A New Prediction Market Lets Investors Bet Big on Almost Anything

Listen to this story

“Congratulations,” boomed Heath Tarbert, a Republican who’d been appointed by President Donald Trump the previous year. “You now stand with markets that have been around since the 1840s. And I have no doubt that in time you’ll grow to be a powerhouse, too.”

Companies had been trying to introduce similar so-called event markets in the US for years, but Tarbert’s agency, the CFTC, had always said no, arguing they were tantamount to gambling and vulnerable to cheating. Now the agency had reversed course, giving its approval to two 24-year-olds who would enjoy first-mover advantage in what could eventually become an enormous new asset class. Within weeks of Tarbert’s call, which a representative for him says was customary, their startup, Kalshi Inc., raised $30 million from venture capitalists. “We believe this will be bigger than crypto,” says Mansour, now 26.

The potential of prediction markets is well known to anyone who’s read James Surowiecki’s bestseller, . Well-designed markets can help draw out knowledge contained within disparate groups, and research shows that when people have money on the line, they make better forecasts. “It’s a tax on bullshit,” Lopes Lara says. That explains why Google, Microsoft Corp., and even the US Department of Defense have used prediction markets internally to guide decisions, and why university-linked political betting sites such as PredictIt, where wagers are capped at a few hundred dollars, sometimes outperform polls.

The size of these markets had been limited because regulators worried that Wall Street-scale trading could create incentives for investors to meddle with reality. If the stakes are high enough, it’s not difficult to imagine traders leaning on congressional staffers to stall the passage of a bill, or someone in Kanye West’s orbit betting on whether the rapper’s latest album will drop this week. These were among the issues that senior figures within the regulator flagged when Lopes Lara and Mansour pitched the CFTC. But their concerns were overruled by the agency’s politically appointed commissioners, one of whom has since joined Kalshi’s board.

That Kalshi prevailed was less a testament to Silicon Valley-style innovation than it was to persistent lobbying and legal wrangling. Lopes Lara and Mansour didn’t invent anything from scratch; they took a well-established concept and forced a change in the way it’s governed. The result, depending on whom you ask, could usher in a new era of market-based enlightenment, or it may push Wall Street’s most destructive tendencies even further into the real world.

If Kalshi’s founders lacked experience to bolster their application to the CFTC, they’d at least enjoyed an almost comical amount of youthful success. Lopes Lara studied ballet at the Brazilian Bolshoi before moving to the US on a math scholarship. Mansour finished second in France’s math Olympiad. They met in an advanced computer science class at MIT, where they bonded over their shared work ethic.

Lopes Lara spent part of her senior year interning at Five Rings Capital, a hedge fund in New York, when the idea for Kalshi came to her. Whenever the traders around her weren’t working, she noticed, they seemed to be gambling with each other on whatever was going on in the news: Would Apple’s market valuation hit a trillion dollars? Was Kylie Jenner pregnant? “It could have been absolutely anything,” she recalls.

When Lopes Lara suggested to Mansour this might be a business, he saw the potential right away. He’d interned at Goldman Sachs Group Inc. on a desk helping investors protect themselves against the possibility of the UK leaving the European Union. Goldman sold clients complex combinations of stocks and derivatives. As he talked it through with Lopes Lara, they both agreed that surely it would be better to let investors hedge their risk by betting on Brexit itself rather than some imperfect proxy.

Over time, Lopes Lara and Mansour found themselves hypothesizing about how such a marketplace might work in practice. What they settled on was called an “event contract,” an instrument tied to a question with a binary outcome, such as “Will inflation hit 5% by the end of the month?” The contract would eventually settle at either $1 (if the event happened) or zero (if not), but before that its price would fluctuate, reflecting how likely the market considered the occurrence. A contract on whether a politician will be elected, for example, might jump from 50¢ to 55¢ after a particularly good debate performance. Kalshi would make money by charging a commission on every trade and potentially by selling data to traders, political campaigns, businesses, and anyone else hoping to get a glimpse into the future.

In October 2018, five months after graduation, the pair flew to California to compete in a hackathon for wannabe tech founders organized by the Silicon Valley incubator Y Combinator. Over the course of a day and a night, they built a website and presented it, bleary-eyed, to a panel of entrepreneurs the following day. Their prototype barely worked, but they still found themselves among the winners, which meant they’d earned entrance into a three-month mentorship program and a $150,000 investment. A Y Combinator managing director, Michael Seibel, told them their idea was “so unlikely to succeed, I had to take a chance!”

Seibel’s skepticism—and, paradoxically, also the reason for his interest—was rooted in America’s historical wariness of anything resembling gambling. Roulette, poker, and other online casino games remain largely outlawed, and, until a May 2018 Supreme Court decision, sports betting was permitted in only a handful of states. Setting up Kalshi as a place for hedging risk rather than as a bookmaker seemed to be a possible way forward, but convincing the CFTC wouldn’t be easy. (Rejecting one proposed exchange in 2012, the CFTC wrote that trading on politics had no “economic purpose” and was “contrary to the public interest.”)

One day during their time at Y Combinator, Lopes Lara and Mansour say, they cold-called 60 lawyers they’d found on Google. Every one of them said to give up. “It was painful,” Mansour recalls. Eventually they tracked down a former CFTC official, Jeff Bandman, who assured them the landscape was changing; he agreed to help them navigate the agency and its characters.

When they weren’t busy trying to recruit lawyers, Lopes Lara and Mansour were meeting early-stage investors. Among them was Alfred Lin of Sequoia Capital Operations LLC, who was known for backing companies that eventually overcame government scrutiny, including Airbnb, DoorDash, and Uber Technologies. Lin told the founders he saw potential in their idea, both to capitalize on the explosion in retail trading and, one day, even to challenge how the financial world manages risk. “Come back when you’ve got regulatory signoff,” he said.

For a long time, speculation on most events wasn’t possible in the US, no matter how small the wager. Under the Commodity Exchange Act, the CFTC has the power to stop exchanges from listing contracts relating to “terrorism, assassination, war” and “gaming” if it deems them to be “contrary to the public interest,” which turned out to be the case a lot of the time.

In 1988, as academic interest in the field increased, the agency granted a legal exemption to the University of Iowa, allowing it to set up a prediction market for research purposes—provided it didn’t turn a profit or advertise—and it limited bets to $500. PredictIt, the biggest and best-known political betting platform in the US, also got an exemption thanks to an association with Victoria University of Wellington in New Zealand, but it’s been less cautious about expansion. Today it’s a sprawling marketplace full of professional traders with their own subculture and lingo. (For instance: “Rules Cuck Panther” is what PredictIt users call it when a bet goes against them on a technicality.) PredictIt users trade tips on Discord and the podcast, and major news outlets cite the website’s odds.

In deference to the CFTC, bets on PredictIt are limited to $850. But to keep traders happy, PredictIt will often run multiple variations of the same question—listing separate contracts for two dozen or so Democratic primary candidates, for example—meaning a trader could have more than $10,000 riding on a single outcome. The site’s markets are mostly suggested by the traders themselves, some of whom are current or former campaign staffers who might be well positioned to answer questions such as “How many tweets will Donald Trump post from Nov. 20 to 27?” or “When will Anthony Scaramucci’s role as White House communications director end?”

According to PredictIt co-founder John Phillips, the prevalence of politicos helps explain why the site so often gets it right. “Prediction markets function as well as they do and are as accurate as they are … because these markets, like any markets, attract people with superior information,” he said on a podcast in 2016. “In the financial stock market, it’s called inside information.”

Beyond academia, of course, trading on nonpublic information can be illegal, which represented a conundrum for Lopes Lara and Mansour. On the one hand, they wanted Kalshi’s forecasts to be accurate. On the other, as a regulated entity, Kalshi would have a responsibility to stamp out insider trading across its markets. In essence, Lopes Lara and Mansour wanted to build a vast high-stakes version of PredictIt, but with none of the anarchy or blurred legal lines—a “New York Stock Exchange for Events,” as they pitched it. First, they needed to convince the regulator that event trading could even be safe.

When Lopes Lara and Mansour approached the CFTC in the spring of 2019, some officials in the part of the agency responsible for reviewing their application, the Division of Market Oversight, were skeptical, according to interviews with people who were directly involved in the process and spoke on the condition of anonymity because the details are confidential. There was a sense among the DMO’s seasoned regulators that, for all Kalshi’s talk of revolutionizing finance, this was just a turbocharged iteration of something that had previously been rejected, and with good reason.

But it wasn’t the DMO’s job to look at the big picture. The staff review was supposed to be limited to ensuring Kalshi could complete a checklist, “23 Core Principles of a Designated Contract Market,” which included keeping good records and having the requisite financial resources, among other items. The five commissioners would then make a decision. With Trump in the White House, three of them were Republicans ideologically predisposed to the further proliferation of markets into American life.

For more than a year, Lopes Lara and Mansour, along with their lawyer Bandman, the ex-CFTC official, answered the DMO’s questions while lobbying the commissioners on Zoom about the potential of event markets to help the world mitigate risks and make better decisions. Before each meeting, they would write a script and memorize it word for word.

Kalshi’s case was bolstered by the emergence of a number of prediction markets that hadn’t bothered to try to get regulatory approval. One upstart, Polymarket, let its customers bet hundreds of thousands of dollars anonymously using cryptocurrencies, making it difficult to track. Another market, Augur, which facilitates private wagers between parties using the blockchain, couldn’t regulate bets at all and thus hadn’t stopped users from betting on whether public figures would be assassinated. Kalshi, by comparison, argued it was doing everything right. (The CFTC would go on to fine Polymarket $1.4 million in January 2022 for running an unlicensed exchange. Polymarket says it’s now complying and is “excited to help pioneer the next phase of smart contract-based financial solutions in collaboration with regulators.”)

Despite the overriding reservations of some DMO members about the riskiness of event contracts, Kalshi demonstrated it could meet the 23 criteria and was approved unanimously. “Once they check all of the boxes, they’re in,” says one person with direct knowledge of the CFTC’s review process.

In February 2021, three months after getting CFTC approval, Kalshi issued a press release announcing the funding from Sequoia, which was joined by a group of investors that included Charles Schwab and Henry Kravis. “Tarek, Luana and team have created an entirely new model for how we invest in and engage with what’s happening in the world,” wrote Sequoia’s Lin, who had joined the board.

Up to this point, the CFTC’s process for vetting their exchange hadn’t involved asking exactly what markets they planned to run. That discussion came after approval, by which stage momentum was already firmly on Lopes Lara and Mansour’s side. When Kalshi sent across a preliminary list of 30 proposed contracts in March, it unleashed chaos within the already overworked DMO. The division was set up to deal with exchanges that might create two or three new markets a year. Kalshi’s business model called for new ones practically every day.

Some of the proposals were uncontroversial, such as questions tied to the weather or gross domestic product. Others, among the initial list and submitted subsequently, seemed troubling. DMO officials worried that contracts tied to Covid-19 numbers, for example, amounted to gambling on human suffering, which is one reason markets on war and terrorism are prohibited. (Similar logic doomed ex-admiral John Poindexter’s Policy Analysis Market, a controversial George W. Bush-era plan to uncover intelligence by getting security analysts to bet on events in the Middle East.) Regulators also couldn’t see how speculating on who would win the Grammys, say, was any less a form of gambling than betting on the New England Patriots to win the Super Bowl. Futures contracts are supposed to allow traders to protect themselves against economic risk, and it was hard to understand who, apart from perhaps John Legend, might need to hedge the winner of best R&B album.

For the dozen or so members of the DMO’s product review team, event contracts raised novel questions. Under the Commodity Exchange Act, the regulators could block contracts tied to gaming that weren’t in the public interest, but no one had taken the time to define what gaming actually was. Beyond that, it was unclear whether the CFTC had a right to consider whether a contract was in the public interest or if the agency had an obligation to do so in every case. And how was it supposed to decide what was in the public interest anyway? “To be honest, the whole thing was a mess,” another person with direct knowledge of the CFTC review says. The agency declined to comment for this story.

An even bigger worry among CFTC staff was that some event contracts might be vulnerable to cheating. Kalshi, for instance, wanted to run a market on whether bees would be placed on the endangered species list. The DMO pushed back, saying it saw two problems that were symptomatic of the whole asset class: one, that traders might try to press government officials for information; and two, that the officials themselves could delay adding the insects to the list so they could cash in.

The idea that traders might be tempted to interfere in prediction markets to influence outcomes wasn’t some paranoid fantasy. In 2013 academics David Rothschild and Rajiv Sethi discovered that an unidentified party had lost $7 million buying Mitt Romney contracts on Intrade, a now-defunct, unlicensed Irish platform, in the runup to the 2012 election, even after it was clear that the candidate’s prospects had faded. The authors speculated that the trader, whom they dubbed the “Romney Whale,” may have been looking to boost morale and keep donations coming in.

Kalshi played down such risks, noting that manipulation and insider trading were concerns for any market. It had built a surveillance system and said it would hire a team to monitor for anything improper. “People trade on events all the time—they just use options and other instruments that are harder to track. This is a way to bring all that into the open,” Mansour says, summarizing the argument. Kalshi also didn’t include contracts on elections, which the Democrats on the CFTC might have considered a red line.

By that summer, Lopes Lara and Mansour were poised to introduce kalshi.com, but they remained locked in a stalemate with the DMO. Product review staffers, meanwhile, were exasperated at spending half their time dealing with an exchange that represented such a minuscule proportion of the total derivatives market. During the impasse, Lopes Lara and Mansour continued to press the politically appointed commissioners.

By now, Tarbert, the chairman, had moved on, but Kalshi found another advocate in Republican Brian Quintenz, a crypto-loving former hedge fund manager and staunch free marketeer. He was unmoved by the DMO’s concerns, arguing not only that speculation on the kinds of events that Kalshi proposed was desirable, but also that the agency had no legal standing to prevent it. Earlier in the year he’d lent his support to an unsuccessful bid to allow futures tied to NFL games. Others on the commission were more circumspect, but all were broadly supportive. And given the ambiguities in the law, they worried they would be on shaky ground if they ever tried to block a contract and Kalshi sued. With no permanent chairman in place, there was a vacuum at the top of the agency.

To “stay,” or block, a contract, DMO staff required the backing of a majority of commissioners, and it became increasingly clear that, in all but a handful of cases, they didn’t have it. “To paraphrase ,” one of the people familiar with the review says, “we didn’t have the votes.” By the second half of 2021, new contract requests were arriving at the DMO almost daily, and eventually the division, demoralized and overrun, accepted defeat and stopped fighting back, some of the people say. By the end of the year, three senior DMO officials, including the head of the division and the head of the product review team, had left the agency, making it easier than ever for Kalshi to list its contracts unimpeded.

Today, Kalshi is growing. Its staff of 32 operates out of an office in New York’s SoHo with big windows and exposed brick. It’s also added some new names to its board, including Quintenz, who left the CFTC 10 months after Kalshi was approved. He says he joined because of “my interest in the hedging and risk management opportunities event contracts could provide the market.”

As of mid-May, there were 75 markets on the company’s website, such as “Will there be negative GDP growth by Q4?” and “Will NASA land a person on the moon before 2025?” The exchange recently reached 2 million contracts traded a week, a jump from where it started but still a comparatively tiny figure compared with other futures exchanges. Many of the early adopters are prediction market enthusiasts, lured away from PredictIt and Polymarket. Wagers on the site are currently capped at $25,000, but Kalshi has said it hopes to increase that to $100,000 and then beyond, depending on the market.

Lopes Lara and Mansour understand that, with the regulatory drawbridge down, they need to move quickly. CME Group Inc., the Chicago futures exchange, recently announced it was planning a foray into event contracts tied to financial indexes. Kalshi’s next step in attracting customers will be to release a smartphone app. After that it hopes to announce a partnership with one of the large brokerage firms. Sequoia is a major investor in Robinhood Markets Inc., the retail brokerage phenomenon that now has 22 million customers. Perhaps one day, Robinhood users could be given access to Kalshi so that, after picking up their phones to buy shares in GameStop Corp., they’d be prompted to place a bet on the Oscars or the next Federal Reserve commissioner.

There are those, such as Democratic Representative Sean Casten of Illinois, who accuse Robinhood and its competitors of deliberately gamifying trading to encourage addiction, but that kind of criticism isn’t something Kalshi seems overly worried about. Kalshi’s customers aren’t “at the moment” able to wager more than they’ve deposited, Mansour says, which makes it hard to rack up debt. He won’t, however, rule out introducing leveraged bets at some point down the road.

The tension over event contracts brings to mind another episode in the CFTC’s history. In 1994, Chairman Brooksley Born proposed introducing regulations to curb the nascent market for financial derivatives. She was resisted by Fed Chairman Alan Greenspan and others within the government, who claimed it would stifle innovation and push capital offshore. Unconstrained, derivatives grew into a trillion-dollar industry, until 2008, when they were blamed for playing a central part in the financial crisis.

Today, with a midterm election looming, it seems reasonable to ask whether Kalshi plans to get involved. A record 125 million shares were traded on PredictIt on the outcome of 2020, and elections have historically been the biggest draw in prediction markets. “We can’t discuss specifics,” Mansour says. “All I can say is, you know, we’re always working on expanding the universe of things that people can trade on.”

Any contracts on elections would require a green light from the CFTC, and with three of the five commissioners currently Democrats, that might be difficult. But the calculus would quickly change with a Republican in the White House.

More stories like this are available on bloomberg.com

©2022 Bloomberg L.P.