Are Election Prediction Markets Good for Democracy?

Election prediction markets are neither the infallible truth machine their promoters describe nor the existential threat their critics sometimes claim. They are a market design with real informational benefits and real civic costs. The honest answer to whether they are good for democracy is: it depends on which risk you weight more heavily.

The case for them is that prices aggregate dispersed information and give the public a continuously updated probability. The case against is that they can be manipulated, can create perverse incentives and can turn a shared civic decision into a speculative game. Both cases have evidence behind them.

The case for: prices as information

Prediction markets allow people to put money behind a belief. That creates an incentive to be accurate rather than confident. In 2024, Polymarket's presidential market signalled a Trump victory while many polls showed a toss-up. It also moved sharply toward Biden dropping out weeks before he did.

Those examples are real, but they are selected after the fact. A market can be right once and wrong the next time. Academic work has challenged the idea that election markets are reliably better than polls. Political scientist John Sides has argued that prediction markets can be thin, manipulable and overconfident, especially when a small number of traders dominate price discovery.

A 2026 SSRN paper found that a small minority of informed traders accounts for most price discovery, which is consistent with both the optimistic and pessimistic views. Markets can be informative because a few people know something, and fragile because most people do not.

The case against: manipulation and money

The most concrete criticism is that large traders can move a market. In October 2024, the Wall Street Journal reported that four accounts with about $30 million in Trump wagers had moved Polymarket's odds. Polymarket later said the accounts were controlled by one French trader with extensive experience and found no evidence of manipulation. The trader made about $85 million when Trump won.

That episode does not prove manipulation. It does prove that a single wealthy participant can dominate a market's price. If the public interprets that price as a civic signal, the result can be misleading even without fraud.

A second criticism is insider trading. Election markets can attract people with non-public information about campaigns, polling, opposition research or government actions. The insider trading article documents cases far beyond elections, including military operations.

A third criticism is addiction and consumer harm. Election markets are not the only gambling product, but they are marketed as analysis rather than betting. That framing can understate the risk to inexperienced users. The beginner risk guide covers the base rates.

The case for: hedging and speech

Supporters argue that election markets let people hedge political risk. A business exposed to a candidate's tax policy can buy a contract that pays if the candidate wins. That is a legitimate financial use, even if most users are speculating.

They also argue that betting is a form of speech and that banning it pushes activity to worse venues. The global bans map shows that prohibition does not eliminate demand; it redirects it.

France's regulator took the opposite view, classifying prediction markets as unlicensed gambling rather than financial markets. The French case is a reminder that the "is it gambling or derivatives?" question is not settled, and the answer changes the legal treatment. The France ban article has the details.

The design questions that matter

The debate is often framed as a binary: legal or banned. A better framing is: which design reduces the harms while preserving the information?

Three design choices matter most.

• Position limits. A cap on how much one trader can wager reduces the ability of a whale to move a market. Polymarket has historically been permissive.
• Identity and surveillance. Kalshi argues that banning anonymous accounts makes manipulation and insider trading easier to police. Polymarket's international venue has historically allowed more anonymity.
• Market selection. Markets on election outcomes are different from markets on a candidate's health, a military strike or a death. The war markets controversy shows why the line is hard to draw.

The empirical record is mixed

The strongest evidence for prediction markets comes from liquid markets with many informed participants. The weakest comes from thin markets where a few traders dominate. Most election markets fall somewhere in between.

That is why the academic debate has not produced a clean verdict. A market can be more accurate than a poll in one election and less accurate in another. The design, the liquidity and the participants matter more than the label.

What the evidence does not show

There is no credible study showing that election prediction markets change election outcomes. There is evidence that they can be manipulated in thin markets and that they can produce misleading signals when a few traders dominate. There is also evidence that they can be useful forecasting tools in liquid markets.

The honest conclusion is that election markets are a tool. Like any tool, their value depends on how they are used and regulated. A well-designed market with position limits, surveillance and clear rules can add information. A thin, anonymous market with no limits can amplify noise.

For traders, the practical lesson is to treat a price as a signal with error bars, not a prophecy. The probability guide explains how to read the number correctly.

For policymakers, the useful question is not whether election betting is distasteful. It is whether the market design produces more accurate information than the alternatives, and whether the costs to democratic trust are worth it. That is a judgment, not a calculation.