Betting Markets VS. Polls: The 2024 Presidential Election

Audrey Tan ‘27

​With the 2024 presidential election concluded, and Donald Trump emerging victorious, the comparison between betting markets compared to traditional polling methods have been scrutinized. While opinion polls are commonly utilized, they have often proved to be flawed in past election cycles. As Americans are now allowed to bet on elections, it’s useful to evaluate which method may provide a clearer picture of electoral dynamics, and what the betting markets and polls got right or wrong after the 2024 election. 

Political betting has evolved from a niche activity to a more mainstream practice, fueled by online platforms and a heightened public interest in politics. Just this past election, over $100 million was bet on the election. It’s not new, however, as bets on presidential elections were even common during the 20th century, where betting in New York City controlled the market at the New York Curb Exchange. Nowadays, in the legally administered platform Kalshi, with more than $400 million at stake, platforms like Kalshi have allowed individuals to wager on various outcomes, including presidential candidates, congressional compositions, and specific policy decisions. The odds presented in the betting markets not only reflect the collective sentiment of bettors, who analyze the current events and campaign strategies to inform their wagers. These odds can shift rapidly based on breaking news and political developments, such as the Trump and Harris debate, where on  Polymarket, a crypto trading platform, the odds for both candidates were tied. 

Polling remains the standard method for gauging public opinion, as organizations like Gallup and Pew Research conducted surveys to capture voter preferences and issue stances; however, polling is not without its challenges: factors such as errors, question wording and timing can also influence results. For example, polls typically have a margin of error ranging from 2% to 5%, which can lead to many discrepancies in close races. Furthermore, the accuracy of polls can decline as election day approaches, especially when voter sentiment shifts rapidly due to unforeseen events. An example of this would be in Iowa, where polls unexpectedly showed Kamala Harris ahead of Donald Trump, which caught many analysts off guard. Events such as Trump’s high-profile rally at Madison Square Garden, influenced voter perceptions, as betting markets shifted dramatically. For instance, just after the rally, betting platforms like Polymarket and Kalshi recorded a noticeable increase in Trump’s odds, suggesting that the rally had helped effectively galvanize support for him at a crucial time. Conversely, Harris’s campaign focused on outreach efforts aimed at key demographic groups, including young women and voters, which contributed to her rising numbers, as shown in the graph below.

This shows how polls can reveal insights into different demographic groups, allowing for targeted campaigning. 

Historically, the 2016 and the 2020 presidential elections underscored the varying accuracy of polls and betting markets. In 2016, national polls showed that Hillary Clinton was leading, yet Trump won key battleground states, raising doubts about polling reliability. Conversely, betting markets suggested a closer race than many polls indicated. Reflecting on the 2024 election, with both polling and betting markets active, polls indicated a close race between Trump and Harris, often showing them within a few percentage points. 

The interplay between social media and traditional polling methods is also particularly noteworthy; for instance, during the 2024 campaign, sentiment analysis of social media posts indicated a surge in support for Kamala Harris following her debate performance. This shift was reflected in polling data shortly afterward, where her numbers had increased among suburban voters, particularly in key swing states such as Michigan and Pennsylvania. This shows how social media dynamics can precede and predict changes in public opinion; moreover, betting markets also reacted to social media trends. As Trump’s campaign leveraged social media to rally his base, betting platforms like Polymarket observed a spike in his odds, signaling a growing confidence among betters in his chances of winning. This correlation suggests that social media not only influences voter behavior, but can also affect how bettors perceive candidates’ prospects, thereby impacting betting market dynamics.  

As we move forward from the 2024 presidential election, the integration of betting markets and traditional polling methods could reshape how we understand and predict electoral outcomes. Given the discrepancies observed in past elections, there is a growing interest in developing hybrid models that combine the strengths of both approaches. Betting markets offer real-time insights that can quickly adjust to new information, while polls provide deeper demographic analysis and context about voter motivations. By leveraging the predictive power of betting markets along with the methodological rigor of polling, analysts could enhance the accuracy of electoral forecasts, meaning that there could be models created that account for the inherent biases in both systems, such as the demographic skew in betting markets, and the sampling errors in polling. As America continues to evolve, integrating these methodologies could lead to more nuanced understandings of the electorate, ultimately improving the reliability of prediction in future elections.