday in 1987 also began in October. The most recent 2007-2009 financial crisis bear market began in October. What’s remarkable to me, is that even though Caroline’s interest in the stock market rivals her interest in lampreys, she still identified this historical fact. It is as if bear-markets-start-in-October is some universal truth, known by all. Of course, we’ve enjoyed bear markets that have begun in months other than October. Nonetheless, this October has been another scary time for US stock markets. This bad-stuff-always-happens-in-October theme motivated me to dig a bit deeper to determine to what extent October (or any other particular month) should truly be feared in the financial markets.
In particular, I obtained SP500 monthly price data from July 1986 through a partial month of October 2014 (i.e., ending 10/23/2014). The average monthly return for this period was .73%. How scary is October on average? It turns out that the average monthly return for October was almost average: .63%. So what was the worst month? September was the worst month at an average loss of .6%. However, any good statistician will tell you that it’s important to remove outliers. In this spirit, we should probably remove September, 2001, as this includes the tragedy of 9/11. What if we remove that month? Then, instead, August is the worst month with an average loss of .42%. This naturally means that we should not invest in August. Except this year: The SP500 was up almost 4% in August this year, a close second for the best return month of the year. Unfortunately, in the end, what often seems like a pattern in the stock market proves illusory.
More generally, what often seems to be a pattern to the naked eye is often no different from random occurances. As an example, consider tossing a coin four times, and receiving the following pattern: Heads, Tails, Heads, Tails. How does this compare to instead tossing a different coin that yields the following pattern: Heads, Heads, Heads, Heads? An eyeball analysis, as I call it, would likely lead you astray by thinking that latter was less likely. However, in reality, they both have exactly the same chance of occurring: 1/16.
Returning to stock markets, it has been claimed that when the NFC football team wins the Super Bowl, it’s more likely that the US stock market will go up for that year. Further, this predictor is said to be correct 80% of the time (see the Wikipedia article here). This seems remarkable! To test this claim myself, I used all Superbowls listed in the article (from 2000 through 2013). I found that for this most recent period, the Superbowl predictor only worked about 57% of the time, much less than 80%. Meanwhile, what is the chance of a positive market any given year, regardless of who wins the Superbowl? About 65%.[1] An alternative, and more accurate indicator for this most recent period I studied would have been the following: if the planet Earth exists, US stocks will have a positive return year. Since 1900, this would have worked with 65% accuracy (and with 64% accuracy for this most recent period 2000-2013). In short, this Superbowl indicator is a classic case of confusing causation with correlation.
An even sillier example is what some call the Skirt Trade. That is, some suggest that you can predict the returns in the stock market by the average length of women’s skirts that year. And, of course, we could data mine all kinds of other factors to find surprising correlations. I’m thinking about next studying the amount of M&M’s in Omega’s M&M bowl at the beginning of the month as a predictor of that month’s return in the stock market. And I have many other candidate predictors. We are open to any suggestions that you might have, as well. Please let your Omega advisor know, and we’ll provide a prize for the best idea.
Completely missed in the previous discussion is that the typical Omega portfolio has only a minority allocation to SP500 types of stocks, many allocations even less than 25%. So we should really also be talking about predicting the returns of the other investments in Omega portfolios: bonds, emerging market stocks, real estate, merger arbitrage funds, managed futures funds, multi-strategy hedge funds, and so on. But since these are such different types of investments, many times, at least one investment will go up (and down) for the year. So, predicting that at least one slice of portfolio will go up and at least one down in a year will probably prove more accurate than predicting that the stock market will go up as a function of of who wins the Super Bowl, or some similarly dubious predictor. So where will US stocks end this year? Have a peek at our M&M bowl the next time you drop by the office…
[1] This figure comes from a proprietary Omega Financial Group study of the Dow Jones industrial Average from 1900 through 3/11/2013.