It’s never too early to start thinking about next year. Sure, we still have a holiday season and a Fed decision, among other fun-filled events, to get through before 2015 rings out. But now is when we start to focus on our strategic allocation models and plan out scenarios for what the year may, or may not, have in store. Unfortunately, the calendar-gazing habit tends to bring out the silly side of investing. No small number of “experts” habitually hold forth on the airwaves with a gravitas-laced analysis of the January effect or some other variation of “when X (calendar event) happens, expect Y (stocks to go up, down, all around)”. Next year is 2016, meaning of course that it is once again time to elect a US president. And presidential years take calendar-centric silliness to a whole new level. Not only is there an imagined pattern to presidential election years, as the pundits would have it, but there are “rational” explanations as to why this is so. Fact or fable? Let’s do what we always do, and examine the data.
Presidential Years and Market Math
It is true that recent elections have not coincided with great times for investors. We looked at rolling twelve month returns for the Dow Jones Industrial Average going back to 1900, and found that the average nominal return for this period (through the rolling twelve month period ended September 30, 2015) is 7.7 percent. Compared to that average, the twelve month periods coinciding with the calendar years of 2012, 2008, 2004 and 2000 have all been below average (3.7 percent, -33.8 percent, 3.1 percent and -6.2 percent respectively). Case closed, right?
Not so fast. Humans have a tendency to seize on easily-recalled data and make generalized conclusions from them. It’s called the “availability bias” in the lexicon of behavioral economics. Most of us have a better recollection of the Obama-McCain election of 2008 than of, say, the Hoover-Smith campaign of 1928. That does not make the 2008 outcome a whit more statistically relevant than that of 1928. And in case you are wondering, the Dow delivered a 48.2 percent return that year, at the giddy height of the Roaring Twenties.
In fact, the average return for all presidential election years since (and including) 1900 is 7.3 percent – statistically indistinct from that 7.7 percent average for all 1,387 rolling twelve month periods since January 1900. Please keep this in mind the next time some portentous pundit solemnly holds forth on the “facts” of presidential years and the stock market. Lots of things will impact the market next year, of which the election will be just one variable.
The Law of Small Numbers
How about mean reversion? Are we “due” for a better-than-average election year given the recent streak of poor performance? Sadly, no. The presence of an improbable recent streak does not affect the odds of the next outcome. If you roll a die and it comes up with the six facing ten times in a row, the probability of getting a six on the eleventh roll is still 1/6. The law of large numbers says that mean reversion works over time, with large numbers of observations. The law of small numbers, by contrast, cautions against reading anything meaningful into a small handful of outcomes.
So where does that leave us as we look ahead to 2016? As we noted in our commentary last week, we see a reasonably compelling case for further upside before the current bull run peters out. That could take the form of leadership by high quality stocks or an end-of-cycle “melt-up” (or something else entirely). Among the factors we consider critical are: monetary policy divergence, emerging market growth trends, commodity prices, and geopolitical flashpoints in the Middle East and the South China Sea, among others. Yes, we will also be closely following the US election, the policies articulated by the contenders and the sentiments of the voting public. But we discount any kind of a foregone effect the election year will have on market performance.