For several years, I was a regular guest on a cable news financial talk show where the conversation often turned to political matters. So, I found myself boning up on the historical relationship between politics and the stock market - for reasons of self-preservation more so than intrinsic interest.
The recent elections have brought this topic to the fore once again, with both sides making completely ridiculous claims about the effect their party has had on historical equity returns.
Here's the short version for those readers pressed for time: There is not one shred of conclusive evidence that stock market returns can be predicted by knowing which party controls Congress or the White House.
If you're still with me, here's the longer version: Since 1900, the Dow Jones Industrial Average has returned about 8 percent per year after inflation under Democratic presidents, and about 5.6 percent per year after inflation under Republican presidents. Digging a little deeper, the best returns - 9.5 percent per year after inflation - occurred when a Democratic president was accompanied by a Republican Congress, and the worst returns 4.7 percent per year after inflation - occurred when we had a Republican President and a Republican Congress. (Source: Ned Davis Research via a piece published by Edward Jones.)
But wait if we assume that Congress is what matters most, since it controls the public purse, we get a different result. The average after inflation return under Democratic Congresses was 6.2 percent per year, compared with 6.5 percent under Republican Congresses. And if we lag the data by a year, assuming policies take time to be implemented, the Republicans start to pull ahead.
The bottom line: If you tell me your political leaning and the desired result, I can torture the data until it confesses whatever you want it to say. Depending on whether you adjust for inflation, use total returns or price returns, analyze who controls Congress versus who controls the White House, or assume the market reacts instantly to elections versus requires some lag period for policies to take effect the data can show pretty much anything.
This should not come as a surprise. First, the equity market is an incredibly complex system, composed of tens of millions of agents with different sets of preferences whose behavior is driven by both external data and the concurrent behavior of other investors. It beggars belief that such a complex system could be predicted by just knowing the political allegiance of one politician. (Or 535 politicians, if you think Congress is what matters most.)
Second, the U.S. equity market is a global system 40 percent of earnings from the S&P 500 now come from outside the U.S. Why would Starbucks decide whether to open more or fewer locations in China based on who's in the White House?
Finally, and most importantly, there have only been 28 presidential terms since 1900. Any statistician would laugh out loud at the prospect of making a robust prediction based on only 28 data points. It's just not a large enough sample. Would you try to predict an election after polling 28 people? (Even if we assume that each House election creates a new political context, we still only have 56 data points.)
If the party in power has little (if any) impact on stock market returns, what does? The fine folks at Vanguard took a look at this question recently, and generated some empirical results that confirm my own horse sense. The full paper is here (it's quite good), and the takeaways are:
n Nothing predicts the market over the short run.
n Many commonly-used predictors GDP growth and profit margins, for example do a poor job of explaining long-run equity market returns.
n Just about the only thing that does even a halfway decent job of forecasting long-run market returns is valuation lower price/earnings ratios have generally presaged periods of better equity returns, and higher price/earnings ratios have generally presaged periods of poor equity returns.
So, the next time you hear a pundit trying to claim that Party A or Party B is able to sprinkle pixie dust on the S&P 500, remember that long-run market returns are driven by three things, and three things only: Earnings growth, dividends, and valuation. If you pay a reasonable price for companies that grow and pay dividends, you've hit the trifecta, regardless of who's in office.
After all, if beating the market were as simple as watching election returns and investing based on who wins, wouldn't the money management industry be a lot smaller?