By Doug Pinkham
Public Affairs Council President
February 8, 2012
It’s often said that Washington is a town that defies logic. And the most common logical fallacy is that “correlation” means the same thing as “causation.” In other words, say many D.C. insiders, when two events occur at the same time, one must be the cause of the other.
That’s why we hear politicians proclaim that rising public cynicism has reduced voter turnout or that negative campaigning has increased partisanship. Each of these variables represents a real trend, but neither assertion is true.
Mixing correlation with causation becomes even more ridiculous when pundits start tying presidential politics to the financial markets. Investor’s Business Daily recently reported that when the stock market increases by 6 percent or more in January of an election year, the challenger beats the incumbent president. Apparently this pattern has held true since 1936. “Indeed, the stock market typically looks ahead and reacts before headlines become reality,” says IBD’s Paul Whitfield, creator of the IBD January Incumbent Barometer. “So there may be more at work than an uncanny streak.”
Tell that to Joe Weisenthal and Jon Terbush of Business Insider, who have traced the rise in the president’s popularity to gains in the S&P 500. Responding to a post in Talking Points Memo that tried to link President Obama’s higher approval rating to specific policy decisions, they suggest the real cause is the improving stock market. And they have visual evidence to back up their claim:
They’re both wrong, says Barry Ritholtz in The Washington Post, who argues that an incumbent’s poll numbers have very little to do with stocks. “Instead of assuming that one is causing the other, we need to look for broader forces that are driving both elements,” he writes.
When the economy is doing better and earnings, hiring and spending are up, voters feel more secure, which helps a sitting president. The opposite is also true: Markets go down not because a challenger is polling well but because economic conditions are weak.
The correlation/causation fallacy is rampant during elections, he notes:
These simple facts never seem to get in the way of the op-ed writers at various journals who seem to favor arguments along these lines: “Worries about possible policy changes are weighing on markets ahead of the year’s presidential elections. Candidate X’s rise in the polls is a risk that is giving the stock market jitters. Stock prices are wobbling, all leading to uncertainty. (And the markets hate uncertainty.)”
Here are some other logic-free observations Ritholtz says we’ll hear regarding politics and markets this year.
Misplaced credit and blame: Presidents get more credit than they deserve for strong economies and good markets, and more blame than they deserve when times are bad. “This is true regardless of which party wins the White House or where the economy is in its cycle.”
The Obama bull market: People who think the markets will respond positively to a pro-business president should ask themselves why stocks didn’t soar throughout George W. Bush’s entire term. (The S&P declined about 37 percent while he was president.) On the other hand, the markets have gained substantially since Obama was sworn in.
Anthropomorphizing markets: Markets don’t prefer one candidate over another, and poll numbers don’t spur short-term rallies. “It is a trick used to frame issues, and it is disingenuous at best,” writes Ritholtz. “Indeed, with these silly claims, pundits manage to combine all of the analytical errors discussed above.”
So why are people sucked in by irrational reasoning? Jonah Lehrer, author of How We Decide, argues that causal explanations are useful because they help us understand the world at a glance. This approach has worked reasonably well for centuries, but in both the natural and social sciences, “the reliance on correlations has entered an age of diminishing returns,” he writes in a recent article in Wired magazine.
First, all of the easy cause-and-effect relationships have been discovered, which means science “is getting harder.” Second, looking for correlations “is a terrible way of dealing with the primary subject of much modern research: those complex networks at the center of life.” There are simply too many variables, and they can’t be isolated.
And yet — in Washington and elsewhere — we continue to see two unrelated trends and conclude that one is causing the other. Perhaps, as the world’s problems have become more complicated, we keep hoping for easy answers. It’s clear the public has an appetite for this sort of thing. That’s why Investor’s Business Daily computes its IBD January Incumbent Barometer and why we see sports bloggers claiming that a president’s reelection chances are related to which team wins the Super Bowl.
But the best illustration of the correlation/causation fallacy came in a recent edition of Bloomberg Businessweek, in which Vali Chandrasekaran paired a series of trends with similar trajectories. Among the many “conclusions” he reached was one tying the number of active Facebook users to the yield on 10-year Greek government bonds:
“Correlation may not imply causation,” he writes, “but it sure can help us insinuate it.”
Comments? Email me.