To get an idea of how an election will shake out, a person might look at polling data. Increasingly, they might also look at a candidate’s odds on a prediction market, where users place bets against each other on the outcomes of real-world events.

Prediction markets have soared in popularity following a landmark 2024 court victory for Kalshi Inc., a leading platform. In short order, they have become an influential tool for predicting elections. A growing number of media outlets feature real-time market odds in their coverage, and several have entered deals with Kalshi or Polymarket, another major platform.

Prediction markets have gained a reputation in some tech-friendly circles for being more accurate than traditional election polls. This perception was burnished by markets on both major platforms predicting that Donald Trump would win the 2024 presidential election. Mainstream polls slightly overstated support for his opponent, Kamala Harris.

Though the picture is unclear, some early academic research suggests this perception is accurate for major elections. At the same time, critics say turning a civic process into the object of financial speculation poses new risks for democracy. Here’s what’s important to know.

Accounts of election betting in the US date to more than a century ago. In the early 1900s, bettors wagered on elections outside of the New York Stock Exchange, creating odds that were reported in newspapers. The practice declined as state laws and court rulings made election betting widely illegal, and scientific polling became the favored predictive tool for elections.

For decades, the Commodity Futures Trading Commission — the federal agency that regulates derivatives markets — only tolerated election-related markets, such as the Iowa Electronic Markets, that operated as not-for-profit research tools with caps on traders and individual bets.

Roughly 20 states have picked legal fights with prediction market platforms, arguing that they are sportsbooks disguised as derivatives markets and should be regulated by states, which oversee gambling. Federal courts have split on the question of whether prediction markets should fall under the purview of states, teeing up the conflict to potentially reach the Supreme Court.

The legal battles have primarily focused on sports contracts, but more than half of US states explicitly restrict election betting, according to the National Conference of State Legislatures. Those prohibitions could pose another legal hurdle to prediction market platforms.

More people use electoral prediction markets to understand how candidates are performing than to try to turn a profit. Kalshi estimates that on its midterms markets, non-trading visitors outnumber traders five to one. This year, the platform launched a midterms hub catering to that crowd, which it “designed as the one-stop-shop for state and federal election forecasting.”Prediction markets have become popular among some big dollar donors — especially a subset of Silicon Valley funders — who are consulting them to guide their support to campaigns where they believe their dollars will go the furthest, said Cooper Teboe, a fundraising strategist who works in the region.

In the run-up to the 2024 presidential election, news outlets began citing odds from prediction markets the same way they might have reported on the latest polling data. A candidate’s “Polymarket odds” were held up as another data point in the horse race-like media coverage of the election.

Some of those relationships have since been made official. In the past year, Polymarket has entered deals with media organizations including Google Finance, Substack and Dow Jones. Kalshi has minted its own partnerships with CBS News, CNN, CNBC and Fox News.

These deals are bringing prediction market odds to a wider audience. As part of CNN’s deal with Kalshi, the network has introduced a segment called “The Odds,” where presenter Harry Enten explores how Kalshi markets are pricing the outcomes of various midterm races and other real-world events. CNBC regularly publishes articles on what Kalshi traders expect from various events, including midterm elections. The business news channel can receive a one-time fee when people create Kalshi accounts through its website.

“Journalism is better when it’s backed by live markets,” Polymarket’s X account posted when it announced its partnership with Substack in February 2026.

“The media has always relied on third-party data and we see ourselves as one data source among many — a tool to complement reporting,” a Kalshi spokesperson said in an email. Kalshi’s partners have full control over which markets they report on, the spokesperson said.

For nearly a century, opinion polls have been the primary instrument for gauging public sentiment ahead of US elections. The practice took a reputational hit in 2016, when polls underestimated Trump in several crucial battleground states. In the years since, some polling skeptics have embraced prediction markets.

But polls and prediction markets are not one-to-one — they are fundamentally different tools. Election polls are not explicitly predictive, even though they are widely interpreted as forecasts. Rather, they offer a snapshot in time, approximating how the electorate feels about an upcoming election at a given moment. It’s a different aim than prediction markets, which invite people to wager on what they think will happen in the future and profit if they’re correct.

Each has its shortcomings. Prediction markets, especially illiquid ones, are vulnerable to manipulation by those seeking to influence perceptions of an election; strategically placed bets can temporarily move the market’s odds, creating the appearance of a shift in a candidate’s prospects. (More on this later.) Polls conducted by reputable, nonpartisan organizations don’t have that problem, but they can be vulnerable to methodological flaws such as biased sampling, leading questions or flawed weighting. In some cases, online polls without safeguards can be skewed by synthetic respondents generated by artificial intelligence.

For some, prediction markets have a couple of clear advantages over polls. First, they react to events in real time. By the time polls are released, they’re slightly dated, because it takes time to weight responses and publish results. Polymarket and Kalshi odds react instantaneously to major news developments, which published polls can’t do.

Second, when people answer polling questions, they are stating who they intend to vote for, not who they think will win, and their desire to be perceived a certain way might influence their response. When people bet on prediction markets, they are incentivized to predict correctly, not express a preference.

Some research suggests that prediction markets fare better than polls at predicting the outcomes of high-profile elections. A Vanderbilt University analysis found that markets on Polymarket were more successful than polling at predicting the outcome of the 2024 presidential election, particularly in swing states.

Other academics argue that the comparison is flawed, because prediction market odds are informed by the latest polls. A 2012 analysis from Columbia and Temple Universities concluded that historically, prediction markets have been heavily dependent on what credible polls show and follow their lead, even when polls get it wrong.

The 2026 US midterm elections have renewed debate about the reliability of prediction markets to forecast election results — though polls have faced similar scrutiny.

The night before Wisconsin’s gubernatorial Democratic primary, a Kalshi market gave progressive candidate Francesca Hong a 95% chance of winning. The afternoon before the results came in, Polymarket posted on X that Hong was a “near-lock” to win the race (the company subsequently deleted the post). In the end, Hong lost to establishment candidate David Crowley by less than one percentage point. Crowley’s late surge, which came after numerous candidates dropped out in the final months of the race, wasn’t anticipated by prediction markets or polls.

Critics of prediction markets pointed to the Wisconsin miss as evidence of their shortcomings, but their biggest champions argued that the result was consistent with an accurately calibrated market. Calibration measures whether events assigned a certain probability happen about that often over time.

“A 5% probability doesn’t mean it won’t happen,” Kalshi Chief Executive Officer Tarek Mansour posted on X following the results. “If 5% candidates never won, the markets would be broken.”Prediction markets are named for what their biggest cheerleaders consider their core value: predicting the outcome of future events. People are more truthful, the argument goes, when they put their money where their mouth is. But on election markets, bets can be placed in an attempt to influence the results, not just predict them.

A surge in Polymarket and Kalshi odds, even a temporary one, can generate media coverage and awareness for a campaign. Name recognition is a valuable currency in politics and can translate into donations and votes. A candidate’s surge on prediction markets might influence voter behavior: Research suggests that some voters cast their ballots for whichever candidate they expect to win.

Election market surges can be bought. Markets on major races, such as a US presidential election, tend to be highly liquid, meaning they attract a lot of dollars in trading volume and are therefore difficult to manipulate. But the majority of election forecasts on Kalshi and Polymarket are far less liquid. An election market for a smaller race might attract just thousands of dollars in betting. If a user strategically places bids, they can significantly boost the market’s estimate that a particular candidate will win.

In theory, prediction markets are designed to correct for such an influx. If a candidate’s odds have suddenly risen due to market manipulation, the payoff for correctly betting against them rises, giving traders an incentive to take the other side of the trade until the market corrects. But in the modern news environment, even a temporary spike on Kalshi or Polymarket before the market corrects can inflate a candidate’s visibility.

San Jose Mayor Matt Mahan didn’t have a lot of name recognition in January 2026, when he entered the race to be California’s next governor. The morning that Mahan announced his candidacy, a user named Fine-Tributary spent $44,000 placing Polymarket wagers on him. The market for the gubernatorial race was small enough that the purchases tilted it significantly in favor of Mahan, indicating the following day that he had a 36% chance to be California’s next governor. Traditional polls had him around 4%.

The surge was large enough to grab headlines in local and national publications. “Gamblers have placed their bets on Matt Mahan,” stated a New York Post story that ran the next day. The article did not mention that the Polymarket move was driven almost entirely by one anonymous account.

Mahan’s odds on Polymarket fell back in line with the polls months later, and the market correctly predicted Xavier Becerra’s primary victory nearly a month before it took place. But Mahan’s temporary surge and subsequent media coverage illustrates how illiquid election markets can be meaningfully manipulated in the short-term.

Bellwether, an online tool built by a Stanford Graduate School of Business research lab, monitors how much money it would take to meaningfully move political markets on prediction market platforms, including Kalshi and Polymarket. A single bet of less than $100,000 could move 99% of the 15,000 US midterm election markets that Bellwether tracked in mid-October 2026 by at least 5 percentage points, according to a Bloomberg analysis of Bellwether data.

Highly liquid markets aren’t immune to manipulation: A sufficiently well-funded bad actor could use large, strategically-placed bids to temporarily distort market prices, security experts warn.

Asked whether illiquid election markets are vulnerable to manipulation, a Kalshi spokesperson said they are not because of the incentive for markets to self-correct. “Mispriced markets are equivalent to free money lying on the ground,” the spokesperson said, and pointed to a Kalshi study of the market for the 2026 Los Angeles mayoral primary. The study found that a large purchase of “yes” contracts on candidate Spencer Pratt briefly drove the contract’s price from 29 cents to 45 cents, but trading returned it to 29 cents just nine seconds later.

People looking at election markets should be responsible for considering their liquidity in evaluating their reliability, Robert DeNault, head of enforcement and legal counsel at Kalshi, told Bloomberg in August 2026. “A user should also be intaking, ‘I’m looking at these odds, but I’m also seeing there’s not that many people trading in this market, so they might not be reflective of the wisdom of the crowd at this time,’” DeNault said.

Less than a month before the 2026 US midterm general elections, Kalshi began tagging midterm election markets with less than $10,000 in trading volume as “low volume.”The measure was introduced alongside other transparency features to “make election market prices easier to understand,” according to a company statement.

According to Bellwether, mainstream media outlets regularly fail to distinguish between liquid and illiquid markets. Of the prediction markets cited by news outlets over 30 days between September and October 2026, nearly 60% lacked the liquidity needed for reliable reporting, Bellwether found.

Election markets can be fertile ground for insider trading. Campaign staffers might have access to nonpublic information that they can profit on, such as an unreleased opinion poll or knowledge that a candidate is dropping out of a race.

A number of cases underscore the potential for markets that offer bets on elections to facilitate trading on sensitive information. In the spring of 2026, Kalshi fined three political candidates who had placed bets on their own elections. One of them was Mark Moran, a former contestant on the reality show “FBoy Island” running to represent Virginia in the US Senate. Moran wagered on himself in the “Who will run for public office this year?” market before announcing his candidacy, then traded on whether he would become the Democratic nominee, according to a notice of disciplinary action by Kalshi.

Kalshi banned Moran from the platform for five years, the filing says. The company explicitly forbids candidates and campaign staff from trading on their own elections, and bans people from trading in markets where they are “decision makers” or can influence the outcome. Polymarket prohibits users from trading on confidential information and wagering on events where they can influence the outcome.

Both Kalshi and Polymarket have guardrails in place to monitor and prevent suspicious activity, representatives said.