Cryptocurrency Prediction Markets in the 2024 Presidential Election

By Mitchel Zilbershteyn

Introduction

The 2024 presidential election was considered a tipping point for the crypto industry within the United States. Many crypto natives have been critical of the Biden administration and its mishandling of the industry, particularly the unethical enforcement practices by the SEC, accusing the administration of operating through enforcement rather than through regulation. This is evident through the numerous lawsuits brought against crypto businesses - from the beginning of 2023 through June of 2024, the SEC has dished out over 70 lawsuits to companies engaging with the cryptocurrency industry (Cornerstone Research, 2024). As a result, many crypto natives favored Trump to win the election, citing his openness towards the crypto industry and his more loose approach to market regulation.

Polymarket, a predictions market platform founded in 2020, was in a prime position to capitalize from this sentiment. Prior to June of 2024, this application was seeing under $50 million in monthly volume. Fast forward to October of 2024, Polymarket saw $2.5 billion dollars in trading volume (The Block, 2024).

How did this platform jump from being an unknown betting market for gamblers to place bets on future events, to a legitimate research tool that was continuously referenced as a serious predictor of the outcome of the election?

DISCLAIMER: This is purely an opinionated piece meant to analyze how financial markets can create a new level of insight into an event such as the 2024 Presidential Election by providing an unbiased view of the market’s reaction to the event. It is not written to be pro or against any of the candidates, and all of the reflected opinions are my own.

How does Polymarket work?

Polymarket is a market maker that allows users to bet on the outcome of an event. Users are able to buy shares of what they believe the outcome of an event will be, where the shares are based on the percentage chance of that outcome occurring. For example, if a user wants to bet on a basketball game between the Philadelphia 76ers and the Golden State Warriors, they can place a bet on any of the following outcomes: the 76ers winning, the 76ers losing, the Warriors winning, and the Warriors losing.

The price of each individual share will always be between $0 and $1 and is directly determined by the percentage likelihood of that outcome occurring. For example, if the 76ers have a 45% chance of winning then the price of the shares would be the following: 76ers winning is $0.45, 76ers losing is $0.55, Warriors winning is $0.55, Warriors losing is $0.45 per share.

Once the event has concluded and an outcome is determined, the winning shares would go up to $1 while the losing shares would go down to $0. In the basketball example, if the 76ers won, then the shares would have the following values: 76ers winning is $1, 76ers losing is $0, Warriors winning is $0, and Warriors losing is $1 per share.

A user that wants to bet on an event would by an x amount of shares of their perceived outcome. Seeing as this is an open market, the user can buy or sell shares at any given time and doesn’t have to wait until the event has concluded to take profits or cut losses.

For a more detailed explanation, check out this YouTube video by Polymarket as well as the full Polymarket guide.

Polymarket Volume Verification

The 2024 election outcome was by far the most bet on event on Polymarket. It saw a total betting volume of over $3.68 billion on the outcome of the election (Polymarket, 2024). The betting volume includes all of the deposits and withdrawals made to and from this event on the available candidates. The table below shows the full spread of this volume across all of the candidates that users placed bets on.

Source: Polymarket Presidential Election Winner 2024 (Volumes as of 11/8/2024)

All of these volumes can be concretely verified to be accurate because, as a cryptocurrency based prediction market, Polymarket operates onchain. This means that any bet placed can be verified directly on the blockchain database to have occurred. For example, this transaction shows a user who bought 20,000 shares of Trump winning the election at $0.54 per share. The transaction information contains the exact amount the user paid, when the transaction occurred, as well as the address of the contract that the user deposited into.

Why is this important?

This level of transparency would be impossible to achieve with traditional election predictors. News sources such as CNN, Fox, and AP that were all covering election outcomes and displaying their own percentage predictions for each candidate would never openly source their data. Unlike with Polymarket, there is no direct way to determine what their predictions are based on and how they are derived. That is why different news sources had such drastically different predictions and could present their own data, with viewers having no choice but to believe what the media was showing them without any proof behind these numbers.

Besides full transparency, having a prediction market that’s built on blockchain creates a way for users to buy or sell their positions at any time and with very minimal cost. Since Polymarket operates through completely automated smart contracts that don’t require any human intervention (learn more about smart contracts in the Overview of Blockchain section), this market can be traded 24/7 with no restrictions. That means that a user would be able to follow the price trends of an event and buy/sell shares whenever they see fit, and these trades would be settled instantly at the given price instead of having to wait for the next business day and risking price slippage. While users do have to pay fees per each transaction, these fees average only $0.01 to $0.03 per transaction.

How Polymarket Defined Market Sentiment Towards the Election

It’s important to make the distinction between market and political sentiments before discussing the odds presented by Polymarket. The view of a main stream media news source towards the odds of one candidate winning compared to the other can be directly impacted by political bias, where that news source can have an inherent left or right leaning bias that impacts their predictions.

Market sentiment, which is what is being displayed on Polymarket, is generally not impacted by political bias but rather financial bias - those placing bets may be doing so purely out of financial motivation based on odds, as opposed to genuinely liking and/or voting for that candidate. For example, they could have voted for Trump but bet on Harris to win purely based on speculation and better odds.

Regardless of whether it’s financial or political bias, each poll conducted will always contain some degree of bias. There will always be a certain margin of error, no matter who conducts the poll. Polymarket’s system is nowhere near perfect, but many people believe that the odds that they display are overall less prone to bias and can be a more accurate predictor for the outcome of an event like the 2024 election.

If you’d like to see some of this data for yourself, here’s a great dashboard from The Block that looks at all of the statistics behind Polymarket including volume, open interest, and interest specifically in betting on the election. Here’s another great dashboard from Dune Analytics that covers even more data.

Does looking at the all-time view on Polymarket suggest that it anticipated a clear winner throughout the entire election, or was it just reflecting the odds given by other media sources which makes it biased?

The odds presented on Polymarket were directly linked to market sentiment, where the live odds at any given moment specifically reflect the opinions of the bettors at that time. However, these odds increased in accuracy with higher volume. The more people that bet, the more accurate the odds will become because of the increased diversity of opinions. For example, if 80 out of the first 100 bets were pro-Trump, the accuracy of the given odds would be nowhere near as realistic as if there were already 10,000 bets placed, where the odds even out as more people with different opinions placed their bets as well.

Source: Polymarket Presidential Election Winner 2024

Source: Polymarket Monthly Volume (The Block)

Looking at Polymarket’s volume compared to the odds of each candidate demonstrates this exact theory:

• January to June: Polymarket experienced relatively low volume, around $50 million per month across the entire platform. Because the platform wasn’t as well-known yet, few people bet on the presidential election which resulted in skewed odds - Trump had a 45-60% chance of winning, while Harris was at 1%.
• July: Total volume increased significantly to $385 million, with the majority being bet on the election event. This influx led to more competitive odds that more closely resembled the ones given by mainstream media, with both candidates exchanging leads in the 45-55% range.
• August: The upward trend continued with Polymarket seeing $470 million in total volume. The odds continued to range between 45% and 55%.
• September: Total volume was over $530 million, where Polymarket was becoming more well-known and more people began to place bets on the election. The odds continued to fluctuate in the 45-55% range.
• October: Polymarket saw a substantial surge to $2.5 billion in total volume - a fivefold increase from September. This surge created much more defined odds, showing Trump having a 50-65% chance of winning and Harris having a 35-50% chance, indicating a more clear favorite as more people placed bets.
• November: Election day and the 4 days leading up to it saw over $1 billion in total volume. The odds saw one more brief period of uncertainty before surging for Trump and diving for Harris as results began to come in.

Unlike mainstream media that was able to define their odds for each candidate early into the race, Polymarket did not indicate a clear winner until October. However, that is because it came to a natural conclusion based on incoming betting volume. The larger the volume, the more opinions were consolidated to show who bettors believed was more likely to win. As a result, Polymarket was able to achieve a much higher diversity of thought where the odds were not inherently prone to bias.

Outsized bets can skew the prediction odds, favoring one candidate and distorting the perception of the actual odds. How vulnerable is Polymarket to this issue?

While a betting market will always be susceptible to short term volatility created from large bets, these bets are generally not sufficient enough to alter the market’s overall trend direction. Polymarket was able to aggregate a broad range of different sentiments towards the election, creating a natural trend that depicted the odds for each candidate. There are two key points that back up this belief:

1. Several different media sites reported that there was an outsized bet of $30 million placed on Trump, pointing to potential insider interference in an attempt to swing the odds in his favor. What they didn’t highlight was that (a.) this was actually the cumulative value of 4 different accounts and not just one person, and (b.) that these bets were placed by European residents and not US based users, where these bettors claimed to be purely financially motivated in their decision (Reuters, 2024). They also failed to mention that there were multi-million dollar bets placed on Harris to win as well, such as this user who placed a $3.8 million dollar bet, this user who placed a $2.8 million dollar bet, and this user who placed a $2 million dollar bet.
2. Trump’s surge in his odds of winning was not directly correlated to these outsized bets, but rather were part of the larger market trend that already concluded that he is more likely to win than Harris. Kalshi, a US based prediction market that depicted very similar trends compared to Polymarket, issued this statement on why their given odds for each candidate differed from those given by mainstream media sources: "Our stance on Trump's surge in odds is that it's all part of normal market activity. Trump is simply gaining popularity, and prediction markets aggregate information from a wider audience at a faster pace than polls” (Reuters, 2024).

If a large bet is placed there may be short term volatility, but as other bettors react to these odds and place their own bets, the market tends to correct itself. Additionally, if the market is already trending in a specific direction (such as Trump being more likely to win than Harris), a single large bet is unlikely to create a trend reversal unless it’s part of a much larger shift in sentiment among a significant number of participants.

A bet doesn’t necessarily reflect a person’s actual beliefs and the average bet size may not be done in relation to that person’s net worth, meaning that it could simply be a financial decision that doesn’t account for their level of conviction. How can a market based on such financial bias accurately reflect the odds of each candidate?

The chart below shows the average bet size throughout the entire campaign:

Source: Polymarket US Election - Avg Bet Size (Dune)

While there was significant volatility prior to June, the average bet size stayed under $1,000 throughout June to November. When comparing this data to the odds presented on Polymarket (from this previously shown chart), the data demonstrates how diverse opinions amongst the bettors were, which was shown by the fluctuating odds between June and November.

Normal people were placing small to medium sized bets on both candidates. the large volume of independent bettors with differing opinions on the election is what shaped the odds, not large bets placed by whales. Polymarket does not account for whether the reasoning was politically or financially motivated, only that the person betting had enough conviction in that candidate to believe in their bet to hit. This principle applies to both small and large bets - there’s no way to measure the size of the bet in proportion to the person’s net worth, so their bet on a candidate shows the same level of conviction regardless of the size.

Conclusion

While prediction markets like Polymarket are heavily influenced by betting volume, they ultimately reflect the collective sentiment of all of their participants, creating a real alternative to mainstream media for viewing the predicted odds for each candidate.

Comparing the volume of the platform to the odds given for each candidate demonstrated that as more participants engaged and volume increased, the odds stabilized and offered a clearer picture of the likely outcome. Although large bets can create short-term volatility, they rarely shift the overall market trend, as the aggregated views of numerous smaller bets correct the odds.

Furthermore, the motivation behind a bet (whether purely financial or political) matters less than the act of betting itself, as it inherently indicates the bettor’s level of conviction in the candidate's chances. This protects prediction markets from being directly subject to the political bias that is common in mainstream media polls.

This system is nowhere near perfect, but it effectively demonstrates how blockchain, an open source technology that aims to create full data transparency, can act as a fair system for tracking the odds of an outcome based directly on public opinion.

Works Consulted

The Block. Polymarket Volume (Monthly). https://www.theblock.co/data/decentralized-finance/prediction-markets-and-betting/polymarket-volume-monthly

The Block. Prediction Markets and Betting. https://www.theblock.co/data/decentralized-finance/prediction-markets-and-betting

Cornerstone Research. SEC Cryptocurrency Enforcement. https://www.cornerstone.com/insights/reports/sec-cryptocurrency-enforcement/

Dune Analytics. 2024 US Presidential Election Polymarket. https://dune.com/lindyhan/2024-us-presidential-election-polymarket

Polymarket. What is Polymarket? https://learn.polymarket.com/docs/guides/get-started/what-is-polymarket/

Polymarket. Presidential Election Winner 2024. https://polymarket.com/event/presidential-election-winner-2024?tid=1731097111276

Reuters. Large Bets on Election Prediction Market Are from Overseas, Source Says. https://www.reuters.com/world/us/large-bets-election-prediction-market-are-overseas-source-says-2024-10-18/