PredictIt is a current-events stock exchange that poses hundreds of questions — each its own “market” — at any given time. Users can plunk down as much as $850 to tap the next justice to leave the Supreme Court (Ruth Bader Ginsburg at 73 percent), forecast whether William P. Barr will still be attorney general at year’s end (91 percent say yes) or envision a 2020 presidential run by Dwayne “The Rock” Johnson (a hard no at 96 percent).
Then there’s Trump, whose impeachment battle has dominated the site. The odds of the Democrat-run House of Representatives voting to impeach him stood at 78 percent Wednesday morning, just ahead of the start of public hearings; they’d been as high as 80 percent in October. The site doesn’t see the Republican-held Senate voting for conviction.
John Aristotle Phillips, the Washington-based company’s co-founder, contends that the very nature of such platforms is a better picture of what people are thinking than what pollsters can extract from statistics.
“When people have a little skin in the game, they are more likely to distill facts from fiction and discount fake news,” he said. “The theory here is that markets can be effective at pricing future risk.”
An asset’s value is based on the size of the trading pool. Traders, much like their Wall Street brethren, make wagers based on their interpretation of news reports, public statements and other factors that can affect performance.
Only one option can be correct, and it’s winner-take-all.
Prediction markets seek to harness the “wisdom of the crowds” — the theory being that large groups of people are collectively smarter than experts and polls.
Though PredictIt’s spokesman Will Jennings says it’s difficult to make a “blanket accuracy assessment, given the variety of markets,” he said the D.C.-based site’s handicappers are correct 70 to 80 percent of the time. It forecast, for example, that Brett M. Kavanaugh would be on the shortlist to succeed Justice Anthony M. Kennedy on the Supreme Court. It also signaled that Kavanaugh would get the nomination hours before it was formally announced.
Though research generally supports the idea that prediction markets are better forecasters than polls, Justin Wolfers, a University of Michigan professor who has studied prediction markets, said to take that with “a big asterisk.”
“Saying markets are better than the alternative is not saying markets are perfect. Most ways of predicting the future are bad. It’s possible markets are the least bad.”
PredictIt was on the wrong side of Brexit, when the Britain invoked its separation from the European Union in a letter from then-Prime Minister Theresa May. Nor did it call Trump’s 2016 election victory; it said Democratic rival Hillary Clinton had a 95 percent chance of winning Florida, for example, but she didn’t. The site erroneously gave Martha McSally a 92 percent chance of winning Arizona’s 2018 U.S. Senate election. The former Air Force pilot lost, but was later appointed to the other Senate seat that had been vacated by Sen. Jon Kyl.
PredictIt correctly anticipated that the GOP would win the Senate and Democrats reclaim the House in 2018 midterms, that Milwaukee would host the 2020 Democratic convention, and that Boris Johnson would lead Britain’s Conservative Party.
But Washington’s political scene dominates the site, where 3,000 active traders are participating in markets every day.
PredictIt pegs Trump’s odds of reelection at 42 percent, which makes him the favorite one year out. As the incumbent, Trump has the historical advantage. Democratic Sen. Elizabeth Warren (Mass.), meanwhile, has a 17 percent chance, former vice president Joe Biden is at 13 percent and South Bend, Ind., Mayor Pete Buttigieg stands at 12 percent, according to the website.
Trump made his own prediction Tuesday during a speech at the Economic Club of New York. He told the business-friendly crowd that they would vote for him.
“The truth is look, you have no choice, because the people we are running against are crazy,” he said to laughter. “Okay? They are crazy.”
The Democratic presidential debates generated a lot of interest, especially after the Oct. 15 slugfest. PredictIt carried a number of active markets during that debate and continues to hold markets on who fared the best. Traders forecast that Bernie Sanders would get the biggest bump in the polls and that Biden would slide the most.
Though PredictIt comes across as quiz show meets Caesars Palace sportsbook, the five-year-old site doesn’t have bookies and oddsmakers like the sports gambling parlors that populate Las Vegas or local racetracks. There are no “betting lines” like those found on the Betfair Betting Exchange in Europe. And it’s not a traditional pollster.
Rather, it’s a gambling-polling hybrid that essentially asks people to back up their political prognostications with money.
“Polling asks you what you hope will happen,” Phillips said. “Investing in a market is what you actually think will happen.”
PredictIt allows traders to make predictions on future political events by buying 1- to 99-cent shares in binary outcomes. The resulting share price can be read as the probability of a specific event occurring, based on the collective intelligence of the market participants, according to the company.
The odds are set by the traders and what they are willing to pay. If a contract that pays $1 if “Candidate X” wins an election is currently priced at 53 cents, the market says “Candidate X” has a 53 percent chance of winning and 47 percent chance that they will lose.
PredictIt’s team formulates questions that must have two sides to the bet, known as a “contract.” Only one option can be correct, and it’s a dollar, winner-take-all.
The staff gathers every Monday to spitball what questions will be irresistible to traders. Will Pope Francis finish out his term? Will 2019 be the hottest year on record? Which party will take crucial swing state Ohio in next year’s presidential contest?
Political consultant Pratik Chougule is the author of the 2016 book, “How to Make Money from Political Predictions.” He agrees that prediction markets are more valuable than pundits, talking heads and politicians when it comes to discerning political outcomes. But Chougule said PredictIt has deep imperfections because of a demographic dominated by young, highly educated men. They are predisposed, for example, to gravitate to a presidential candidate such as Andrew Yang — a successful entrepreneur who’s polling at about 3 percent — as opposed to lifetime politicians such as Biden or Sanders.
“If markets are so rational, why are they putting such a high likelihood of Andrew Yang or Hillary Clinton winning the Democratic nomination,” Chougule said. “The market is overstating their odds because the people who are betting on political markets are not anywhere near a random sample of the population.”
Most of PredictIt’s 200,000 subscribers are well-educated millennial males from large cities on the East and West coasts. They earn $100,000 and $200,000 a year, according to the company. Many are employed in finance, law and politics. Not surprisingly, a big chunk of traders are in the fields of economics, statistics and mathematics.
Phillips, the company’s 63-year-old co-founder, first made national news in the 1970s as the “A-bomb Kid” when he designed an atomic bomb for a seminar while attending Princeton University. The term paper elicited contact from a Pakistani official looking for the plans, sparked attention on the floor of the U.S. Senate and led to a book called “Mushroom: The True Story of the A-Bomb Kid.”
Phillips later tried his hand at politics, twice unsuccessfully running for a congressional seat in Connecticut. During those campaigns, he spotted a business opportunity to develop technology around political fundraising and election compliance. That led to Aristotle, the company he co-founded with his brother, Dean Phillips, in 1983. Aristotle is the parent of PredictIt.
“Aristotle was a happy byproduct of my congressional campaigns,” he said.
The brothers branched out into political forecasting with PredictIt in 2014 after teaming up with Victoria University of Wellington, New Zealand. Victoria University created and operated the site that Phillips introduced to the United States.
“There are limitations on polling that can somewhat be addressed by something like a market,” Phillips said. The question of whether markets are more accurate than polling has been the subject of intense academic research, and that is why Victoria University was studying it.
The university received a letter from the U.S. Commodity Futures Trading Commission giving PredictIt permission to allow trading on the site at a limit of $850 per trade. The cap is designed to stop a trader from skewing the prices by flooding the bets with cash.
“Instead of trying to arrive at a price of oil 12 months from now, you are using the same market forces to ascertain who is going to win an election,” Phillips said.
The company makes money, but not a ton, Phillips said. Its revenue comes from a 10 percent fee on the profit of any trade. There is also a 5 percent fee when a trader withdraws money from their account. There is no charge to open a PredictIt account or to deposit funds.
Jason Hemmendinger, a software engineer and top predictive trader, has been trading since June 2017, when he created a PredictIt account with $40. He said it hit five figures after the 2018 midterms.
“I have made some mistakes, and I was below $5 at one point, but I kept at it and rebounded,” said Hemmendinger, who is probably more proficient in current events than the average citizen.
Hemmendinger said his biggest PredictIt “gets” were predicting that North Korea would participate in the Olympics and that Trump would meet with North Korean leader Kim Jong Un. He also guessed correctly that Sen. Lisa Murkowski (R.-Alaska) would be the only Republican to oppose Kavanaugh’s nomination.
“It’s arrogant to say, but I’m right more often than I feel,” he said.
Hemmendinger checks dozens of prices every day, and he sometimes spends hours on PredictIt. He isn’t sure that the site is a great prediction vehicle, but he enjoys playing the game.
“Constraints on PredictIt’s market structure, combined with participants self-selecting, can lead to distorted prices and therefore profitable plays,” Hemmendinger said. “You can even tell people the correct play and they will still throw their money away.”