With just two weeks until election day, races around the country are heating up. With Kalshi crossing the $70 million mark in elections trading volume, our markets are only getting more liquid and our information more accurate. All information is as of Oct 22, 5 PM EST. 

Note: The images you see are all from our private internal metrics, but feel free to republish them.

Who is winning the election in the presidential market? Volume on this market alone is ~$50 million, and Trump keeps breaking his own records for the lead, surging to 61% last night before stabilizing around 60%. Admittedly, this does seem quite high for an election many in the press have been calling a toss-up. So, what’s going on? Is there something wrong with how the market is operating?

Contrary to claims in the media about prediction market manipulation (although I can’t speak for Polymarket), betting activity on Kalshi seems to indicate it’s operating exactly how it should. 

Trump is high in the markets simply because he has more support. The number of people betting on Trump is ~1.5x the amount of Harris bettors, and the overwhelming majority of states have more directional bettors who believe in a Trump victory than Harris. 

Directional Bettors by State

Okay, sure, so he has more support…that doesn’t necessarily mean the market is acting properly. What if it’s because the pro-Trump demographic is more prone to signing up for political betting sites? Or simply because pro-Trump bettors have a more deeply ingrained and confident belief in a Trump victory than the Harris bettors have in a Harris victory? 

Fortunately, unlike polls or pundit analysis, markets have a self-corrective mechanism to eliminate these types of potential biases: profit incentive. If there’s some sort of bias in the market, that means there’s money to be made, and sharp traders (usually institutions and large bettors) will come in and take balancing positions until the market settles. 

Lo and behold–this is exactly what we see happening. While Trump has more bettors, large bettors overwhelmingly favor Harris, driving up median bet size. On average, Harris bettors bet ~$800 more than Trump bettors, and the largest net position on Harris exceeds the largest net position on Trump by ~$300,000. I don’t have access to who these users are for privacy reasons, but the indication is that these are the sharps and market makers balancing out the grassroots Trump support. 

Median Bet Size by Candidate

Trading Positions by Candidate

All of this suggests that the market is acting exactly as it should. Broad Trump interest drives up prices, which is bid down to equilibrium level by sharp traders who come in with larger bets. Prediction markets, especially liquid and regulated ones, simply work. 

Elsewhere, Trump has (to put it lightly) surged in key swing states. In the past few days, Trump has broken out past 60% in Nevada, is flirting with touching 70% in both Georgia and North Carolina, and took a dominant lead in Pennsylvania at a shocking 57%.  

The one swing state that Harris seems to be holding her own in is Michigan. Here’s the interesting part: this is being driven by local traders. Michigan is the only state in which more bettors have placed pro-Harris bets than pro-Trump bets, which seems to be buoying her market chances there. Theoretically, this information is more accurate, as traders have on-the-ground information that out-of-staters wouldn’t. (Think of this graph as on a scale of dark red to gold, it means a higher percentage of Trump bettors as opposed to Harris bettors).

Local Trading Volume Ratio by State (>30 unique traders)

Finally, Trump has opened as a massive favorite in ME-2, which, in one of the at least semi-realistic scenarios in which the electoral college ends in a 269-269 tie, could decide the election. This mostly rural country in a secluded corner of the country could very well be the tiebreaker in deciding the fate of the free world–and right now, the market is predicting a landslide Trump victory. 

This market is especially notable because shockingly little is known about the district. 538 has only aggregated 5 polls since Harris became the nominee and they have wild swings between them, with some saying Trump is up by 7 points and others claiming Harris is up 5. Although Trump carried the district by 7 points in 2016 and 2020, the Harris campaign is acting like they’re competitive, and the Cook Political Report has it listed as a toss-up. 

With only ~$16k in volume, this market isn’t all that trustworthy either, adding to the general confusion about this potentially important district–but it’s at least worth taking into account. This is one of Kalshi’s newest markets, so we’ll keep you updated as trading increases over the next two weeks. 

Final note: none of this is financial advice.