Can ‘wisdom-of-crowds’ prediction markets outperform institutional economic forecasts?
Jump into the bullpen
So you live in Melbourne.
Every day you bring your shoes out, you take on the binary gamble of “Will it rain today?”
Let’s get more adventurous, actually. “Will it rain within the next two hours?”
You will notice our rain question has a binary outcome. Yes or no. One or the other. If we devised a game where you can monetise your correctly predicting this rain, you will have signed an event contract.
These games are the best ones to bet on because they have fixed payouts and valid expiration dates. The crowd only cheers while the bull strikes its horns in the pen. It is invisible when it simply runs around.
Consider yourself the bull. You’re in the pen, wondering where to make a dent. You can’t hit wherever; it’s no fun. So let’s learn what the optimal spot is.
Let’s go back to the rain question. To shoot your shot, you must first hold a position on the answer to this question. “Yes” or “no.” You lock your prediction in. But, you’re a bull, so you make a stupid decision:
You monetise your bet.
Welcome to the weather prediction market: a place to trade predictions about the weather.
You draw up a contract and tag a price to it that reflects the market’s expectation of its outcome (for example, 70 cents for “yes” and 30 for “no”). If you purchased a “yes” position worth 70 cents on a contract called “Will it rain tomorrow?” worth $1, and it rained the next day, you would earn a 30-cent profit. If you predicted incorrectly, you would lose the amount you invested in the contract.
Simple, right? You can draw up more complex event contracts by offering multiple-choice options or even offer partial payouts. The more complexity your contracts stack, the fewer participants it may draw, a simple reason being that we prefer simplicity and straightforwardness when investing what we feel are valuable assets to us. Ultimately, though, prices offered in a contract reflect a market’s perceived likelihood of all possible outcomes, not what they particularly mean to you.
This makes betting risky, where you wager money on an event contract whose outcome is beyond your control and often incalculable.
One way to offset the risks of the uncertain and largely unpredictable real world is by hedging: placing multiple bets on various potential outcomes as an insurance mechanism to better your odds of receiving a return regardless of the actual outcome.
Picture yourself betting on Spain to win the FIFA Final. You start watching the 90-minute game, and just over the 45-minute mark you see that the team hasn’t recorded any goals yet. You begin to contemplate in bated breath, and await the outco— wait, what if you placed a price on the other team winning as well? I mean, there are two teams, only one can win, but you can bet on both, right?
Yes, but no. No, but yes.
Firstly, this is what hedging a bet means. We know that two contradictory outcomes here cannot happen simultaneously; both teams cannot win. Here’s where hedging can help. If you don’t hedge, and Spain loses, you get nothing. If you do hedge, you diminish your losses and your pockets see some dough either way.
The good, the bad, and the ugly of betting is the uncertainty involved in it, but participating in any market carries financial risk, especially for future events.
Putting names to faces, the risk we have been trying to understand so far is called speculation. The markets we have been discussing so far are speculative platforms regulating sales of event contracts called prediction markets.
The global trading volume on prediction markets was about $24 billion as of April 2026 (1). Remember our rain example? Simply put, that is exactly how these markets work. All our sports examples? Since 2024, Sports made up 80% of total trading volume on Kalshi (1).
Seems like massive fun!
But can all this fun be enough to retire a need for forecasting? If we can bet professionally, do professionals need to bet?
The boring way
Economic forecasts, in particular, are projections of everything that could possibly happen in the future, constructed on past and present evidence that forecasters analyse.
All policies, legislation, management decisions, environmental modelling, and market volatility trends rely on forecasting methods developed over decades, and practitioners of these methods all stoically approach their work.
Humans are optimistic, though. We place our bets often in what we hope would bring us successful outcomes. How do we ensure we arrive at them?
Crowds to the rescue
Crowds are not your local pigeon group that attacks your fries on a walk across the South Wharf.
These crowds are smart, considerate, and actually let you walk away with something.
It has been said that a group’s collective input is more coherent than clashing individual inputs (2); a group of 1,000 people can perhaps make a prediction about the weight of a big rind of parmesan, and be within a very agreeable margin of error from the rind’s true weight. All the group has to do is ask a lot of questions and employ their varied expertise.
While this may not be deemed wisdom by the ancient greats, this guess-timate game grows more interesting.
When we speak of the “wisdom of a crowd,” we imagine mass opinions and collective responses to important questions like:
“Will Anne Hathaway win Best Actress in a Leading Role at the 2027 Academy Awards?”
“Will mint chocolate chip ice-cream be recalled within the next month?”
“Can Tom Cruise pull off a 17-flip cartwheel at the 2028 Olympics?”
“Will the New York Knicks take the 2026 NBA title?”
You don’t have to be an expert in any of these fields to pick a side.
In a time where we usually resort to self-fulfilling decisions, we seem to place a good deal of faith in people’s judgement on worldly events, even with the possibility that the crowd might be completely wrong.
But hey, at least we’re all wrong together!
There is a presumed comfort in believing that 10,000 people have the right side of the coin on a binary dilemma, which brings us to the essence of trusting the wisdom of crowds; that their average judgement likely converges to the right solution (3).
In the 1840s, statistician Francis Galton observed the average of all entries in a “guess this ox’s weight!” competition and found the result to be astoundingly accurate to the actual weight, which also beat the individual guesses of alleged cattle experts.
Hmm, we might be wiser if we’re ‘together’
How do we make out the wiser crowd from the… well… not-so-wise crowd?
We start by analysing how our candidate crowd might think together and by ourselves.
James Surowiecki, in his 2005 book, The Wisdom of Crowds, pointed out the fallibility of a crowd as contingent upon its members’ decisions being independent from other members in the crowd (4).
Let’s say we observe a group of 5 people, all tennis enthusiasts, visiting the Australian Open. Incentivised by the allure of betting on the Final winner, they sit together and strategise. Three are Nadal fans, while two absolutely adore Djokovic. But they must all go with the same player.
The simplest path forward would be to average out wins for both players and see who has performed better in past matches; to look up statistics on the player’s performance over their last 2-4 games; how the player performs against specific opponents, serves, match-day conditions, and so on.
Excessively motivated, they all agree to go with Nadal because Tom, Vita, and Pen said, “Trust me.”
Djokovic wins.
A number of things went wrong here. Firstly, a group’s decision isn’t necessarily accurate simply because it is a consensus. In general, a crowd can agree on an outcome but be completely off the mark. Secondly, if everyone in the group made varied choices in the tennis player they would bet on, the group as a whole would be largely unsuccessful in being collectively accurate.
We learn from the tennis example that the average guess is likely to shift towards a misplaced bias when individuals within the group allow themselves to be influenced by each other’s guesses.
The tennis example also shows us how important collective informed decisions are. Researchers at ETH Zurich found that a group’s guesses narrowed down the more they came to understand each other’s choices. Social influence majorly defines a collective decision because its effectiveness stems from sound judgements from the group. Uninformed group consensus ends up far beyond an accurate response if the starting point is not chosen intentionally.
This is called ‘herding.’
The dangers of being ‘together’ and the benefits of making more friends
It shouldn’t be hard for a herd to decide if they like the field or not. But herding effects tend to be greater in problems for which no objectively correct answer exists (3).
Could knowing about the yield of the field help? A 2011 study by the Yale School of Management found that group predictions about football results were markedly different from the real outcomes due to overconfident decisions by fans with distinctly favourite teams.
Representation is thus important. People hailing from different cultural and professional backgrounds take varied approaches to a problem or a decision. Large-scale discussion platforms like Reddit, for example, have become hubs for discussions, arguments, as well as decision-making influenced by opinions from individuals and groups you’ve probably never met before.
When individual members of a crowd are cognitively diverse and their views properly aggregated, we can minimise collective error (5). Diversity is the single most determinant factor of the success of a crowd’s prediction, which also makes it the most likely condition to be violated in prediction and other financial markets (5).
Theoretically, a diverse group of problem-solvers make better collective guesses than a group of high-performing experts. A great team has each member bringing unique and valuable skills, depth of expertise, and a healthy willingness to contribute. Diverse minds thus do better when their decisions are averaged, than experts conservatively do.
Can a group of people coming together at random form a convincing enough market?
A team that hasn’t won a championship in the history of the game, can still win the title at an upcoming one. We can imagine forecasts that depend on historical behaviour for information having a difficult time predicting such an outcome. The underdog team winning would be a miracle for the fans, and an anomaly in the trend.
Such a miracle is an unfair advantage that crowds participating in prediction markets hold over forecasts, because historical successes and trends are not altogether representative of future behavioural processes.
In the same vein, a crowd can turn mad without prior notice and no longer think strategically about the conditions of their advantage. Like old, ancient dudes have said: “if you can keep your head when all about you are losing theirs… (6)” maybe you haven’t heard the news.
Could we make a crowd wise?
Davis-Stober et al (7) defined crowd wisdom as:
“A crowd is wise if a linear aggregate, for example a mean, of its members’ judgements is closer to the target value than a randomly, but not necessarily uniformly, sampled member of the crowd.”
Phew.
All of this is to say that a simple average of a group’s constituent individual judgements will be significantly more accurate than a single member of the crowd.
To make a crowd wise would be to create an optimal group whose collective ‘wisdom’ exceeds in accurately predicting outcomes than any randomly sampled member. For this, the group must be ‘maximally informative,’ meaning the more varied and unique bouts of information, knowledge, expertise, and domain experience members bring to their groups, the wiser the crowd they make.
Ready to join the prediction markets with your crowd?
The colossal, gargantuan power of prediction markets
Prediction markets hold an information aggregation advantage over other markets because they provide incentives for event outcomes. If we further use them as inputs for future decisions, this advantage may instead incentivise the manipulation of market prices.
Market prices tend to respond rapidly to new information. Leaks of early polling numbers after an election could reverse the course of a market abruptly and detrimentally alter results.
Prediction markets thus perform considerably better than alternative approaches to forecasting outcomes or futures if we simplify their range to binary outcomes and control for diversity requirements.
Now.
While you may think knowing better about the market, than the market, exempts you from its oblivion of magnificence, and heralds you with an unfair advantage, you might try to escape its clutches and step out into the sun without an umbrella.
With your best pair of shoes.
But then you remember you’re still in Melbourne, and so…
… you’re still in the bullpen.
References
Radde K. Trading volume on prediction markets has soared in recent months. Pew Research Center. May 27, 2026, Kalshi and Polymarket trading volumes dramatically increase since mid-2025 | Pew Research Center.
Dai M, Yanwei J, Kou S. The wisdom of the crowd and prediction markets. Journal of Econometrics 222, no. 1 (Part B): 561-578. doi: 10.1016/j.jeconom.2020.07.016. The wisdom of the crowd and prediction markets - ScienceDirect.
Ball, Philip. ‘Wisdom of the crowd’: The myths and realities. BBC. July 8, 2014, ‘Wisdom of the crowd’: The myths and realities.
Surowiecki, James. The Wisdom of Crowds: Why the Many Are Smarter Than the Few and How Collective Wisdom Shapes Business, Economies, Societies and Nations (Doubleday; Anchor, 2004).
Mauboussin M J, Callahan D. The Wisdom of Crowds in Markets: Crowd Behavior in Prediction, Betting, and Stock Markets. Morgan Stanley Investment Management. August 5, 2026, The Wisdom of Crowds in Markets: Crowd Behavior in Prediction, Betting, and Stock Markets | Morgan Stanley | Morgan Stanley.
Kipling R. If—. Poetry Foundation, A Choice of Kipling’s Verse. 1943, If— | The Poetry Foundation.
Davis-Stober C, Budescu D, Dana J, Broomell S. When Is a Crowd Wise? Decision 1, 79-101. doi: 10.1037/dec0000004. (PDF) When Is a Crowd Wise?
Back to
Issue 11
