By Phil Huber, ADVANCE COLUMNIST
For two summers while I was in college, I worked as a groom at The Meadows, the harness-racing track near Washington, Pennsylvania. Today it is called Hollywood Casino at The Meadows, but when I was there, it was simply The Meadows. There were horses in the barn, drivers on the track, trainers watching every detail, and bettors studying the program for an edge.
I learned something about betting in those summers.
A wager can make a race more exciting. It can make a person pay close attention to a horse’s form, the driver, the track, and the odds. Many bet on the horse’s name or number or the stable colors and hope for the best. But betting can also make people see what they want to see. It can persuade someone that a number on a tote board is knowledge, rather than a reflection of money already placed.
Now, at 79, I am learning about prediction markets—because apparently the country decided that elections were not complicated enough without turning them into a stock ticker.
They are not the same as pari-mutuel horse racing. In a horse race, people wager on horses in an event that lasts a few minutes. In a prediction market, people buy and sell contracts tied to a future outcome: Will a candidate win? Which party will control Congress? How many Virginia House seats will Democrats win?
But the two have something important in common. Both can create the illusion that odds are the same as certainty. That is why political prediction markets deserve our attention before their percentages become just another piece of political shorthand.
I do not know how many people in the Fredericksburg region are betting on elections. It may be very few. But political prediction markets are now available in Virginia, and the leading commercial platforms offer contracts on elections, including individual congressional contests and broader questions about party control of Congress.
The best-known platforms are Kalshi and Polymarket. Their changing percentages can create the familiar campaign vocabulary: a candidate is “surging,” “slipping,” “favored,” or, by lunchtime, pronounced politically deceased by someone with a phone and an internet connection.
We should be careful with that language.
The basic idea is simple. If a person buys a “Yes” contract at 60 cents, the market is usually read as giving that outcome a 60 percent chance. If the event happens, the contract generally pays one dollar. If it does not happen, the contract pays nothing.
That can sound like a forecast. It comes with a percentage, a colored chart, and enough confidence to make a weather forecaster look modest. It is still not the same thing as a poll.
A 60 percent market price does not mean 60 percent of voters support a candidate. It does not mean a reputable poll found a 60–40 race. It does not mean election officials have counted ballots or that a campaign has earned a mandate.
It means only that traders, at that moment, are willing to put a price on an outcome. That can be interesting, but it is not democracy speaking.
There is a respectable academic case for studying prediction markets. The University of Iowa has operated the Iowa Electronic Markets since 1988 as a small-scale, real-money project for teaching and research. Its markets have covered national presidential contests, control of Congress, and other political and economic events.
Researchers have found that the Iowa markets often produced relatively low forecasting errors and that some of their winner-take-all prices were reasonably well calibrated. But that record does not prove that today’s commercial political-betting platforms are equally accurate—or that their changing prices should be treated as a substitute for polling, reporting, or voter judgment.
The Iowa market is a limited-stakes academic project. Commercial political-betting platforms operate in a much louder environment where money, social media, partisan enthusiasm, campaign strategy, and the desire for a dramatic headline all meet. That is why citizens need to understand the difference between a poll and a price.
A poll tries to measure voter opinion. A responsible pollster identifies the people surveyed, the number of respondents, the dates of the survey, and the method used. Polls can be wrong. They often are, but readers can at least examine the method.
A prediction market measures what traders are willing to risk money on. Those traders may be informed. They may also be following the same polls, cable-news panels, campaign advertisements, social-media rumors, and partisan chatter the rest of us see. Some may bet from conviction. Others may be speculating, hedging, seeking a quick profit, or simply treating politics as entertainment.
That does not make a market price worthless. It does mean we should describe it honestly.
The risks are not merely theoretical. A Reuters analysis published in September found that in over 11,000 congressional-election prediction markets, a bet of less than $1,000 could move the displayed probability by 10 percentage points in 94 percent of the markets examined.
At The Meadows, the tote board changed when money came in. That did not make the favored horse a sure thing. It told you where the money was going. Political prediction markets work much the same way. A candidate can appear to move from 48 percent to 58 percent—enough for someone to declare a “surge”—without any demonstrated change in public opinion.
The Anti-Corruption Data Collective reported that 87 percent of the midterm markets it examined were either low-volume, with less than $10,000 wagered, or had higher volume concentrated among only a few bettors. We hear a lot about “the wisdom of crowds.” But if the crowd is small, anonymous, and easily pushed around by a few people with money, how much wisdom are we really seeing?
There is another concern: fraud and conflicts of interest.
In February, the Commodity Futures Trading Commission issued an advisory after two publicly disclosed Kalshi matters involving fraud or misuse of nonpublic information. In one, a political candidate traded contracts involving his own candidacy, even though Kalshi’s rules prohibited someone from trading a market whose outcome that person could directly or indirectly influence. In the other, a person affiliated with a YouTube channel traded contracts after gaining advance knowledge of unreleased content.
The federal government has already had to remind prediction-market companies that people with inside information—or the ability to influence an outcome—cannot be allowed to turn that access into a private betting advantage. The CFTC has warned that fraud, manipulation, wash trading, and misuse of nonpublic information can violate federal law. It also said regulated exchanges are responsible for maintaining audit trails, monitoring trading, and enforcing their rules.
At The Meadows, we did not ask the tote board which horse deserved to win. We watched the race. Our politics could use the same discipline.
The lesson is that the number deserves questions.
Who placed the trade? How much money has been traded? Was the price moved by one bettor or many? Is the market liquid? What, precisely, does the contract say? How will the result be settled? Does the supposed trend match reliable polling, local reporting, campaign activity, and public evidence?
Most importantly, does the number cause us to stop doing our work as citizens?
That is where the real civic risk lies.
If voters are told a candidate is a “90 percent favorite,” some may decide there is no reason to volunteer, donate, attend a forum, ask questions, or vote. If a lesser-known candidate is assigned a low probability, reporters and donors may decide that candidate is not “viable” before the public has had a fair chance to listen.
A forecast can then help produce the result it claims only to predict.
The people most vulnerable are not necessarily the people who place bets. They are busy citizens who see a neat percentage and mistake it for settled political fact. They are small donors and volunteers who may redirect their energy.
Democracy should not be a spectator sport in which the best-funded or most-talked-about candidate is declared inevitable before the voters have spoken.
There is nothing wrong with being curious about prediction markets. They may provide one imperfect clue about what some traders expect. But a clue is not a conclusion.
Before repeating the odds, ask whether you are looking at a poll, a market price, or a well-packaged piece of political theater.
Then do the harder work. Read local reporting. Attend a candidate forum. Compare records. Ask candidates how their plans would affect the family around your kitchen table: the cost of groceries and housing, health care, veterans’ care, schools, federal employment, military families, and honest government.
A bet may be entertaining, but a ballot is how democracy speaks.
***
Phil Huber is a retired Army Reserve colonel, federal civil servant and consultant who writes on civic education. He lives in Fredericksburg.
