Prediction markets — a key to understanding the future
An interview with Oaklins’ Andrew Eaddy
Attempts to predict the future are as old as civilization itself. Historically, these efforts have taken many forms, including placing bets on the outcome of certain events, such as who would become the next pope or whether the price of grain would pass a certain threshold. In the 20th century, Austrian economists analyzed these practices and formalized the idea of using “prediction markets” to forecast political and financial outcomes.
In 1988, the University of Iowa’s Tippie College of Business introduced what may be regarded as the first modern prediction markets. Since then, the Iowa Electronic Markets have often been used to predict the results of political elections with greater accuracy than traditional polls. In 2004, the American journalist James Surowiecki championed the use of prediction markets in his seminal book The Wisdom of Crowds, and the term came into vogue.
Andrew Eaddy is an associate at Oaklins DeSilva+Phillips in New York and a co-founder of baserate.io, which publishes Global Guessing, a geopolitical forecasting site, and Crowd Money, a podcast and newsletter about prediction markets.
Eaddy is an authority on prediction markets and how companies like Google are using them to improve their forecasts and understanding of important trends. In this interview with Oaklins editor Elliot Kass, he explains how prediction markets work, the value that they bring to business, and why the demand for the tools and services that comprise them is likely to grow.
You and your baserate.io co-founder, Clay Graubard, spoke recently at Consensus 2022. Why did you attend this conference and what was the subject of your talk?
Andrew Eaddy: Consensus is one of the premier events in the USA for the blockchain community, and it attracts many of the leading practitioners in this space. We were invited to speak about prediction markets and how they’re related to blockchain technology. Our talk was about the societal good that can come about through the use of prediction markets at scale.
What exactly are prediction markets? How do they work?
At the most basic level, prediction markets are markets where the participants are able to trade on the outcome of future events. Will the temperature be above 85 degrees tomorrow? A market participant can either buy shares of “yes, it will be above 85 degrees tomorrow,” or “no, it won’t.” Upon resolution — in this case, the next day when you see what the temperature is — those who made predictions are paid out based on their wagers.
What are the different kinds of prediction markets?
One major distinction among different types of prediction markets is the type of currency that market participants use to place their wagers. Some operate on the blockchain and participants use cryptocurrencies to place bets; with others, participants use fiat currencies like dollars or euros to make their wagers. There are also markets that don’t use any currency. Instead they use points or some kind of reputational system, which allows people to establish their reputation for accuracy within a certain community.
This ties in with the idea that prediction markets are about accountability. They help reduce bias and encourage people to “put skin in the game.” You only bet on an outcome if you truly believe that it’s going to occur. Being accurate is more important than saying something that will be well received by a particular audience.
What do the participants in these different markets bet on?
Many prediction markets focus on economic and financial matters. There are some, for example, where you can place bets on whether or not the Fed will raise interest rates above 2.5% following the next FOMC meeting. Others are tied to US GDP, and participants attempt to predict how fast it will rise or fall in a given quarter. There are also markets where people wager on the price of oil or some other commodity as of a given date. In some respects, these can be very similar to the futures markets, with the major difference being that a prediction market participant is betting on the outcome of a specific event.
But there are also prediction markets that address public health concerns, geopolitical issues and many other types of questions and trends. There are some markets, for example, that allow people to wager on who’s going to win the UK Conservative Party leadership election or whether Space X will land people on Mars by 2030.
What is the significance of prediction markets? Why do they matter?
We live in what’s been called “the age of punditry.” All sorts of journalists, analysts, politicians and other self-appointed experts make predictions all the time, and it becomes very difficult to keep track of them all and whether or not they come true. But unless we’re able to do this, we don’t know who to trust or who we should be listening to.
One of the most important societal benefits of prediction markets is that they can provide a check on all this punditry and the media’s tendency to sensationalize certain viewpoints. Based on the market sentiment registered by various prediction markets, it could turn out that what The New York Times wants us to believe about a certain event is not all that accurate. So prediction markets provide a way of evaluating these prognostications.
Prediction markets can also help companies make better business decisions. Companies like Google and Microsoft run their own private prediction markets to gather business intelligence. They use these markets in lieu of surveys to gauge employee sentiment on important business matters — such as whether a major project will be completed on time or whether a certain business decision will have a positive impact. Utilized in this way, a prediction market can provide an uncensored view of employees’ beliefs about various matters of importance to the company.
It's widely understood that markets are efficient means of obtaining good information. That recognition underpins Google’s view that an excellent way to obtain good insights is to crowdsource and then average out the different viewpoints within the company. A single individual’s opinion, regardless of the person’s expertise or how well-informed they are, may be an outlier. Taking stock of all the sentiments held throughout the company can provide a much clearer and more accurate picture of what’s really taking place.
Google, in fact, has just turned its internal prediction market function into a B2B tool and now offers prediction markets as a service. This lets other companies use Google’s technology to create their own prediction markets for gleaning insights about their business. These insights are generally better and more meaningful than what they could obtain by simply surveying their employees.
Ford is another very large company that runs its own prediction market. For companies that operate at this scale, it can be a big challenge to track what’s happening throughout the business, and they have discovered that prediction markets are an effective tool for staying on top of their far-flung operations.
Prediction markets are also a good way to promote social engagement, which is why both Facebook and Amazon — the latter through its Twitch video game streaming platform — have been experimenting with prediction market apps. In the case of Twitch, for instance, streamers can bet on their favorite players and place odds on whether they’ll make it to the game’s next level. If they bet correctly, Twitch rewards them using its own currency, which members can use to buy merchandise on the site.
Do the different types of prediction markets work differently?
Apart from the different type of currency that can be used to make a bet, markets differ in terms of the kinds of bets you can place. Some markets are binary, you either bet “yes” or “no,” while others make use of a sliding scale, where the participants can wager on the degree to which something will take place. So, for example, in a binary market, you can place a bet on whether or not currency rates will rise, while in a non-binary market you can wager on just how big the rate increase will be.
The binary markets tend to be centralized. Kalshi, for instance, is the first federally regulated exchange where you can trade on the outcome of events, and the markets it sponsors are binary and highly centralized. This privately held startup is backed by Sequoia, Y Combinator, Henry Kravis and Charles Schwab, among others, and has raised US$30 million to date.
With a centralized platform like Kalshi, all disputes are settled by the market operator, which charges participation fees but also provides liquidity. Other platforms are decentralized, which allows for lower fees and non-binary trading. These markets tend to run on the blockchain, which enables participants to keep a very accurate record of every transaction and close tabs on how accurate different forecasters are over time. Decentralized markets are akin to open-source software — they are run by their users, who also provide the markets’ liquidity.
But the most important mandate for any prediction market is volume, because you need large numbers of participants to generate accurate information. And the key to getting enough people to participate in a market is having the right incentive structure. So, that’s what many of these differences come down to — how these different platforms incentivize their traders.
What sector do prediction markets fall into? Where is the investment in these firms coming from? Have they experienced much M&A activity?
Prediction markets tend to fall into the fintech and the financial software space. This is especially true of companies like Kalshi, which have gone through a very traditional startup journey and have extensive VC backing.
The investment in these companies comes primarily from VCs and growth equity firms. We haven’t seen much by way of M&A activity as of yet.
Right now the space is very fragmented — there are many different prediction markets with more starting up every day. And that fragmentation makes it difficult for many of these markets to achieve the requisite trading volumes that they need to reach critical mass.
Over the next few years, M&A activity is liable to increase as a lot of these companies that are getting funded have to hit certain metrics and increase their valuations, and this will very likely force consolidation to take place.ANDREW EADDY, ASSOCIATE/NFT LEAD, OAKLINS DESILVA+PHILLIPS
What types of companies make up the space?
The types of companies that participate in the prediction market ecosystem include the companies that provide the blockchain backend as well as market platform providers like Kalshi. But there’s also the publishing content that’s posted to these sites to attract market participants, and the role of content is increasing, since this is a very cost-effective way for a market to build its user base. So, content providers are also becoming a bigger part of the ecosystem.
Speaking of content creation, there’s a lot of buzz right now about creating an alternate digital reality, referred to as the “metaverse.” Is there a relationship between this concept of the metaverse and the growing interest in prediction markets?
I see the metaverse as a collection of trends, mostly having to do with content creation and the way that businesses operate and organize their employees in a digital environment. I think the connection to prediction markets has to do with economic and social interoperability. Less so, perhaps, with the Kalshis of this world and their centralized trading platforms. But for some of these blockchain-based, decentralized platforms, the metaverse provides a very interesting framework for how social hierarchies will be created and the different ways that people will participate in political and business decision-making.
So, I think it’s very exciting that we’re starting to see prediction markets implemented on Twitch and some other social platforms like Reddit. As we see these companies move more deeply into the metaverse, they will probably increasingly turn to prediction markets to help them engage more fully with their users.
How active is Oaklins in the segment? Are there deals that you’re striking right now among market players?
It’s still a bit early for the M&A side of the market’s dynamics, so we’re not playing in the prediction market space just yet. Right now, the funding for these companies is coming from series A, series B and angel investors.
It is, however, a space that we want to be active in. It’s important that we be smart in this area, because of the impact that it will have — and is already having — on the other sectors that we move in.
Right now, in terms of prediction markets, we’re focused on exploration, education and thought leadership. That said, we’re also ready to have a serious conversation with any clients that come along.ANDREW EADDY, ASSOCIATE/NFT LEAD, OAKLINS DESILVA+PHILLIPS