Everyone “knows” that a Joe Biden presidency would be bad for the economy and the stock market.
And they very well may be right, of course. But, as Humphrey Neill, the father of contrarian analysis, constantly reminded his clients: “When everyone thinks alike, everyone is likely to be wrong.”
Humphrey’s admonishment inspired me to subject the conventional wisdom to statistical scrutiny. We are able to conduct such an analysis because of data provided by online gambling sites, such as PredictIt.org, which allow traders to invest in futures contracts tied to various outcomes. Because those contracts trade continuously, it’s possible to analyze changes in their prices with contemporaneous changes in the stock market.
If the stock market were to plunge every time the Joe Biden contract rises, for example, then that would be strong statistical support for the conventional wisdom.
Not to bury my lead too much: Such a correlation does not show up in the data. Consider the accompanying chart, which plots the S&P 500 over the past three months against the price of the PredictIt.org contract pegged to Biden’s winning the election.
Eyeballing the data, you might even be inclined to conclude that the stock market looks favorably upon a Biden presidency. That’s because both the Biden contract and the S&P 500 fell through late March, and since then both have generally risen.
Nevertheless, the jury is still out, according to Eric Zitzewitz, an economics professor at Dartmouth College. Two decades ago, Zitzewitz helped pioneer the use of electronic futures markets to predict the stock market’s eventual reaction to certain events.
In an email, Zitzewitz explained that the past three months are an especially “tricky” one in which to draw any firm conclusions about cause and effect. Take the apparent correlation between the Biden contract and the S&P 500 in the days leading up to the bear market bottom on March 23. “Almost surely most of that correlation was due to COVID-19 causing both an improvement in Biden’s odds and a recession (not to the improvement in Biden’s odds causing the recession/market decline).”
Zitzewitz says that it’s equally hard to draw any conclusions from the apparent correlation since then. “Somehow, Trump has figured out a way to not benefit politically. But even if he had, and there was a correlation between the market and Trump’s odds, I’d interpret it as mostly due to the decline in COVID-19 risk causing both, rather than the improvement in Trump’s odds helping the market.”
Conducting these statistical tests has not been worthless, however, even if no firm conclusion emerges. Because there is no unambiguous pattern in the data, we are betraying our own political biases when we confidently say what the stock market would do if Biden wins or loses. If we’re being rigorously empirical, we need to say that it’s impossible at this point to know.