Ah, we're baaaack with THIS wonderful topic It's been, like, three weeks (almost... #1572831) without Den musing on/ranting about prediction markets? (#1559479 #1556891
(Look, someone almost tries to steal my thunder! #1578574)
Bloomberg knows how to ROPE US BACK IN
To get an idea of how an election will shake out, a person might look at polling data. Increasingly, they might also look at a candidate’s odds on a prediction market, where users place bets against each other on the outcomes of real-world events.
You know why, right?
BECAUSE MONEY IS THE MODERATOR!
Yeah, we know: #1527191, #1530730
Prediction markets have soared in popularity following a landmark 2024 court victory for Kalshi Inc., a leading platform. In short order, they have become an influential tool for predicting elections. A growing number of media outlets feature real-time market odds in their coverage, and several have entered deals with Kalshi or Polymarket, another major platform.
Prediction markets have gained a reputation in some tech-friendly circles for being more accurate than traditional election polls. This perception was burnished by markets on both major platforms predicting that Donald Trump would win the 2024 presidential election. Mainstream polls slightly overstated support for his opponent, Kamala Harris.
Makes perfect sense, #1497608.
For nearly a century, opinion polls have been the primary instrument for gauging public sentiment ahead of US elections. The practice took a reputational hit in 2016, when polls underestimated Trump in several crucial battleground states. In the years since, some polling skeptics have embraced prediction markets.
"Though the picture is unclear, some early academic research suggests this perception is accurate for major elections."
critics say turning a civic process into the object of financial speculation poses new risks for democracy.
Perhaps, but the critics are morons and democracy is overrated so whatever. Though, this history I wasn't aware of!
In the early 1900s, bettors wagered on elections outside of the New York Stock Exchange, creating odds that were reported in newspapers. The practice declined as state laws and court rulings made election betting widely illegal, and scientific polling became the favored predictive tool for elections.
Kalshi ruling in 2024, and Polymarket settlement in 2022 and then... voìla: "a boom in the prediction market industry followed"!
Today, Kalshi and Polymarket users trade on everything from whether the Los Angeles Rams will win the Super Bowl to whether the federal government will confirm the existence of aliens this year.
Nearly half of US states have picked legal fights with prediction market platforms, arguing that they are sportsbooks disguised as derivatives markets and should be regulated by states, which oversee gambling. Judges have split on the question of whether prediction markets should fall under the purview of states, teeing up the conflict to potentially reach the Supreme Court.
funky, fuuuuunky. US legal battles are fun and exciting. (I wonder if there'll be a contract on first, whether the question makes it to the Supreme Court, and secondly whether they'll strike down pred-markets)
Here's a not-so-shocking observation for those of us in the know:
More people use electoral prediction markets to understand how candidates are performing than to try to turn a profit. Kalshi estimates that 75% of visitors use its site only to see what markets forecast, not to trade.
Even I made that argument in my belated "Social Value of Prediction Markets" article (#1556891).
...In the run-up to the 2024 presidential election, news outlets began citing odds from prediction markets the same way they might have reported on the latest polling data. A candidate’s “Polymarket odds” were held up as another data point in the horse race-like media coverage of the election.
"Journalism is better when it’s backed by live markets," -- Polymarket, Twitter
PREDICTION MARKETs:
especially illiquid ones, are vulnerable to manipulation by those seeking to influence perceptions of an election; strategically-placed bets can temporarily move the market’s odds, creating the appearance of a shift in a candidate’s prospects.
ELECTION POLLS:
[especially when] conducted by reputable, nonpartisan organizations don’t have that problem, but they can be vulnerable to methodological flaws such as biased sampling, leading questions or flawed weighting.
Prediction markets are faster, really immediate; polls are slow, and by the time results come out, dated. Polls have social desirability/self-delusion problems ("when people answer polling questions, they are stating who they intend to vote for, not who they think will win"); markets have illiquidity or manipulation problems.
Election market surges can be bought. Markets on major races, such as a US presidential election, tend to be highly liquid, meaning they attract a lot of dollars in trading volume and are therefore difficult to manipulate. But the majority of election forecasts on Kalshi and Polymarket are far less liquid.
True, perhaps, but the more interesting observation is how much dough it'd take to make them stay there (for e.g., 24-72 hours)
....and mispriced markets are FREE MONEY
What About Insider Trading?
I mean, at some level it's THE POINT and PURPOSE with these markets... financially induce insiders to come forward with hidden info.
The company explicitly forbids candidates and campaign staff from trading on their own elections, and bans people from trading in markets where they are “decision makers” or can influence the outcome. Polymarket prohibits users from trading on confidential information and wagering on events where they can influence the outcome.
Boring.
For high salience and high liquidity markets like presidential elections, I absolutely do believe that prediction markets can predict better than polling.
It's the low liquidity low salience markets that I'm highly skeptical of.
On the salience point, markets on very low or very high probability events probably won't be very good because bettors won't find the mispricing to be worth the risk.
Something like will there be a 9+ point earthquake in Seattle this year is low probability, but whether it's 1% or 0.1% makes a huge difference to people because it would be such a catastrophic outcome.
https://twiiit.com/Polymarket/status/2024217326065783058