The Election Odds Are Becoming Campaign Material

The danger is not that prediction markets are always wrong. The danger is that they look too right.

A prediction market is a trading venue where people buy contracts that pay out if a future event happens, such as a candidate winning an election or a government report landing above a certain number. If a contract costs 62 cents and pays $1 if the event occurs, readers often treat that as a 62 percent probability. Kalshi is a U.S.-regulated event-contract exchange, operating as a Commodity Futures Trading Commission, or CFTC, designated contract market, according to Kalshi’s market-integrity materials⁶. Polymarket is the crypto-native rival whose trades are recorded on-chain and whose prior U.S. legal history includes a 2022 CFTC settlement over offering event-based binary options without the required exchange designation or registration, according to the CFTC’s order⁵.

That legal distinction matters. It does not solve the political problem.

My view is straightforward: election prediction markets can be useful private forecasting tools, but they should not be treated as neutral public information infrastructure until platforms disclose, in real time, who is being paid to promote them, how deep the market actually is, how concentrated the trading is, and which insiders are barred from participating. Without that, political odds do not merely measure expectations. They can help manufacture them.

The Los Angeles mayoral race made the mechanism visible. After the June 2, 2026, primary in Los Angeles, influencers used changing odds on Kalshi and Polymarket to suggest something suspicious was happening as Spencer Pratt’s position worsened and Nithya Raman’s improved, according to KPBS/NPR reporting². Some posts carrying election-fraud claims also carried “paid partnership” labels tied to the companies, and NPR reported that the platforms moved to tighten or apply rules for paid creators after the controversy, according to the same KPBS/NPR account². Semafor separately reported that Kalshi asked some paid political influencers to remove posts that promoted Kalshi odds while casting doubt on the integrity of the Los Angeles election, according to Semafor³.

That is the feedback loop in miniature: (1) a market price moves, (2) a paid or affiliated account amplifies it, and (3) the price is reframed as independent proof that the official process is suspect. Liquidity, the amount of trading depth available before a price moves sharply, becomes a democracy issue because a shallow market can create a dramatic-looking chart. Market manipulation, meaning deceptive conduct meant to distort prices or create false signals, becomes a media issue because the distorted signal can travel faster than any enforcement action.

The strongest defense of these markets is real. Prediction markets have a forecasting pedigree. A study of Iowa Electronic Markets presidential vote-share markets from 1988 through 2004 found that market predictions were closer than 964 polls to the eventual two-party vote split 74 percent of the time, and that the markets significantly outperformed polls more than 100 days before Election Day, according to Berg, Nelson and Rietz’s study⁹. That evidence matters because it shows that money-backed forecasts can aggregate information better than some surveys under some conditions.

But the Iowa Electronic Markets were not the modern attention economy. They did not live inside a real-time influencer system where paid creators can turn a price tick into a fraud narrative before election officials finish counting mail ballots. The lesson from the older research is not that every election contract deserves public trust. It is that markets can be informative when the structure around them is disciplined enough for users to understand what the price means.

Modern political markets are not there yet.

Concentration is the quiet problem behind the shiny percentage. Bloomberg reported during the 2024 presidential campaign that 670 “power traders,” about 0.7 percent of analyzed Polymarket accounts, accounted for nearly half of purchase volume in the period it examined, according to Bloomberg⁴. Bloomberg also reported that a small cluster of large pro-Trump positions drew scrutiny because a few accounts appeared capable of shaping heavily cited odds, according to the same analysis⁴.

A concentrated market can still be smart. Sometimes one large trader is better informed than a crowd. But that is exactly why the public presentation needs context. “Candidate X has a 67 percent chance” sounds like collective intelligence. “Candidate X has a 67 percent market price in a contract where a small number of wallets drive a large share of volume” sounds like financial data. The second sentence is less viral. It is also more honest.

Insider trading risk is no longer theoretical. Kalshi said on June 9, 2026, that it would start collecting employment information for customers trading in certain high-risk markets, assign risk scores to new markets, and block presumptive insiders from trading where they may have material nonpublic information, according to the Associated Press¹. AP also reported that the move followed incidents including an investigation into former Rep. George Santos over alleged trading tied to his own State of the Union attendance and charges against a U.S. Army soldier accused of using classified information to make $400,000 on Polymarket trades related to U.S. military operations in Venezuela, according to AP¹. Kalshi says its new screening tools have stopped at least 100 potential insider trades and that it has made at least 20 referrals to law enforcement or regulators for manipulation or insider-trading concerns, according to AP¹.

Those reforms are good. They are also an admission that prediction markets now cover events where inside information is obvious, valuable and sometimes easy to possess. Campaign staff know campaign plans. Election administrators know procedural details. Government employees may know unreleased data. Platform employees may know market mechanics. If those people can trade, the price may become a leak wearing a probability costume.

The platforms know the problem. Kalshi says CFTC rules and its own exchange rules prohibit insider trading, fraud and manipulation, according to Kalshi⁷. Polymarket says it prohibits insider trading, fraud, market manipulation, self-dealing, front-running, wash trading and information misuse, and says trades are publicly viewable on-chain, according to Polymarket’s market-integrity page⁸. These rules are necessary. They are not enough for politics, because the political harm often happens before a case is investigated, a wallet is banned, or a post loses its paid label.

The counterargument I take seriously is that regulated markets are better than rumor. That is often true. A tradable price can discipline pundits, expose overconfident narratives, and create an auditable record. A CFTC-supervised exchange with know-your-customer rules and surveillance is more governable than a meme account with no paper trail. I do not want election forecasting pushed entirely into opaque offshore venues or private donor briefings.

But that argument sets the bar too low. Prediction markets seek authority because prices look impersonal. That appearance is the product. A poll comes with a sample size, margin of error and methodology, at least when reported responsibly. A forecast model usually has inputs and assumptions. A prediction-market price often arrives as a single clean number, stripped of bid-ask spread, open interest, order-book depth, wallet concentration, sponsorship relationships and insider-screening status. That stripped-down number is not transparency. It is branding.

So the fix is not to ban political markets tomorrow. The fix is to demote their public status until they earn it. Any election contract cited in news coverage should come with a standard label: volume, open interest, bid-ask spread, largest-holder concentration, share of volume from top traders, paid-promotion activity, affiliate-policy violations, trader eligibility rules and whether the venue is CFTC-supervised or operating outside ordinary U.S. exchange oversight. Paid creators should have to disclose the relationship clearly in the post, not through tiny labels or retroactive cleanup. Campaign staff, candidates, election officials and platform employees should be categorically barred from trading on markets where they can influence or know the outcome.

My prediction is that the next serious controversy will not come from a market simply getting an election wrong. It will come from a thin, fast-moving local or primary market whose odds are amplified as evidence that a count is illegitimate. The indicator to watch before the 2026 midterms is whether Kalshi and Polymarket publish live concentration and liquidity warnings on political contracts. If they do not, the odds will keep being sold as measurement while functioning, at the worst moments, as campaign material.