The Idiocy Of Prediction Markets Illustrated With 2016 Republican Contest
For a while now, there’s been an economic school of thought that supposes that people in crowds are inherently better decision makers than individuals can be. This idea comes out of the capitalist faith in an invisible hand that somehow guides markets to the optimum outcomes.
For over a decade now, this idea has been applied in the political sphere, with prediction markets, where people bet on the chance that certain political events, such as the victory of a candidate in a particular election, will take place. In these markets, people can trade options on these outcomes, so their value fluctuates as the invisible hand of the market moves slowly but surely toward its optimized prediction.
That’s the theory, anyway. The reality, as shown in this graph from a prediction market for the 2016 Republican presidential nomination shows.
The graph, from ElectionBettingOdds.com, shows that just a half a year ago, the prediction market gave Donald Trump a 10 percent chance of winning the Republican presidential nomination, and said that Jeb Bush was most likely to win the race. However, by the beginning of this month, the market had changed its mind almost completely, giving Trump an 80 percent chance of victory in the Republican contest. Now, after Donald Trump’s latest round of off-the-cuff absurdities trampled on the GOP’s holy-of-holies (abortion), the prediction market gives Trump just over a 50 percent chance of winning, although the chances of Trump’s main opponents, Ted Cruz and John Kasich, hasn’t grown by much in response. Apparently, the markets believe that some mysterious person not currently running for President has a good chance of winning the Republican nomination.
The current dip really isn’t a low point for Donald Trump, as far as this prediction market is concerned. Trump’s likelihood of winning the Republican nomination, in the market’s eyes, plummeted from 50 percent to below 25 percent back at the beginning of February. Remember what happened then? Trump lost the Iowa caucuses.
The rapidity with which the market’s predictions changed then reveals what kind of market it really is. It’s not making predictions. It’s showing reactions. The market supposed Donald Trump was the favorite on the day before the Iowa caucuses, and then, suddenly, after Trump lost that contest, the market had a new version of reality to work with.
The repeated failure of ElectionBettingOdds.com to provide a consistent predictive view of the Republican presidential race shows that there really isn’t any invisible hand guiding markets in a reliable way. Crowds of people don’t have particularly reliable, or even constant, predictive insight. What we see in this prediction market is about the same as what we would get from watching the TV news, or asking random people on the street what their opinion is – and both of those sources are frequently inaccurate.
There isn’t any magic that happens when you expose an idea to a marketplace. A marketplace, it turns out, is just a bunch of people making guesses about the future, based on inaccurate and incomplete information.
They’re gamblers, trading options based on what they think the odds are on any particular day.
If there’s anything we know about gamblers, it’s that, on average, they’re wrong. On average, they lose.
Why should we hold any trust in groups of gamblers like those at ElectionBettingOdds.com to predict what our political future will bring?