Research Finds Thin Political Markets Can Shift on Less Than $1,000
Political prediction-market prices can move by double-digit percentage points on surprisingly small trades, according to new research that raises questions about how investors, journalists and political observers should interpret sudden changes in election odds.
An analysis by the Anti-Corruption Data Collective found that 94% of more than 11,000 markets tracking U.S. congressional races across Kalshi, Polymarket and Polymarket US could experience a price movement equivalent to 10 percentage points following a single wager of less than $1,000.
The finding does not mean that 94% of the markets have been manipulated. Instead, it measures how sensitive their available liquidity is to new orders and shows how little capital may be required to change the probability displayed to users.
That distinction matters as prediction markets become increasingly prominent as financial and political indicators rather than simply places to wager on events.
The study found the vulnerability was particularly pronounced in thin, low-probability markets. Among 1,094 contracts priced at five cents or less, 806 could reportedly be moved by five percentage points for less than $100.
A $3,500 trade could move 97% of the midterm-related markets studied by at least five percentage points, while most could move by 25 points following trades of around $25,000, according to the research.
The results add another dimension to previous research questioning how broadly prediction-market prices reflect collective intelligence. A separate academic study of Polymarket price discovery found that a small minority of traders played a disproportionate role in moving prices toward their eventual outcomes.
ACDC’s latest research also went beyond hypothetical order-book sensitivity.
The organization identified 353 occasions during the current election cycle where one or two Polymarket wallets moved a political market by at least five percentage points. Of those episodes, 211 markets remained around the new level, another 62 continued moving in the same direction and 80 returned to approximately their previous price.
That means most of the observed moves did not immediately reverse.
Reuters reported that in hundreds of Polymarket cases studied by ACDC, changed prices remained at their new levels for at least 24 hours and typically persisted for around four days.
The data challenges the idea that every unusual move in a thin market should automatically be interpreted as new information entering the market.
Kalshi and Polymarket dispute the broader implication that these conditions make their markets meaningfully vulnerable to persistent manipulation.
Kalshi pointed to a company case study involving a market tied to former Los Angeles mayoral candidate Spencer Pratt. The company said prices corrected in nine seconds even after a trader placed more than $1 million into the market.
Polymarket made a similar market-efficiency argument, saying an incorrect price creates a profit opportunity for traders willing to take the opposing position.
That mechanism is fundamental to prediction markets. If an artificial trade pushes an outcome from 40% to 55% without any change in the underlying information, other traders can theoretically buy the underpriced side until the probability returns toward fair value.
Recent cross-platform pricing gaps between Kalshi and Polymarket show that traders do respond to such differences, although those discrepancies have sometimes remained visible for hours rather than disappearing immediately.
Market Concentration Adds Another Layer
Liquidity is not the only concern highlighted by ACDC’s research.
An earlier analysis of 2026 congressional prediction markets found that the top 1% of Polymarket wallets generated 68% of trading volume, while just 10 wallets accounted for 17%.
About 80% of Polymarket’s congressional markets had fewer than 100 participating wallets, while only 10 markets had reached 1,000 participants as of the August analysis.
ACDC also found that 87% of the markets it examined were either low-volume, with less than $10,000 traded, or had relatively high volume concentrated among a small number of participants.
The figures are notable because headline industry volumes can create an impression of much deeper liquidity than exists in individual markets.
Midterm-election wagering had already reached $133 million by Aug. 10, exceeding the $92.4 million traded across the entire 2024 congressional election cycle. But activity was heavily concentrated in a handful of prominent races rather than distributed evenly across thousands of contracts.
This creates the possibility that a large market category can appear liquid in aggregate while an individual congressional seat, primary or candidate contract remains extremely shallow.
The issue is becoming more important as regulators focus on market manipulation and insider trading risks in prediction markets.
Recent enforcement cases have already demonstrated that political markets can attract traders with privileged information. The CFTC has pursued Kalshi trading cases involving people with access to non-public political information, while federal authorities have separately investigated alleged use of classified information on Polymarket.
A U.S. serviceman has also been charged in a case involving more than $400,000 in alleged profits from trades connected to the capture of Venezuelan President Nicolás Maduro, increasing scrutiny of insider trading on Polymarket.
A 10-Point Move Is Not the Same Thing as a 10-Point Change in Reality
This is where the research becomes particularly useful for investors and journalists.
A prediction-market price looks authoritative because it is numerical.
Candidate A is at 62%. Candidate B is at 38%. One number goes from 45% to 57% in an hour, and suddenly the natural assumption is that something happened.
Sometimes something did.
But in a thin market, the explanation might simply be that one person bought a few hundred dollars of contracts.
That makes volume almost as important as price.
If a highly liquid presidential market moves 10 percentage points while millions of dollars change hands across multiple platforms, that is potentially meaningful information. If a congressional primary with $4,000 of lifetime volume jumps by the same amount after one wallet spends $700, the informational value is completely different.
The displayed probability does not tell you that distinction.
This also explains why arbitrage is only a partial answer to the manipulation concern.
Kalshi and Polymarket are right that an obviously wrong price offers other traders an incentive to correct it. In deep markets, that mechanism can be extremely powerful.
But arbitrage only works quickly when someone notices the distortion, has sufficient capital available on the correct venue, trusts the contract’s resolution terms and believes the profit is large enough to justify entering the trade.
There may simply be nobody watching a lightly traded House primary in real time.
And by the time the price corrects, the distorted number may already have escaped the trading platform.
That is the part I think matters most.
Prediction-market probabilities increasingly appear in headlines, social-media posts, campaign discussions and investor commentary. Once a screenshot saying a candidate has suddenly gained 12 points begins circulating, the price itself becomes information — regardless of why it moved.
The possibility of deliberately buying that perception therefore creates a different incentive from normal trading. Someone might accept a financial loss if temporarily changing the displayed odds helps a campaign narrative, influences donors or generates favorable media coverage.
The ACDC research does not prove that this is occurring systematically. It does show that doing it could be remarkably cheap in some markets.
At the same time, suspicious-looking activity should not automatically be treated as misconduct. Recent scrutiny of unusual Polymarket wallet activity illustrates how difficult it can be to distinguish sophisticated trading, automated strategies, informational advantages and genuine abuse from the transaction data alone.
For anyone using prediction markets as an informational signal, the practical response is straightforward.
Do not look at the percentage first.
Look at what moved it.
A sudden political-market swing should be checked against trading volume, order-book depth, the number and concentration of participating wallets and prices on competing platforms. If Kalshi, Polymarket and conventional polling or news indicators all move together, the signal becomes considerably stronger.
If one thin contract jumps 10 points while everything else remains unchanged, treating that move as breaking political information may be giving a small trade far more authority than it deserves.
Shane Neagle is a financial markets analyst and digital assets journalist specializing in cryptocurrencies, memecoins, prediction markets, and blockchain-based financial systems. His work focuses on market structure, incentive design, liquidity dynamics, and how speculative behavior emerges across decentralized platforms.
He closely covers emerging crypto narratives, including memecoin ecosystems, on-chain activity, and the role of prediction markets in pricing political, economic, and technological outcomes. His analysis examines how capital flows, trader psychology, and platform design interact to create rapid market cycles across Web3 environments.
Alongside digital assets, Shane follows broader fintech and online trading developments, particularly where traditional financial infrastructure intersects with blockchain technology. His research-driven approach emphasizes understanding why markets behave the way they do, rather than short-term price movements, helping readers navigate fast-evolving crypto and speculative markets with clearer context.