On a crisp Tuesday in September, primary election day in New York City, volunteers are handing out mayoral campaign flyers on the fringes of Zuccotti Park, and Laurence Lau sizes up the odds. “I don’t let my ideology get in the way,” he says. “You can’t — that’s how you lose money.” On a concrete table, a spreadsheet listing candidates and prices is open on a laptop and Lau explains how his computer model once turned raw data into positions on Intrade, a now-defunct site that billed itself as “the world’s leading prediction market.”
“I like betting because it’s the ultimate form of truth,” Lau says. “If someone says something to me and won’t put money on it — forget about it.”
Our meeting place is famous due to its association with Occupy Wall Street, but Lau picked it because it was one his regular lunch spots when he worked inside the nearby World Financial Center. By day, as a 26-year-old mutual fund analyst, he pored over billion-dollar investments, and by night, he traded tens of thousands from his own savings on Intrade, making bets with counterparties around the world. The market’s mechanics mirrored the proposition of a handshake wager: Traders bet directly against one another, winner takes all, with contracts always paying off in increments of $10. If a seller was willing to risk $6, the buyer would put up $4, and people would say the contract was “trading at 60,” giving the price the appearance of a percentage probability.
Intrade’s system was designed to evoke sophisticated investment, not a casino or a horse track; it rewarded analytical calculation over partisanship and wishful thinking. In theory, the traders’ collective wisdom, reflected in the market’s prices, would augur the ultimate electoral outcome. In 2008, for instance, it had come within one electoral vote of nailing Obama’s margin of victory. Its record had won it a following among journalists, Washington insiders, and a small cadre of economists who believed in a radical-sounding proposition: that free markets could be used to reliably forecast world events.
Though Intrade was tiny, and based in Ireland for regulatory reasons, its backers included several billionaires and a passel of American hedge fund managers. The site reflected their gospel of quantification — the numerical ethic that has spread, in recent years, from the finance world to nearly every competitive facet of American life. Intrade says its users wagered a total of $230 million on the 2012 presidential campaign, which made it nothing more than a flyspeck in comparison to actual financial markets. But it attracted an avid community of traders — finance guys, card sharks, political junkies — most of them young men in their twenties, the Xbox and Red Bull demographic. They scoffed at the rubes, like the tea partiers who bet real money on Herman Cain. They competed for inside information, cozying up to political reporters and campaign workers. A few traders, the evidence suggests, employed the sort of bucket-shop tricks that might have gotten them handcuffed by the Securities and Exchange Commission if they had tried them in their day jobs. When they made a killing, the wolves of Intrade howled about it on the site’s message boards, and on Twitter.
On primary nights, Lau would sometimes call county clerks to pry out the latest voting numbers, before they reached news organizations. “That’s what I think a lot of Wall Street is,” Lau says. “Have an edge and don’t tell anyone about it.”
In his wagering, Lau gave similar weight to the teachings of Paul Tudor Jones — a Wall Street legend who was an early Intrade investor — and Billy Beane, the statistically minded baseball executive who was the hero of the book and movie Moneyball. His computer system owed something to Black-Scholes, a financial model for derivatives, and something to Nate Silver, the poll-crunching star of the 2012 election, then based at the New York Times. Though Lau’s winnings were modest — “five figures,” he says, money he was setting aside to pay for an MBA program — he thrived on the competition, and the stories, perhaps apocryphal, of traders who had hit big jackpots. “Some other guy,” he tells me, “bought a million-dollar condo in the Financial District with his winnings.”
The night of the 2012 election, Lau camped out in his office adjacent to his firm’s trading floor, taking advantage of the ultrafast internet connection to get an extra nanosecond’s advantage over the guys using their living room Wi-Fi. Meanwhile, down in Virginia, Joe Schilling — a communications professional who drove a car with the vanity plate “BETSMRT” — was sitting in front of three computer screens, monitoring the markets for individual states, which you could also bet on Intrade. “We were buying, hand over fist, Obama-Virginia contracts in the 40s,” he said, “and rode it all the way up to 100.”
Along with the reelected president, new modes of quantitative analysis were vindicated by the 2012 campaign season, which sometimes seemed as much a contest between dueling statistical models as ideologies and candidates. Journalists like Silver and the Washington Post’s Ezra Klein, and the Obama campaign’s own sophisticated numbers operation, were hailed as avatars of a new force of rationality in media and society. The nerds won.
But Intrade wasn’t given much of a chance to capitalize on the moment. Within months of the election, the company had collapsed beneath the weight of a U.S. government lawsuit and a crippling financial scandal. Some of the factors in Intrade’s demise — regulatory ambiguity, erratic management, a flawed business model — were ills common to many internet startups. Other travails, including the mysterious death of its longtime CEO on Mount Everest, were like plot twists from an adventure novel. But the root source of the market’s problems, defenders say, was a gap, perhaps unbridgeable, between a promising model for predicting the future, and the legality of Intrade’s betting parlor.
As New Yorkers cast their ballots for mayor, Lau could do nothing but rue the wasted profit opportunities. “What is gambling? Can someone define gambling for me?” he asks. “Is it putting money on an uncertain outcome? Then that sounds like Wall Street.”
In a federal lawsuit filed in Washington, D.C., the U.S. government accuses Intrade of flouting regulation and violating securities laws. The regulators contend that beneath its innovative façade, the site was really just another grubby offshore bookmaker. Koleman Strumpf, an economics professor at the University of Kansas, says the powers that be could never accept the promise of prediction markets. “It was socially unseemly,” says Strumpf, who has shown that the practice of betting on elections has a long history, and an equally a persistent habit of attracting the wrath of the authorities.
Roman bookies ran numbers on the election of Renaissance popes until Gregory XIV banned the practice on penalty of excommunication. Around the turn of the 20th century, Wall Street brokers openly traded election futures and newspapers quoted their prices like modern opinion polls. Strumpf estimates that at the peak of this practice, in the election of 1916, around $10 million was bet on these markets — more than $200 million in today’s dollars. By the end of the New Deal era, though, the electoral markets had all but disappeared, due to both competition — modern polling pushed the betting lines out of the newspapers — and legal crackdowns.
American law has traditionally drawn a distinction between hedging risk for an economic purpose and playing games of chance. Putting $100 on the Broncos is illegal, but speculating on the future price of a barrel of oil, or the timing of your own death (via life insurance), is perfectly acceptable. “Lots of ordinary financial things are gambling,” says Robin Hanson, an economist at George Mason University. “But they are carved out in the public mind.”
Hanson argues that laws against gambling suppress a useful social purpose. Economic studies have shown that point spreads, which shift with the actions of numerous bettors, are a reliable predictor of sports contests. More than two decades ago, Hanson began to wonder: Why couldn’t the same mechanism work for other types of clashes, like contentious scientific debates? “I was a contrarian hanging out with a lot of other contrarians in Silicon Valley,” Hanson said, describing people who were both “rabidly libertarian” and interested in exploring the internet’s capacity to solve problems. “I sort of pulled those two things together,” he explained, “and said how about betting markets?”
Hanson called his concept “idea futures,” and devised a game where people could wager play money on questions like the chances of developing cold fusion. In a 1995 article for Wired, he wrote that he hoped the prototype would inspire real investors, with real money, to create “a radical, market-based alternative for reaching scientific consensus.”
The idea ended up resonating, initially, with a different audience: the U.S. intelligence community. After Sept. 11, the Defense Advanced Research Projects Agency, or DARPA, the military’s research arm, commissioned a group involving Hanson to develop a market to forecast events in the Middle East. The resulting firestorm was, perhaps, predictable. “The idea of a federal betting parlor on atrocities and terrorism is ridiculous and it’s grotesque,” said Sen. Ron Wyden, a Democrat from Oregon. The head of DARPA, Adm. John Poindexter, was forced to resign and the project was squelched.
In a recent interview with the New Yorker, Wyden suggested that he seized on the gambling experiments as pretext to kill a more sinister surveillance program Poindexter was pursuing. But Hanson says his research was a collateral victim. “Once you have a moral principle like that,” he says, “factual analysis doesn’t really matter.”
The logic of the marketplace has many proponents. In his popular 2004 book The Wisdom of Crowds, financial journalist James Surowiecki extolled the accuracy of collective predictions, citing the success of experiments like a rudimentary election market run out of the University of Iowa. But lawmakers have shown extraordinarily little sympathy. “We idolize Einstein and Isaac Newton,” says Emile Servan-Schreiber, a prediction market pioneer who now consults on corporate applications. “The idea that the expert human mind could be outclassed by a bunch of amateurs getting together is very disturbing, as disturbing as saying the Earth is not the center of the solar system.”
The irony, of course, is that even as academics struggled to establish their fledgling markets, Wall Street was conducting its own experiment, with much more dangerous stakes. In the late 1990s, Congress passed a series of laws relaxing regulations on financial derivatives. Some of them had practical applications — a utility could buy weather futures to hedge its financial exposure to a heat wave — and some proved disastrous. Infamously, bankers invented derivatives to place trillion-dollar bets on home mortgages and the risk of credit defaults, which led to the near-implosion of the world economy when the real estate bubble collapsed in 2008. But long before any of that, in 1999, some guys from the pits of New York’s commodities exchange decided to start a website where you could speculate on current events.
“Anything that you could answer yes or no to, or measure objectively with a number, could be traded,” says Ron Bernstein, one of Intrade’s founders. “That was our philosophy.” Why couldn’t the same pricing system that allowed him to trade sugar futures, for instance, be repurposed to predict the outcome of a football game or a civil war? The idea proved bewitching to Wall Street professionals, who were used to seeing the world in terms of probabilities, whether they were betting on interest rates or the Jets.
“You have a couple thousand testosterone-laden young men who are all mad about sports,” says Patrick Young, author of the book Capital Market Revolution and an early investor in Intrade. “So it’s easy to draw the corollary.”
Intrade secured an early $6.5 million investment from Paul Tudor Jones, which bequeathed instant credibility. One day, Young, Bernstein, and others met at Jones’ estate in Greenwich, Conn., to pitch Rupert Murdoch’s son, Lachlan, who flew up in a helicopter. Murdoch bought in, as did many others. By 2001, Intrade was flush with $13.5 million in venture capital. The site’s initial focus was on sports, but it also offered markets on elections, box office returns, and geopolitical events. “You talk about the wisdom of crowds,” Young says. “This was the ultimate mass-market operation.”
Intrade began, essentially, as a wager in itself — on the possibility that the U.S. government would take a relaxed attitude toward gambling on the internet. But it launched into the maw of the dot-com crash, and instead of loosening up on gambling sites, U.S. law enforcement tightened the screws. Though Intrade operated out of Ireland, a country with a permissive attitude toward bookmaking, Bernstein and his partners had always seen their home country as the site’s primary market. (Betfair, a publicly traded and very respectable British company, dominates Europe, but excludes American customers.) But U.S. prosecutors began pressing cases against other offshore betting sites, arresting some owners when they were on American soil, and effectively forcing others to live in exile. Bernstein and Intrade’s other founders distanced themselves from management, ceding control to their top Irish executive, an accountant named John Delaney. The new CEO was left to run the ailing shop with minimal interference.
The startup was limping toward a quiet death until it stumbled into an unforeseen bit of luck: the war in Iraq. In the excruciating buildup to the 2003 invasion, the world media discovered a curious website was giving odds on the chances of war. Later that year, for reasons never explained, Intrade’s market for Saddam Hussein’s capture began to move two days before he was actually seized. Economist Justin Wolfers, now at the University of Michigan, was trying to devise a method to predict the reaction to war within financial markets. He examined the site’s trading patterns and was excited by what he found. “That was the start of what became a symbiotic relationship,” says Wolfers, a longhaired Australian with a loquacious social media presence. He became one of Intrade’s most vocal boosters within academia, and used its data to conduct important research on prediction market theory.
During the 2004 election campaign, Intrade vastly outperformed the infamously faulty polling, showing a consistent lead for George Bush and picking every state correctly on the day before the election. For prediction market theorists, the result was precious proof of concept. Their scholarly enthusiasm pointed the way to a potential new business model. The company phased out sports betting, and Delaney declared that “hedging political risk is the next frontier in asset management.” Press coverage of the site’s success brought it to the attention of Dean LeBaron, an eminent Wall Street investor credited with inventing the index fund, who sought out Delaney after a talk at Dartmouth University. “As someone who has been in the investment business for 50 years,” he says, “it never occurred to me that there was any difference between what Intrade was doing and what I was doing, buying and selling securities.”
Eventually, LeBaron became Intrade’s largest single shareholder. “What Intrade was doing,” he says now, “was it really gambling, or was it extracting information?”
The U.S. government offered a hostile answer to LeBaron’s question. In 2005, the federal Commodities Futures Trade Commission brought an action against Intrade for soliciting predictions on financial questions like the future price of gold. (The company paid a fine.) The following year, Congress passed a law further restricting offshore gambling sites. Delaney had become a regular at American academic conferences, often concluding conversations with the enthusiastic economists with the promise of a future Guinness, but after the feds started locking up gaming website proprietors, he abruptly stopped visiting. In early 2007, Delaney appeared via Skype at a Washington conference co-sponsored by the American Enterprise Institute and the Brookings Institution, apologizing for his absence and joking that, as an Irishman, he preferred to be sitting in bars, not behind bars.
“Even though there’s just this tiny, tiny little risk,” Intrade’s chief executive explained, in terms all the economists understood, “because that risk could have big personal costs for me, until things are a little bit clearer, I’m going to stay in Dublin.”
How good were Intrade’s predictions? Though its accuracy varied with its trading volume — the more bets, the more collective wisdom — the site’s predictions appear to have consistently beaten individual pollsters. Sometimes, the market was proved wrong: Intrade gave an 80% chance that the Supreme Court would overturn Obama’s health care law. But as Justin Wolfers grew tired of pointing out to reporters, an 80%-certain prediction should be wrong one out of every five times.
Probably the most exhaustive study of Intrade’s performance was conducted by David Rothschild, an economist with Microsoft Research. He looked at 74 statewide races during the 2008 campaign season, and found that Intrade yielded probabilities that were more accurate than polls, and was particularly good at picking the winner early and in close races. Rothschild found that Intrade was roughly as accurate as the vaunted algorithmic model that Nate Silver used to make predictions on his personal website, FiveThirtyEight, which he later took to the New York Times. But the two systems arrived at their accurate forecasts by different routes.
Silver’s formula was complex, but it fundamentally relied on opinion polls, which typically ask voters which candidate they prefer. Prediction markets were designed to elicit a potentially more revealing opinion: Who is going to win? Although economists disagree about the effect of monetary stakes, they agree that a competitive marketplace — whether over cash, prestige, or the pleasure of saying, “I told you so” — will typically yield more accurate assessments than passive statements of individual preference.
At the most elemental level, Intrade was a useful tool to test one of the most important questions in economics, how markets absorb information. Prediction market proponents argued that, in addition to providing a gauge of the betting public’s sentiment, Intrade also served a second function: what Wolfers calls “information discovery.” Intrade would respond to gossip of clandestine negotiations, or an extramarital affair, or an “oops” moment in a debate. That was why many political journalists watched the site’s twitches. A former Intrade employee told me that, from IP addresses, it was clear that many election season visitors were U.S. government employees. The day before Mitt Romney made his own vice presidential pick, Paul Ryan’s price mysteriously surged, leading traders to presume a leak.
Intrade’s management winked at suggestions of insider trading. “I want to get people who know too much into the market, rather than keep them out, so that their wisdom is reflected in prices,” LeBaron says. Intrade’s message boards were always full of accusations of manipulation, which some took as a perverse sort of endorsement. “That’s a sign of a mature market,” says a hedge fund manager who invested in Intrade as a business. “All big markets have someone thinking about how to manipulate them.”
In the heat of the political moment, though, it was often hard to sort the speculators with an edge from the ideologues. In 2004, there was a deep-pocketed trader who bet a bundle on John Kerry, sparking outrage among conservative bloggers, who inevitably accused George Soros. “During the Obama-McCain contest, a single millionaire tried to pull the market in one direction,” the former employee says. The company would usually contact traders in such situations. “He said, ‘I really want McCain, I really want to buy these for McCain.’ He went on for another eight days, at $150,000 a day, and it didn’t even show up as a blip.” The market had a natural way of punishing the foolish: It took their bets.
By contrast, the market rewarded those who acted quickly on solid information. Ken Fitzpatrick, a former professional poker player who lives in Las Vegas, is proudest of a score he made through intrepid reporting. In 2008, as John McCain prepared to announce his running mate, Fitzpatrick and his fellow trader Joe Schilling were monitoring the movements of all the contenders, calling their press secretaries and checking a flight-tracking website. They noticed a Gulfstream jet from Anchorage, Alaska, bound for the site of a McCain rally, and bought Sarah Palin, a longshot. When the surprise pick was unveiled, Fitzpatrick took home $25,000.
Laurence Lau credits his biggest single profit during the 2012 season to his swift reaction to a friend’s outraged post on Facebook. It related that Missouri Republican Senate candidate Todd Aikin had claimed that victims of “legitimate rape” seldom become pregnant. “I went against my algorithm,” Lau says, “and because this guy made these idiotic comments, I put money on Claire McCaskill,” the Democrat. By moving quickly, Lau was able to buy McCaskill futures cheap. Not long afterward, he was able to use the strategy again, after Richard Mourdock, a Republican candidate in Indiana, said something similarly offensive about rape. “I made a lot of money on those guys,” Lau says.
Intrade’s prices reflected the impact of these news events faster than any poll could; this is what the academics meant by “information discovery.” And how could anyone be opposed to knowing more information? In 2008, a group of 22 academics called for loosened regulations in an open letter to Science, describing a “virtually limitless” range of applications for government policy, business and public health. Four Nobel laureates were among the signatories, including 2013 economics co-winner Robert Shiller. Yet the reform effort went nowhere. Over time, federal authorities curtailed Intrade’s ability to use PayPal and other financial mechanisms, forcing U.S.-based traders to rely on a cumbersome bank wire system.
For a company that was geared toward American consumers, the restrictions were stifling. For all the press attention the company managed to attract, Intrade always scuffled financially, barely sustaining a small staff. During the 2012 campaign, it claimed to have around 7,000 active account holders who paid a monthly fee of $5 to trade. But the presidential campaigns were a bonanza that only occurred every four years, and in between the political markets went through long periods of dormancy. Even at the height of the presidential races, Intrade’s volume was puny by financial industry standards, and undercard races were often lightly traded, which contributed to the exchange’s price volatility.
“They were the only real-money prediction market around, but they always felt a little bit like a fly-by-night operation,” says Adam Siegel, chief executive of Inkling Markets, a collective intelligence consulting firm. Even some of Intrade’s investors were incredulous about the company’s grandiose claims of predictive powers. “People were attaching quite a lot of value to hundreds of dollars, as indications of what people think,” says one. “There was part of you that was like, ‘Really?’”
In retrospect, it’s easy to say the economists should have recognized that an offshore betting website with a proprietor who joked about ending up in an orange jumpsuit was flirting with disaster. But the prediction experts never saw it coming.
Intrade’s downfall began at the top of Mount Everest. In 2011, John Delaney, an avid climber in his early forties, died about a hundred feet short of short of the summit. He left a wife and three children, including a daughter born just three days before. His body remained on the mountain, making a funeral impossible, so the family held a memorial service, which some of Intrade’s investors attended. A question hung over the tragic affair: How could the master of probabilities have taken the ultimate risk?
“In hindsight, it’s easy to say he calculated wrong in attempting to climb Everest, but especially among prediction market proponents, we know that decisions cannot be evaluated in hindsight,” researcher David Pennock wrote in a blog post. “My guess is that John knew the risks and felt the climb was a gamble worth taking.”
Intrade’s shareholders brought in new management. “We knew something appalling was wrong very, very quickly,” says Patrick Young. Irish courts were pursuing Delaney over personal debts, including unpaid bank loans and cell phone bills, and he had incorporated a number of mysterious companies, including one named after Everest’s height, “29035 Approx Limited.” Meanwhile, Intrade’s books were in disarray. An internal audit found Delaney had redirected millions of dollars to accounts he personally controlled. Delaney had been battling an undisclosed illness that left him alarmingly thin. But he also spent more than $40,000 on the climbing expedition, according his guide Alex Abramov, who now questions whether Delaney concealed the extent of his health problems.
“Sometimes people come to climb Everest,” Abramov says. “Sometimes people come to die.” Dean LeBaron, among others, came to reevaluate Delaney’s gamble: Maybe climbing Everest looked like a win-win proposition if he was already at the end of his rope. A mountaineering friend of Delaney’s dismissed that interpretation, though, questioning Abramov’s preparations. “John Delaney tried to do it on the cheap, he tried to cut corners, and he didn’t have the experience to cut corners, and he died,” he said, adding that his friend was the same way with business. “John was a risk-taker.”
One risk Delaney took was continuing to defy U.S. regulation. Intrade’s investors had always wanted to move the company out of the gray market: At one point they came close to selling to MF Global, then a respected trading firm, and later they negotiated a merger with a commodities exchange in Minneapolis. But Delaney scuttled the deals, which would have exposed Intrade’s books to scrutiny. Instead, he had reintroduced the sort of financial contracts that the company had promised to give up in its prior settlement with the CFTC. To some, it looked like Intrade was thumbing his nose at the authorities.
“You are dealing with people who were essentially flouting U.S. law for a decade,” says Emile Servan-Schreiber. “So at some level, they had to be sort of gangster-like.”
After Delaney’s dubious machinations were discovered, Intrade went through a protracted behind-the-scenes drama, which culminated in co-founder Ron Bernstein’s return as chief executive. The new regime now portrays the dead CEO as a rogue operator who took advantage of a lack of oversight from the company’s absentee investors. But Delaney’s behavior notwithstanding, there was little question that the exchange itself operated, at best, on the furthest fringes of legality. The Dodd-Frank financial reform, signed by President Obama in 2010, specifically bans futures related to terrorism, assassination, gaming, or anything “contrary to the public interest,” and in 2012 the agency advised a Chicago firm that covers elections. Though a CFTC spokesman declined requests for comment about Intrade, Michael Gorham, a former agency official who now teaches at the Illinois Institute of Technology, told me he looked in vain for a way to legalize such exchanges when he headed the agency’s market oversight division. “I don’t see how you can get around the fact that these things are gambling,” he said.
The case against election betting isn’t merely moralistic. Since the 19th century, lawmakers have consistently worried about the corrupting effects of attaching explicit monetary stakes to political decisions. Strumpf says his historical research shows that “the record is rife with accusations that parties tried to boost their candidates” through betting market manipulation. He found little evidence that such ploys had any sustained effect on public opinion. But by making a market for political information, Intrade did create a potential temptation for insiders. After all, someone made a bundle on that Paul Ryan bet.
During the final days of the 2012 election season, Intrade’s message boards were abuzz about an anonymous trader who came to be known as the “Romney Whale.” His motives were mysterious, but his movements were unmistakable, making huge waves through the market. In the two weeks before the election, even as the polls moved in Obama’s favor, the Whale added $3.8 million to his already enormous bet on the Republican.
Sitting in his office at Barnard College, economist Rajiv Sethi was intrigued. According to the efficient markets hypothesis, the price of a Romney contract should have reflected the reality of the race, which seemed to be heading toward a comfortable Obama victory. Nate Silver’s operation for the New York Times, for instance, was putting the president’s chances at better than 90%. But Obama’s Intrade price was holding steady at 70. When Sethi looked closer at the order book that listed all the contracts up for sale, he saw the many thousands of bets offered by the Whale effectively propping up the market.
Sethi thought there had to be a hidden agenda behind the Whale’s seemingly irrational position. But what was it? The Whale never revealed his identity, disappearing from Intrade as soon as the last polls closed. The market immediately corrected, predicting the final outcome in all but one state: Florida. In total, the Whale lost nearly $7 million. “In the end the fog clears and reality asserts itself,” Sethi wrote on his blog. “Or so one hopes.”
The fog that obscured Intrade’s shaky finances began to dissipate immediately after the election. First, the CFTC brought its lawsuit to federal court in Washington, specifically citing Intrade’s decision to go back to offering (comparatively tiny) markets on financial predictions like the future unemployment rate. The evangelists of quantification responded with predictable outrage. “Out of all things the CFTC could be doing to protect consumers and investors,” Nate Silver tweeted, “it chooses to sue Intrade?!?”
Shortly afterward, though, in March 2013, Intrade announced that it was suspending operations, and the less innocent truth began to emerge. Ron Bernstein disclosed a cash shortfall affecting customer accounts. The company managed to avoid insolvency by obtaining a forbearance from customers while it attempted to recover the funds allegedly diverted by John Delaney. In a series of audits publicly filed in Ireland, Intrade disclosed that some $4.2 million was missing from its accounts and those of a related company, and the company sued Delaney’s widow, Orla, in the Irish courts.
As Intrade’s business imploded, Sethi and David Rothschild, his colleague at Microsoft Research, continued to investigate the market’s anomalous behavior during the closing days of the 2012 elections. The Romney Whale’s actions raised a troubling question: How did Intrade actually settle on its prices? Economists sometimes say they know how markets work in practice, but not in theory. According to the prevailing model, the microstructure of the market is broken up between smart traders, who wager based on good information, and lots of little “noise traders,” who decide based on transitory factors like momentum.
But that couldn’t explain the actions of the Whale, who seemed to be taking a manifestly dumb position. Was he a Romney true believer? Some exotic arbitrageur? Sethi and Rothschild eventually settled on the tentative hypothesis that the trader was a manipulator, trying to sustain Republican hopes — and turnout — at a time when all signs pointed to a demoralizing loss. They could only guess, however, as to whether the motives of the strategy were political or financial. “If there is an opportunity there,” Sethi tells me, “it raises the question of how seriously we should take Intrade’s prices.”
When Sethi and Rothschild got a look at all the transactions on Intrade over the final two weeks of the 2012 campaign, what they found didn’t look much like the economists’ clockwork model of the marketplace. Intrade’s contracts could be closed any time, allowing traders to take incremental profits — for instance, buying Romney after his “47%” comments leaked and selling after his victory in the first debate — as prices shifted.
It turned out that few traders followed this strategy. Only 6% of account holders showed a willingness to take either side of the wager. They weren’t willing to bet, however profitably, on a candidate they didn’t like. This was true even of the big fish: The market’s seven largest traders, including the Whale, all bet almost exclusively on one candidate. Inside, it turned out that Intrade didn’t look like a dispassionate machine. It appeared to function a lot like our messy, irrational, uncompromising political system.
Maybe that was the flaw in the theory — the uncontrolled variable of belief. But we’ll probably never know. Since the demise of Intrade, many prediction market proponents have come to the reluctant conclusion that their experiment is finished. However promising the idea of measuring political futures through real prices might be, it’s not likely to be put into large-scale practice again, given the CFTC’s intransigent interpretation of the law. “As an economist,” Koleman Strumpf says, “it’s hard for me to understand.”
Over the years, many prediction market pioneers have given up on the notion of trying to sell their ideas to the general public and skeptical legal authorities. Some, like Servan-Schreiber, have turned to applying the technology within big corporations, allowing employees to, say, bet on the success of products in development for play-money stakes. Research he and Wolfers conducted indicates that prediction markets function even without the promise of financial gain. Wolfers lost a bet on that, and had to buy his French colleague a bottle of Australian wine. “The problem with real-money markets,” Wolfers acknowledges now, “is that some people are dumb but rich.”
Many other academics, including Wolfers and Robin Hanson, are involved in projects that cater to another clientele: the intelligence community. Recently, Intelligence Advanced Research Projects Activity, IARPA, the Directorate of National Intelligence’s research wing, sponsored an academic competition meant to “dramatically enhance the accuracy, precision, and timeliness of forecasts.” Teams with names like the Good Judgement Project developed prototype markets for world events, eliciting predictions from a wide range of volunteers. (No money was at stake.) Hanson, who heads an IARPA-funded project called SciCast, said spies could derive an obvious benefit from trading on their collective knowledge through exchanges.
“The mechanisms we already have are plenty good enough to help them,” Hanson said. “It’s just hard for them socially to actually use them.”
Ron Bernstein, meanwhile, is still struggling to prove that Intrade itself has a future. While the company is still fighting the CFTC’s lawsuit, late last year it came to an undisclosed financial settlement with Orla Delaney. Bernstein now talks of “turning the corner from defense to offense.” He has moved the company’s headquarters back to New Jersey, where he lives, and is currently hiring staff to a relaunch a sports betting site, Tradesports.com, in a sense returning to the company’s original business plan. The Tradesports website touts a “new and LEGAL way to play fantasy sports,” structured around close-ended contests that award prize money. Bernstein says he plans to be up and running in time for college basketball’s March Madness.
As for the political markets that made Intrade famous, Bernstein still believes there’s a use for the trading platform technology, but he says the company is done with exploring gray areas. “Prediction markets in the United States for real money are not legal,” he says, emphatically. On a recent appearance Bernstein made on the Fox Business Channel, host John Stossel brought up Predictious, a new Irish-based marketplace that denominates trades in Bitcoin. “We’ve learned from our own experience,” Bernstein replied, “that regulatory avoidance isn’t a good business model.”
A few of the most voracious Intraders have moved on to playing Predictious or one of the handful of competing offshore exchanges. But most now consider themselves retired from politics. Laurence Lau has struck up friendships with some of his former competitors, such as Joe Schilling. “We really operated behind the veil of secrecy,” Schilling says, “but now I’m Twitter buddies with these guys.”
There has been a lot of gossip about Intrade’s corporate intrigue, and half-serious talk of a reunion, or a hedge fund. The traders have also watched with bemusement as Nate Silver has successfully marketed his algorithmic approach to all of life’s compelling pursuits: politics, sports, the weather, even dating and sex, moving his operation from the New York Times Company to Disney, the parent company of ABC and ESPN.
“The thing about Nate Silver is, there are a million other guys just like him,” Lau says. “As information becomes more readily available, it becomes harder to find that edge.” If Intrade was flawed, he and other traders argue, it wasn’t that there were too many dumb whales in the marketplace; it was because there were too many people betting exactly the way that Silver and his statistically minded acolytes told them to. When even the rubes are playing Moneyball, it gets hard to make a buck.
“God bless Nate Silver,” Schilling says, “but he kind of ruined Intrade for a lot of us.”
CORRECTION: An earlier version incorrectly stated which state Sen. Ron Wyden represents. (h/t to commenter Miguel Arkaitz Arbusto.)