Everybody knows that money talks. In one corner of the internet, it tells the future. That’s PredictIt, where anonymous bettors can buy and sell contracts on political events, such as who will win an election or whether a bill will pass. Each event has its own market, framed as a yes-or-no question, where shares trade between 1 and 99 cents and the winning option is paid out at $1. The price arrived upon by traders buying and selling reveals what the wisdom of crowds expects the future will bring. As a user since PredictIt’s first year in operation, I have participated in dozens of markets as a hobby—and walked away with thousands of dollars in winnings for my trouble.
But now federal regulators are about to shut down PredictIt, which is actually an academic research project. I hope they change their mind—not just because I would like to continue deriving entertainment and profit from questions like, “Will any Republican Representatives who voted to impeach Trump be reelected,” but because PredictIt has become an invaluable source of information for journalists and the public about future political outcomes. Instead of shutting it down, regulators should change the rules it operates under to more fully and ethically harness the predictive power of market forces for society.
The forecasts on PredictIt have been referenced countless times by outlets around the world. In 2015, it was cited by the New York Times to argue that Donald Trump might really run for president, back when many dismissed the possibility. It was even mentioned in the 2018 Joint Economic Report to Congress. Many topics in politics command broad but mostly shallow attention; the theory behind PredictIt is that betting markets incentivize deep, reliable knowledge that can cut against—and outperform—popular perception. After all, the thinking goes, real people will lose real money if their predictions prove wrong.
A vibrant, geeky community has sprung up on PredictIt’s comment boards since its founding in 2014. Super-users dive deep into arcane political questions implicated by markets, often going well beyond the efforts of mainstream journalists. When the Iowa Democratic caucus collapsed into uncertainty last election cycle, PredictIt users doggedly called up union locals and nursing homes that had hosted caucus sites to ask how many voters had attended. The market assessed likely outstanding delegate tallies and correctly determined Pete Buttigieg held the lead. Similarly, when most English-language news organizations wrongly reported in November 2019 that Israeli Prime Minister Benjamin Netanyahu had been indicted, PredictIt users hired professional translators and correctly determined that he not been. (Not yet, as it turned out.)
But this generative ecosystem has as little as six months of life remaining. PredictIt, which is owned by Victoria University of Wellington, New Zealand and run through a contract with political resources firm Artistotle, Inc., is able to operate in the U.S. thanks to a no-action letter from the Commodity Futures Trading Commission. (PredictIt’s markets are considered to be binary options, a kind of futures contract that falls under the CFTC’s purview.) While the CFTC has rejected previous applications for political-outcome futures, it granted PredictIt a limited, revocable exception to run a small-scale, university-affiliated research project testing whether the real-dollar incentives of a prediction market can produce accurate, actionable predictions. PredictIt data has been cited in dozens of academic papers, with implications for the burgeoning economic theory of “superforecasting,” which states that financially motivated laymen can predict events with greater accuracy than credentialed experts. “Superforecasting” techniques could be brought to bear on policymaking—and PredictIt stands as one of the theory’s biggest tests to date.
The CFTC announced it was revoking PredictIt’s no-action letter earlier this month, effective this coming Februaary. In some sense, the decision was unsurprising. PredictIt had long since exceeded its intended small scale, growing to see tens of millions of dollars in annual trading volume across its hundreds of markets. PredictIt itself remains a nonprofit, but its vendor Aristotle, Inc. is likely earning millions of dollars in profits for operating the platform on behalf of Victoria University. The markets featured on the site have also become progressively wilder since inception: Some are effectively grim deathwatches, while for over a year PredictIt even ran markets on how many times Trump would tweet in a given week. (Bettors discussed coding alert systems to wake themselves up if the then-president started a tweetstorm at 4 a.m.). Should thousands of dollars in payouts really hinge on whether Trump would invite A$AP Rocky to the White House?
The CFTC did not give specific concerns in its statement, but it likely questioned whether PredictIt was legitimately being run as a nonprofit simply for research purposes. There are also broader moral risks posed by political betting. Users could develop gambling addictions and lose money they can’t afford to part with. Bad actors with strong financial incentives could try to not just predict but actively manipulate political events—when you posit what they might do to prevent an outcome, it quickly gets grim.
Of course, these concerns can feel a little silly in a world where legalized sports betting has surged to become a billion-dollar for-profit industry. Americans certainly seem comfortable with that risk to gambling addicts. (And to be fair, the CFTC can’t do anything about sports betting. The agency’s regulatory ambit extends to contracts “dependent on the occurrence…of an event or contingency associated with a potential financial, economic, or commercial consequence,” a standard which so far has been interpreted to include political events but not sporting outcomes.)
At any rate, the solution is not to shut PredictIt down. Instead, the CFTC should recognize the value the website provides, but alter the terms under which it operates to maximize the benefits it offers and minimize the risks it poses. The objectives of a new no-action letter should be to strengthen the efficiency and accuracy of the market, limit downside for individuals and gambling addicts, and allow companies to hedge political outcomes that may affect their business.
First, PredictIt should update its fee structure. Currently, the company charges a 10 percent fee on the net profits of any given trade. This means that a user that wins one bet at 50 cents but loses another at the same price will wind up losing money, since they will pay a fee in the market they won. This should be changed to instead be a dynamic fee on 10 percent of a user’s overall net winnings across all markets at any given time. This will lower aggregate fees and allow players to better arbitrage across related markets (thus improving pricing accuracy), while also placing the fee burden entirely on the net winners. According to economic theory, more efficient pricing should create more accurate predictions.
Second, PredictIt should be allowed to increase the current $850 limit that a user can wager in a given market. In private securities markets, investors who have sufficient wealth or experience, called “accredited investors,” are deemed to be sophisticated enough to buy assets that have not been publicly registered. If an accredited investor can invest $100,000 in a risky startup, they should be able to use that money on PredictIt too. This will allow the “superforecasters” who add the most knowledge to have a greater impact on markets.
Third, PredictIt should have a mechanism to automatically ban any user who loses more than a certain amount of money, perhaps $10,000. Such a limit would not impact casual users, but would provide protection for those who develop gambling addictions (which is indeed a major problem on sports betting sites like DraftKings and FanDuel).
Finally, PredictIt should be allowed to offer corporations participation in its markets, for the purpose of hedging political risk. An oil company may want to bet in favor of the election victory of a candidate who has pledged to tighten environmental protections. If the politician is elected, the company’s prediction-market winnings will help offset the potential hit to profits from greater regulation. This helps strengthen economic activity (by lowering political risk) and transfers money from corporations, who would in effect pay a premium for their hedge, to the individual bettors who take the other side of the trade and bring the market back into predictive alignment.
What would this amount to? Hopefully, a world where hobbyists like myself can participate in the most efficient, well-capitalized and ethical political prediction market there has ever been, providing glimpses of the future to everyone from voters to candidates to corporate board room members. Instead of shutting PredictIt down, the CFTC should work to preserve and improve it. You don’t need the wisdom of crowds to tell you that’s the right bet to make.