Election betting odds: the complete guide to election prediction markets (2026 midterms to 2028)
Where you can legally bet on elections in the US, how to read election odds, how accurate the markets were in 2024 and in New York’s 2025 mayoral race, and what Kalshi and Polymarket say about the 2026 midterms and the 2028 presidential election. With charts from the markets’ own price history.
In the US, election betting happens on prediction markets, not at sportsbooks. Since a 2024 court ruling, Americans can legally trade contracts on who wins an election on federally regulated exchanges such as Kalshi and Polymarket US; state-licensed sportsbooks do not take election bets, and Nevada's rules forbid them outright. Outside the US, the global Polymarket exchange and licensed betting exchanges fill the same role.
Election betting odds on these markets are simply prices: a contract that pays $1 if a candidate wins and trades at 67¢ means traders put that candidate's chance at about 67%. This guide covers where you can bet, how the odds work, how accurate election markets have been (with charts of the 2024 presidential race and New York's 2025 mayoral race), and what the markets say about the 3 November 2026 midterms and the 2028 presidential election.
Odds are a snapshot from 22 October 2026. They move every day; live prices are linked throughout.
Election betting odds today
| Market | Polymarket | Kalshi |
|---|---|---|
| Democrats win the Senate (2026) | 67.5% | 67% |
| Democrats win the House (2026) | 92.5% | 91.5% |
| Democrats sweep both chambers | 67.5% | — |
| Democratic Party wins the 2028 presidency | — | 59% |
| JD Vance wins the 2028 presidency | 20% | 20% |
Live: Senate on Polymarket · Senate on Kalshi · House on Polymarket · House on Kalshi · balance of power · 2028 winner on Polymarket · 2028 winner on Kalshi · 2028 party on Kalshi. Every 2026 race, side by side: our election hub.
Where can you legally bet on elections in the US?
| Platform | What it is | Election markets | Notes |
|---|---|---|---|
| Kalshi | CFTC-regulated exchange | Congress, governors, mayors, primaries, 2028 | Election contracts since October 2024; blocked in Washington state by a court order |
| Polymarket US | CFTC-regulated exchange | Major US races | Polymarket's regulated US app, in dollars, since December 2025 |
| PredictIt | Political market under CFTC relief | US politics, from primaries to cabinet picks | $3,500 limit per contract; 10% of profit plus 5% to withdraw |
| Interactive Brokers | Broker with its own exchange, ForecastEx | Selected political contracts | Commission-free, in a brokerage account |
| Robinhood, Coinbase | Brokers routing to exchanges | A selection of Kalshi's political markets | Convenience; a small commission on Robinhood |
| Polymarket (global) | Crypto exchange outside the US | Elections worldwide | Closed to US residents |
What you cannot do is walk into a Las Vegas sportsbook and bet on the presidency. Nevada's regulations bar licensed sports pools from taking bets on elections, and when the state's Gaming Control Board ordered Kalshi to stop offering sports and election contracts in March 2025, it pointed to exactly that rule. The "Vegas election odds" you see quoted in the media usually come from offshore bookmakers, British betting exchanges or, increasingly, Kalshi and Polymarket themselves. How the main platforms compare overall: the best prediction markets.
How election betting became legal
For most of the last century, Americans had no legal way to bet on elections at scale. The path to today ran through one lawsuit:
- 2014: PredictIt opens under a CFTC no-action letter with an $850 limit per contract (PredictIt review).
- September 2023: the CFTC refuses to let Kalshi list contracts on which party will control Congress, arguing they are against the public interest. Kalshi sues.
- September 2024: Judge Jia Cobb of the federal court in Washington, DC rules for Kalshi, finding that the law did not give the CFTC the review power it used. Weeks before the presidential election, Americans can trade election contracts on a regulated exchange for the first time.
- May 2025: the CFTC drops its appeal, and the appeals court dismisses the case. Election contracts are here to stay.
- 2025: PredictIt's settlement raises its limit to $3,500; Polymarket wins CFTC approval to return to the US.
- October 2026: the CFTC proposes writing event contracts on politics, sports, culture and weather into the federal definition of a swap, which would strengthen its exclusive jurisdiction (the CFTC proposal).
Legal does not mean popular. A Reuters/Ipsos poll taken from 30 September to 5 October 2026 found that about 69% of Democrats and 61% of Republicans do not support using prediction markets to bet on election results. The longer history, from papal elections to the Curb Exchange: a short history of prediction markets.
How to read election odds
Every election contract pays $1 if its outcome happens and nothing if it does not, so the price is the market's probability. Bookmakers write the same thing as American or decimal odds:
| Price | Implied chance | American odds | Decimal odds | $100 buys | Profit if it wins |
|---|---|---|---|---|---|
| 90¢ | 90% | −900 | 1.11 | 111 contracts | $11 |
| 67¢ | 67% | −203 | 1.49 | 149 contracts | $49 |
| 50¢ | 50% | +100 | 2.00 | 200 contracts | $100 |
| 20¢ | 20% | +400 | 5.00 | 500 contracts | $400 |
| 9¢ | 9% | +1011 | 11.11 | 1,111 contracts | $1,011 |
Profits are before fees, which on Kalshi are 7% × price × (1 − price) per contract for takers and on Polymarket's politics markets 0.04 × price × (1 − price) (fees compared). Two practical points:
- Buy and sell, do not just bet. You can sell a contract before the election at whatever the market pays, locking in a profit or cutting a loss (how to hedge).
- Read the rules, not just the title. "Who will win the Senate?" resolves on which party holds the majority when the new Congress sits, not on election-night calls. A race can be decided weeks later in a recount or runoff (how markets resolve).
Converting between formats in more depth: how to read prediction market odds.
Types of election markets
Election markets go well beyond "who wins". The main kinds:
- Winner markets: one contract per candidate, such as the 2028 presidential winner or a Senate race. Prices across candidates should add up to about 100%; when they add up to more, the extra is the market's margin.
- Party and control markets: which party wins a seat, the Senate, the House or the White House. These are often the deepest, because they pool many races into one question.
- Combination markets: Polymarket's balance-of-power market prices all four ways Congress can split, which tells you how traders think the chambers are linked.
- Margin and popular-vote markets: by how much a candidate wins, or who wins the national popular vote. Thinner, and more sensitive to how the rules define the result.
- Primaries and nominations: who wins a party's nomination or a primary. Often the least accurate, because polls are sparse and turnout is low (Kalshi's Wisconsin miss).
- Event markets around a campaign: endorsements, debate moments, rallies, what a candidate says. Small, easy to move, and the ones regulators watch for insider trading.
- International elections: from Brazil to Israel, mostly on the global Polymarket site.
As a rule of thumb, the broader the question, the more money it attracts and the more its price deserves your trust. A party-control market with tens of millions traded aggregates far more information than a primary with a few thousand dollars on the book (slippage, volume and spread).
Election odds vs polls vs forecast models
Polls, models and markets answer different questions, and the useful skill is reading them together:
| Polls | Forecast models | Prediction markets | |
|---|---|---|---|
| What they measure | What a sample of voters says today | A probability built from polls, history and fundamentals | What traders with money at stake expect on the day |
| How fast they update | Days, with each new poll | Daily, as polls come in | Minutes, on any news |
| Strengths | Direct evidence from voters | Transparent, consistent method | Speed, and they price turnout, polling error and events polls cannot see |
| Weaknesses | Sampling error, non-response, house effects | Only as good as their inputs and assumptions | Thin markets, whales, long-dated discounts, rules risk |
Markets are not independent of polls: traders read the polls and the models, then add everything else they know. When a market and a model disagree sharply, as with the 2026 Senate, where markets put Democrats near 67% and Decision Desk HQ about 56%, the gap itself is information. It usually means traders are weighing something (a candidate's momentum, a correlated polling error, early voting) differently from the model (how accurate are prediction markets).
How big election betting has become
Elections made prediction markets famous. Pew Research Center found that in October and November 2024, political markets made up 90% of Kalshi's trading volume and 65% of Polymarket's. Polymarket's presidential winner market alone traded about $3.7 billion.
Source: Pew Research Center analysis of data from The Block (notional taker volume; Polymarket includes Polymarket US); politics includes Polymarket’s geopolitics tag
Show the dataHide the data
| Polymarket (incl. Polymarket US) | Kalshi | |
|---|---|---|
| Jul ’24 | $0.73B | $0.00B |
| Aug ’24 | $0.74B | $0.00B |
| Sep ’24 | $0.63B | $0.00B |
| Oct ’24 | $2.96B | $0.26B |
| Nov ’24 | $2.06B | $1.13B |
| Dec ’24 | $0.93B | $0.09B |
| Jan ’25 | $0.54B | $0.08B |
| Feb ’25 | $0.29B | $0.04B |
| Mar ’25 | $0.32B | $0.03B |
| Apr ’25 | $0.55B | $0.04B |
| May ’25 | $0.63B | $0.02B |
| Jun ’25 | $0.66B | $0.04B |
| Jul ’25 | $0.22B | $0.03B |
| Aug ’25 | $0.29B | $0.03B |
| Sep ’25 | $0.53B | $0.04B |
| Oct ’25 | $1.28B | $0.13B |
| Nov ’25 | $0.86B | $0.23B |
| Dec ’25 | $1.21B | $0.08B |
| Jan ’26 | $2.10B | $0.26B |
| Feb ’26 | $1.96B | $0.30B |
| Mar ’26 | $2.55B | $0.19B |
| Apr ’26 | $2.49B | $0.20B |
| May ’26 | $1.44B | $0.21B |
| Jun ’26 | $1.26B | $0.26B |
| Jul ’26 | $0.73B | $0.17B |
Politics never fell back to pre-2024 levels on Polymarket: Pew puts its political volume at about $2 billion a month on average in 2026, while Kalshi's has stayed under $300 million, because Kalshi's growth has come from sports.
The 2026 midterms are already the biggest midterm cycle on record:
Source: CNN; Reuters and the Anti-Corruption Data Collective
The Anti-Corruption Data Collective projects $1.4–1.6 billion of trading on congressional races by the end of the cycle. The money is concentrated, though: it found that the top 1% of Polymarket wallets account for about 70% of congressional betting volume (the full story).
How accurate are election betting odds?
A market priced at 70% should be wrong three times in ten. Accuracy means calibration (70% outcomes happen about 70% of the time) and resolution (favourites win more often than not), and both are measured over many questions, not one (Brier score explained). Here is the long record, then two recent races up close.
The long record.
- 1868–1940: New York's organised election betting picked the winner a month out in every presidential election but one, 1916, according to economists Paul Rhode and Koleman Strumpf.
- 1988–2004: the University of Iowa's small real-money markets beat 964 opinion polls 74% of the time across five presidential elections.
- 2016: markets gave Donald Trump roughly a 20–25% chance. Unlikely outcomes happen: a forecaster whose 25% calls never come in is badly calibrated too.
- 2024 onwards: analyses of Polymarket's resolved markets found the favourite right about 90% of the time a month before resolution and about 94% a few hours before (how accurate are prediction markets).
The 2024 presidential election
Source: Polymarket price history (daily closes, UTC), “Presidential Election Winner 2024”
Show the dataHide the data
| Donald Trump | Democratic candidate | |
|---|---|---|
| Jan 8 | 47.5% | 33.5% |
| Jan 15 | 47.5% | 35.5% |
| Jan 22 | 48.5% | 36.5% |
| Jan 29 | 54.5% | 38.5% |
| Feb 5 | 52.5% | 35.5% |
| Feb 12 | 52.5% | 32.5% |
| Feb 19 | 53.5% | 32.5% |
| Feb 26 | 53.5% | 32.5% |
| Mar 4 | 53.5% | 32.5% |
| Mar 11 | 51.5% | 34.5% |
| Mar 18 | 51.5% | 37.5% |
| Mar 25 | 49.5% | 40.5% |
| Apr 1 | 48.5% | 41.5% |
| Apr 8 | 46.5% | 43.5% |
| Apr 15 | 45.5% | 43.5% |
| Apr 22 | 43.5% | 43.5% |
| Apr 29 | 43.5% | 44.5% |
| May 6 | 46.5% | 43.5% |
| May 13 | 44.5% | 44.5% |
| May 20 | 51.5% | 41.5% |
| May 27 | 55.5% | 37.5% |
| Jun 3 | 53.5% | 37.5% |
| Jun 10 | 55.5% | 33.5% |
| Jun 17 | 56.5% | 33.5% |
| Jun 24 | 57.5% | 34.5% |
| Jul 1 | 61.5% | 18.5% |
| Jul 8 | 62.5% | 14.5% |
| Jul 15 | 69.5% | 17.5% |
| Jul 22 | 63.0% | 29.1% |
| Jul 29 | 58.5% | 39.3% |
| Aug 5 | 51.5% | 44.6% |
| Aug 12 | 45.6% | 51.5% |
| Aug 19 | 48.4% | 49.4% |
| Aug 26 | 49.5% | 49.6% |
| Sep 2 | 50.8% | 47.8% |
| Sep 9 | 51.3% | 46.4% |
| Sep 16 | 49.4% | 49.6% |
| Sep 23 | 46.9% | 52.0% |
| Sep 30 | 49.0% | 50.2% |
| Oct 7 | 50.7% | 48.4% |
| Oct 14 | 53.9% | 45.5% |
| Oct 21 | 60.7% | 39.2% |
| Oct 28 | 65.0% | 34.8% |
| Nov 4 | 56.4% | 43.5% |
The chart tells the year in three acts. Trump led most of the spring; after the June debate and the July assassination attempt his chance passed 69%. When Biden withdrew on 21 July, Harris rose within three weeks to a narrow lead that held, on and off, until late September. In October the market swung back to Trump, peaking at 65% on 28 October, and on election eve it gave him about 56–59% while polling averages showed a near tie in the swing states. He won.
That does not prove the market "knew". A 58% favourite loses four times in ten. What it shows is that the market moved faster than polls on news and was willing to take a side when polls could not.
New York City's 2025 mayoral race
Source: Polymarket price history (daily closes, UTC), “New York City Mayoral Election”
Show the dataHide the data
| Zohran Mamdani | Andrew Cuomo | |
|---|---|---|
| Apr 28 | 8.0% | 77.5% |
| May 5 | 8.0% | 77.5% |
| May 12 | 7.0% | 83.0% |
| May 19 | 6.0% | 82.5% |
| May 26 | 3.9% | 86.5% |
| Jun 2 | 9.8% | 83.0% |
| Jun 9 | 18.4% | 73.5% |
| Jun 16 | 28.9% | 60.0% |
| Jun 23 | 19.6% | 66.0% |
| Jun 30 | 68.3% | 10.2% |
| Jul 7 | 69.3% | 4.5% |
| Jul 14 | 73.2% | 13.9% |
| Jul 21 | 71.1% | 16.2% |
| Jul 28 | 73.2% | 13.0% |
| Aug 4 | 81.8% | 8.1% |
| Aug 11 | 80.3% | 8.6% |
| Aug 18 | 82.7% | 8.9% |
| Aug 25 | 85.0% | 9.0% |
| Sep 1 | 85.3% | 10.4% |
| Sep 8 | 81.1% | 17.0% |
| Sep 15 | 81.5% | 17.3% |
| Sep 22 | 87.2% | 10.8% |
| Sep 29 | 84.8% | 14.5% |
| Oct 6 | 86.2% | 13.1% |
| Oct 13 | 87.6% | 11.5% |
| Oct 20 | 92.3% | 6.5% |
| Oct 27 | 95.3% | 3.6% |
| Nov 3 | 93.8% | 6.2% |
This one is a lesson in humility. In late May 2025 the market gave Zohran Mamdani under 4% and Andrew Cuomo over 85%. Mamdani climbed through June as polls tightened, then won the Democratic primary on 24 June, and the price flipped within days. From there the market made him a steady favourite and, a week before the general election, priced him at about 95%. He won.
The pattern is typical: markets can be badly wrong months out, especially when the only information is early polls, and become sharp as real results arrive.
When election markets miss
- Thin markets. In August 2026, Kalshi's market on Wisconsin's Democratic primary for governor gave the eventual winner about 4%. Kalshi then added low-volume warnings and accuracy scores to its midterm markets (Kalshi's transparency features).
- Correlated errors. If polls are off in one direction, they are usually off in several states at once, which is why markets price Senate control below what the individual races suggest (Senate and House odds).
- Whales. A single trader can hold a price away from fair value for a while, as one Intrade trader did with Mitt Romney's price in 2012 (whale tracking).
- Long horizons. Money tied up for years earns little, so long-dated markets drift towards 50% and longshots stay overpriced.
The 2026 midterms: what the markets say
Republicans hold the Senate 53–47, so Democrats need a net gain of four seats for a majority. The markets have moved steadily in their direction for a year.
Source: Polymarket price history (daily closes, UTC)
Show the dataHide the data
| House | Senate | |
|---|---|---|
| 1 Aug ’25 | 70.5% | 28.5% |
| 15 Aug ’25 | 70.5% | 29.0% |
| 1 Sep ’25 | 69.5% | 29.0% |
| 15 Sep ’25 | 68.5% | 28.5% |
| 1 Oct ’25 | 67.5% | 27.5% |
| 15 Oct ’25 | 61.5% | 26.5% |
| 1 Nov ’25 | 61.5% | 28.0% |
| 15 Nov ’25 | 69.5% | 33.5% |
| 1 Dec ’25 | 76.5% | 34.5% |
| 15 Dec ’25 | 75.5% | 34.5% |
| 1 Jan ’26 | 79.5% | 34.5% |
| 15 Jan ’26 | 77.5% | 34.5% |
| 1 Feb ’26 | 80.5% | 35.5% |
| 15 Feb ’26 | 82.5% | 37.5% |
| 1 Mar ’26 | 83.5% | 40.5% |
| 15 Mar ’26 | 83.5% | 50.5% |
| 1 Apr ’26 | 84.5% | 49.5% |
| 15 Apr ’26 | 84.5% | 53.5% |
| 1 May ’26 | 83.5% | 51.5% |
| 15 May ’26 | 78.5% | 46.5% |
| 1 Jun ’26 | 81.5% | 46.5% |
| 15 Jun ’26 | 82.5% | 44.5% |
| 1 Jul ’26 | 82.5% | 41.5% |
| 15 Jul ’26 | 83.5% | 45.5% |
| 1 Aug ’26 | 85.5% | 43.5% |
| 15 Aug ’26 | 87.5% | 52.5% |
| 1 Sep ’26 | 88.5% | 50.5% |
| 15 Sep ’26 | 87.5% | 53.5% |
| 1 Oct ’26 | 92.5% | 62.5% |
| 15 Oct ’26 | 90.5% | 63.5% |
| 22 Oct ’26 | 92.5% | 67.5% |
The House has looked like a Democratic pickup since late 2025; the Senate went from a long shot (under 30% in autumn 2025) to a coin flip in spring 2026 and to about two in three in October, much of the latest move coming from Maine. Forecast models are less bullish on the Senate: Decision Desk HQ put Democrats at about 56% in October.
The races that decide it, with live odds: Texas · Maine · Ohio · Iowa · all Senate races · House · governors. On election night itself: how to follow live election odds.
Governors and mayors have drawn surprising money too: by late September Kalshi's Los Angeles mayor market had traded about $97 million and its California governor market more than $50 million (LA mayor odds · California governor odds · Ohio governor odds).
The 2028 presidential election odds
Polymarket's 2028 winner market had already traded more than $700 million by October 2026, two years before the vote. Kalshi's party market makes Democrats about 59% favourites to win the White House.
Source: Polymarket and Kalshi, 22 October 2026
Show the dataHide the data
| Value | |
|---|---|
| JD Vance | 20.1% |
| Alexandria Ocasio-Cortez | 13.0% |
| Gavin Newsom | 9.5% |
| Jon Ossoff | 9.0% |
| Marco Rubio | 8.6% |
| Kamala Harris | 4.2% |
Read long-dated markets with care. Nobody on this list has declared, the field will change after the midterms, and money locked up for two years makes traders reluctant to bet against longshots, so the prices add up to less information than they seem. Notice too how the two exchanges disagree on Ocasio-Cortez and Newsom by 2.5 to 3.5 points: on a market this far out, that kind of gap is normal (why prices differ between exchanges).
How to bet on an election, step by step
- Choose a platform. In the US, Kalshi or Polymarket US for the widest choice, PredictIt for small stakes. Outside the US, Polymarket (Kalshi vs Polymarket).
- Open and verify an account. Expect to give your name, address, date of birth and, in the US, usually your Social Security number. You must be 18 on Kalshi (age rules).
- Deposit. Bank transfer or card on US platforms; USDC on global Polymarket (deposits and withdrawals).
- Read the market's rules. Check the resolution source and the deadline before you buy.
- Place a limit order. Name your price instead of taking whatever is offered; it is cheaper and protects you on fast nights (limit vs market orders).
- Decide your exit. Hold to settlement, or sell if the price reaches your target.
Election betting strategy: what actually works
- Trade the race you know. Local knowledge (how a county counts its mail ballots, a candidate's ground game) is the most common real edge. National headline markets are the hardest to beat.
- Compare platforms. The same race often trades a few points apart on Kalshi and Polymarket; buying the cheaper side is free value (spread scanner).
- Think in correlations. If you like Democrats in Maine, you probably like them in Texas and Ohio too. Five correlated bets are closer to one big bet than to five small ones.
- Size positions sensibly. Even a real edge loses often. The Kelly criterion gives a ceiling for how much to stake (Kelly criterion · expected value).
- Know election-night dynamics. Early counts can mislead: some states count mail ballots first, others last. Spreads widen when results come in fast (live election odds).
- Mind taxes and time. Close races can take weeks to settle, and profits are taxable (Kalshi taxes).
More approaches, from market making to news trading: how to make money on prediction markets.
Insider trading and manipulation
Election markets raise a question sports and economics markets rarely do: what if the people who decide the outcome are also trading it? Kalshi says every account is identity-verified to screen out insiders, and in 2026 it fined and suspended four congressional candidates for trading on their own races. CNN reported in October 2026 that very specific midterm markets, on endorsements, rallies and debate moments, had drawn more than $20 million, exactly where inside knowledge pays. Congress has opened an inquiry into insider trading on prediction markets. If a tiny market moves sharply before news breaks, treat it as a warning, not a forecast.
Election betting outside the US
- Polymarket's global exchange lists elections in dozens of countries, from Brazil's runoff to Israel's vote, but is closed or close-only in the US, UK, France, Germany, Australia and others (where Polymarket is legal). Recent examples: Brazil · Israel.
- Betting exchanges such as Betfair Exchange and Smarkets are licensed for politics in the UK, and bookmakers there offer fixed odds on elections (Polymarket alternatives).
- Play-money and forecasting sites such as Manifold and Metaculus let anyone forecast elections without money (Manifold · Metaculus).
Election betting odds: frequently asked questions
Is it legal to bet on elections in the US?
Yes, on CFTC-regulated exchanges such as Kalshi and Polymarket US, and on PredictIt. State-licensed sportsbooks do not take election bets.
Where can I see live election betting odds?
On our 2026 election hub, which shows Kalshi and Polymarket side by side for every race, or in either platform's app.
Who is favoured to win the 2026 midterms?
As of 22 October 2026, Democrats: about 67% to win the Senate and 91–92.5% to win the House.
Who is favoured to win the 2028 presidential election?
JD Vance leads at about 20% on both Polymarket and Kalshi; Kalshi makes the Democratic Party about a 59% favourite.
Can you bet on elections in Las Vegas?
No. Nevada's regulations prohibit sportsbooks from taking election bets.
How accurate are election betting odds?
Well calibrated on average and better than single polls far from election day, but thin markets and long-dated markets can be far off, and favourites still lose.
What happens if a race is too close to call?
The market stays open until its resolution source decides, which can take days or weeks; your money stays in the position until then.
