Punditry is outdated. We have all the tools necessary to see the future, but we’re leaving them idle. Prediction markets should be legalized.
Punditry Is Broken
There is no need for punditry in 2017. Individual pundits – even “experts” – simply don’t have the insight their positions hint at.
2016 saw expert predictions fail en masse. Brexit, the Presidential election, the Premier League, the NBA Finals, and so on. All coincidences, but all results that shattered the expectations of the punditocracy.
It wasn’t last year alone – we’ve been treated to failed prediction after failed prediction in newspapers and magazines for decades. Not that there’s anything wrong with that – everyone should hope to be more wrong. But as Nassim Taleb notes in The Black Swan (review here), experts have failed to predict almost all of the most impactful events we’ve seen as a society.
We’ve managed to connect billions of people across the entire planet. Why base our predictions on the voices of a flawed few?
What’s A Prediction Market?
Prediction markets pit speculators against one another. Open questions are posted, and participants take a position on the issue from a set of defined options. Only one option will turn out to be true – those that selected that option are rewarded with the spoils from those that picked every other option.
Here’s a hypothetical example: a question could be posted, such as “Will the US economy grow or contract in 2017?”. Two contracts would be offered: a “grow” contract, and a “contract” contract. Participants can choose to buy or sell these contracts, so long as there is another participant on the other end of the transaction. The total value of each contract is $1.00 once the market closes – but only one of the participants receives the value. So contracts are priced in cents – fractions of a dollar. For example, this market could reach equilibrium at $0.70 grow – $0.30 contract. For every grow contract I purchased, I would get paid $0.30 plus my money back if the economy ended up growing. If the economy shrunk, I’d lose my $0.70 stake. It works in reverse, too – the less optimistic participants would get $0.70 on top of their $0.30 if they turned out to be right at the end of the year.
Choices don’t have to be binary – a market can offer many options, such as “Who will win the Democratic presidential nomination in 2020?”. But the market is zero sum – there can only be one correct answer for any given question.
Prediction Markets Make Punditry Obsolete
Have you ever been talking shit with a friend making an outrageous prediction? Ask them to put money on it, and they’ll often back down. This same mechanic is at work with a lot of pundits – who make predictions to satisfy some need other than being right. For example, some pundits are cheerleaders for a position, just trying to lift the spirits of those who agree with them. Others are intentionally inflammatory to attract attention. This type of punditry is excellent memetically – it attracts eyeballs in the war for attention – but it is of little value for accurate forecasting.
The reasons why prediction markets are superior for prediction are summarized by a couple simple truths:
- Many people provide a better prediction, on aggregate, than a single individual does (a.k.a. the wisdom of the crowd)
- A person with money down is more honest than a person without money down (a.k.a. skin in the game)
It follows that many people with money down will almost always be more accurate than any individual can be. Sports fans know that Vegas odds are better predictors than talking heads on ESPN. So a widely-used prediction market would give us an excellent view into what is to come, using the aggregated skin-in-the-game reasoning abilities of millions of people.
One note: prediction markets themselves failed with Brexit and Trump (yours truly made some money betting against them). Part of the reason for that is that we are still so reliant on phony talking heads and motivated reasoning in media to separate wheat from chaff. Brexit/Trump were such “shocks” because the media had led us to believe that they were unpopular, when, in fact, plenty of people supported the eventual winners.
Even then, the pundits didn’t beat the markets. In fact, prediction markets routinely match or outperform the vast majority of pundits. Punditry is one of the only jobs where massive failure has almost no consequences. How many political pundits have failed on almost every major issue in the past 20 months, yet are still in a job?
Experts themselves suddenly get better at predicting when money is on the line. A study showed that experts correctly predicted whether psychological studies would be replicated just 58% of the time; give them money to bet on the results, and that figure rose to 75%.
Punditry as we know it is obsolete, but prediction markets can revive it. Widespread adoption of prediction markets would give pundits the option to display their skin in the game. A pundit could make his prediction as usual, but also reveal the amount they’ve bet on that result. This creates the opportunity for costly signaling.
Prediction markets would give us our very best glimpse into the future:
- Crowd wisdom, enhanced by skin in the game, would be on display for all to see, and all to influence
- Any expert or pundit who really did have special insight could broadcast their confidence level by publicly committing funds to a contract. The level of confidence would depend on how much money, relative to their means, they were willing to commit. And true experts would accrue a visible history of not only what they predicted, but how confident they were, quantified and on record.
The Problem: They’re Illegal
Prediction markets are, ludicrously, illegal in the United States. The sole operator open to the public is PredictIt, which caps total funds at risk to $850, and lacks an option to share predictions. The prohibitions are a holdover from anti-gambling law. The result is that we’re flooded with bad predictions across the media landscape, with no easy way of sorting the signal from the noise.
There’s valid concern that prediction markets would become a vehicle for gambling on everything. Gambling has its own problems, and I won’t wade into that swamp here. But legalizing prediction markets is distinct from legalizing all gambling. We can debate where we want to draw the line, but there’s room between the positions for it to be drawn somewhere. For example, we could restrict markets in things people typically gamble on, like sports.
There are a couple side benefits to legalizing prediction markets:
- An end to professional bullshitters in the media – show me the money, or shut up
- An opportunity for smart but poor people to increase social mobility – bet against the rich and stupid, and climb out of poverty
But the main benefit is to enhance our predictive ability as a society. We have the research proving that an incredible tool to predict the future is sitting in our laps. But we’re ignoring it, at great cost.
I participate in prediction markets wherever possible, to keep me honest. I recommend it! I also recommend joining thousands by subscribing for email updates and following me on Twitter, where I also post daily.