By Victoria Liu and Djavaneh Bierwirth
Crowdsourcing opinions, experiences and answers is not news to the Internet. Per Wolfers and Zitzewitz, “the power of prediction markets derives from the fact that they provide incentives for truthful revelation, they provide incentives for research and information discovery, and the market provides an algorithm for aggregating opinions.”
Despite prediction markets being less well-known than traditional polls, platforms like Kalshi and Polymarket proved their worth during the 2024 Presidential Election. Media outlets and the public were previously unaware of the accuracy of prediction markets, even though academic research had long supported their effectiveness.
Let’s look at how these prediction markets work in terms of their components as well as their implications for the finance industry.
Contract Types
There are multiple types of event contracts, but the common aim is that they allow for traders to bet on the outcome of a specific event. The winner-take-all format is most common, and it reflects the market’s expectation that some outcome will happen. There are also index contracts, in which the payout varies along a continuum of outcome possibilities. The third is spread betting, in which the outcome is a cutoff and traders select on the desired cutoff. These contracts have different determinations – winner-take-all shows the market expectation of the probability of an event, index reveals the market expectation of the mean and spread gives the market expectation of the median.
Market Designs
Market design differences impact liquidity, capital choices and distribution of winnings. Continuous Double Auction is the most traditional system via a manual ledger whereas Automated Market Makers have the platform act as the universal counterparty. The other design choice is real capital vs. virtual because platform-specific currency can improve trader participation and incentives. The lowered barrier to entry increases trade volume. In more recent platforms, the pari-mutuel system of splitting winnings is standard. All bets are gathered into a collective pool and then split amongst winners. Alongside AMMs, these new structures allow for simultaneous predictions as well as a multitude of outcomes which further enhance market dynamics.
Functional Requirements
The way a betting market is designed and operated significantly impacts its accuracy. To create a successful betting market, several key elements are needed:
- Clear event definitions and outcomes that can be easily verified
- Agreed-upon sources of data to determine results
- Well-designed incentives that reward accurate predictions
- Safeguards against market manipulation and spam
Users are more likely to participate when they can earn rewards and when the outcomes being predicted are interesting or controversial enough to generate active trading.
Past implementations
One notable early implementation is the Iowa Electronic Market that the University of Iowa started in 1988 that allows users to trade contracts on the popular vote of the presidential election. Interacting with the platform is a stark contrast to how modern prediction markets now alight screens with real-time trading activity as well as niche, social-media driven prompts.
Large prediction markets such as Polymarket and Kalshi see upwards of $13 million in average turnover for each trading event. For context, there are numerous events available for bids, ranging from the probability of someone’s cabinet appointment to the price of Bitcoin by a date. The 2024 Presidential Election saw a peak of over $3 billion in trading volume just on Polymarket. These providers streamline participation via ease in account setup as well as various attention-grabbing mechanisms, which further encourage trading.
The power resides in friction-reduction: real-time analytics, automated market-making, and easy-to-understand graphics relieve the liquidity concern that plagues any market. And now, the markets are just a few clicks away. The ease of mainstream providers such as Kalshi and Polymarket accelerates the propensity to trade and therefore bolsters the influence of the platforms.
The ability to effectively hedge against specific risks creates market liquidity. Prediction markets are a little bit more nuanced than derivatives markets since they can cover solely event-based outcomes instead of price-based action in traditional derivatives exchanges. This opens up a realm of opportunity for traders both individual and institutional depending on the investments they pursue. This encourages a new market of potential hedges.
However, not all liquidity is created equal. Specifically in the context of prediction markets, uninformed noise bettors can derail the benefits of liquidity and decrease market efficiency per Flepp, Nuesch and Fränck. This plays into the requirement earlier that betting platforms should thoroughly incentivize informed betting. The 1999 research on market efficiency by Athanscoulis, Shiller and Wincoop bolsters the point that insufficient liquidity means being unable to fully diversify away risk in their example of small pension accounts of very young investors. The accounts represent low liquidity and the lack of ability to put up high margins required for risk diversification. Thus, as the markets become more liquid, we expect investors to capitalize on these previously unavailable opportunities.
The possible implications from these modern platforms are significant. Policy must carefully navigate the fine line between innovation and consumer safety. Betting markets could be parallel to gambling – are consumers manipulated into placing orders? What are the implications of betting markets on tangential markets? Given that Polymarket also has a line of competitors, how replicable is the strategy in setting up a sufficiently liquid market? Further, how much trading activity is for-fun retail investors versus traders looking to offload event-based risks?
Increasingly liquid prediction markets in an interconnected, digital today beckon promising utility for tomorrow. As platforms better manage the liquidity lever, users will be able to better hedge risks as well as capitalize upon informed opinions. Kalshi and Polymarket, the most applicable integrations, achieved high engagement across the elections, sports, pop culture and crypto. The platforms wield the ability to synthesize diverse opinions into probabilistic insights and they are remarkable in maximizing participation. As prediction markets evolve, policy will impact their resultant influences on advanced risk hedging, consumer behaviors and directional beliefs.
Victoria Liu and Djavaneh Bierwirth are Wharton Initiative on Financial Policy and Regulation student fellows. The views and ideas expressed in this post are those of the authors and do not necessarily represent those of the Wharton School or the Wharton Initiative on Financial Policy and Regulation.