Polls show a deadlock, but betting markets strongly favour Eisenkot to replace Netanyahu

JTA – Traders on the Kalshi and Polymarket prediction markets are betting that former Israel Defense Forces chief of staff Gadi Eisenkot will replace incumbent Benjamin Netanyahu as Israel’s next prime minister following Tuesday 27 October’s election. 

That assessment comes despite polls released last week showing that neither candidate has a clear path to a governing coalition with a 61-seat majority. 

Most polls predict what has been dubbed the “change bloc” led by Eisenkot at 52 to 56 seats, while Netanyahu’s coalition is polling at around 50 seats. 

Kalshi traders, however, give Eisenkot a 52% chance of becoming Israel’s next prime minister, followed by Netanyahu at 28% and the former prime minister Naftali Bennett at 10%. On Polymarket, Eisenkot’s chances stand at 58% and Netanyahu’s at 27%, with no other candidate above 15%. 

Moran Ofir, a professor at the University of Haifa’s Faculty of Law who has researched prediction markets, explained the discrepancy, telling the Jewish Telegraphic Agency that there are two fundamental differences between polling and prediction markets. 

Polls measure which party voters intend to support, while prediction markets ask a different question: Who will become prime minister? 

Secondly, she said that the methodologies are fundamentally different. While polls rely on samples designed to be representative of the electorate, Polymarket reflects only those who choose to put their money on an election outcome. 

Also, Polymarket’s percentages represent implied probabilities of winning, rather than shares of voter support. 

At first glance, Ofir said, the fact that Polymarket participants have to put money behind their bets should protect the platform against deliberate misrepresentation, to which polls can be vulnerable. 

“The promise of these platforms, like Polymarket, Kalshi, or PredictIt, is to aggregate the wisdom of the crowd faithfully,” she said. 

But that vision is far from the reality on the ground. 

In a study conducted by Ofir, Joshua Mitts, a professor of law at Columbia Law School, and Dan Zimmermann, chief intelligence officer at Betting Intelligence (BETINT), the researchers analysed 1.2 million transactions. They found that 18% of all the money backing Netanyahu on Polymarket was concentrated in a single account, while just seven accounts held 29% of the money bet against him. 

“What we see in our research is that the Israeli election market on Polymarket is not a market driven by the wisdom of the crowd, but one dominated by a very small number of relatively large players putting up very large sums of money,” Ofir said. 

She described the election betting market as “fragile and highly susceptible to manipulation”, adding that participants may not be acting to maximise profits, but rather to influence public opinion ahead of the election. 

“Perhaps Polymarket was, for years, a platform that reflected the wisdom of the crowd. But powerful actors with vested interests identified it, and today, in the world of elections, it has become a platform where large and powerful players step in to shape public perception,” Ofir said. 

Another example identified in the study took place last summer. Two accounts placed bets on Ofer Winter, the leader of the Amcha Yisrael party, which was hovering around 3.25%, Israel’s electoral threshold to qualify for seats in the Knesset. The researchers identified two newly created betting accounts that wagered $3 200 (R53 190) on Winter becoming prime minister. The bets briefly pushed him above Avigdor Lieberman, an established Israeli political figure who heads the Yisrael Beiteinu party, in the market’s ranking of candidates’ chances of becoming prime minister. 

“Two days after the party’s launch: Ofer Winter overtakes Lieberman,” a major Israeli newspaper declared just hours later. Ofir suspects the bets were intended to create a misleading impression that would help persuade Winter’s supporters that his party could clear the electoral threshold. 

Ofir said the researchers interpreted betting patterns by the account that placed large wagers on Netanyahu as evidence of possible access to non-public information. Other accounts that wagered substantial sums showed similar patterns related to the war with Iran, she said. 

Brazil and Serbia held elections over the weekend. On Polymarket, Serbian President Aleksandar Vučić had a 72% chance of winning. In Brazil, incumbent President Luiz Inácio Lula da Silva had just a 17% chance of victory. 

Vučić lost. Lula won.