Take political betting markets literally and not seriously

The US election betting markets are having a moment. Relative newcomers Polymarket and Kalshi are attracting billions of dollars in transactions, while the more established PredictIt continues to operate some markets amid a legal battle with the CFTC.

These markets all work in broadly similar ways: bettors can trade stocks tied to a specific outcome, such as the winner of the US presidential election. The shares trade at prices between $0 and $1. When the underlying event is resolved, the shares tied to the correct outcome are paid out at $1 and the rest become worthless.

However, betting markets have exhibited strange behavior in recent weeks, with the implied odds of a Trump presidency increasing in a way that is inconsistent with neither polls nor electoral models.

Of course, the markets could be right in the end. Trump could win. Whether the current odds are fair is a more philosophical, unanswerable question.

Polymarket is the largest election prediction market by volume and is reportedly only open to traders outside the United States. As I wrote on MainFT, the price action was largely driven by a small group of anonymous accounts on Polymarket that use disproportionately large limit orders.

The largest account in this group, using the username @Fredi9999, owns more than 20 million shares in the President's main market alone. Together, the four largest holders of Trump stocks have invested nearly $40 million across all markets on the site, almost all in holdings that imply a big Trump victory.

By analyzing trading and deposit patterns, traders and internet experts have speculated that the four accounts (hereinafter referred to as FrediGroup) could be a single person or group. It is difficult to verify a connection with certainty, but there is plausible evidence that the accounts are linked. Furthermore, their public comments have a similar (pretentious) tone and (messy) writing style.

Comments and messages to other users have revealed some scant biographical details. One account claims to be French and to have lived in New York from 2000 to 2006 and worked as a trader. Another describes himself as an “investor and statistician” and repeatedly claims to have “no political preference,” but rather a sincere belief that Trump is likely to win the election and that the true odds should be “75 percent Trump.” . Very fascinating!

Still, there are some reasons to believe this FrediGroup FrèdiGroup could be acting a little stupid.

One commenter seemed a bit confused about the definition of a basis point, an odd mistake for an “investor and statistician.” The reports used at least four different spellings or abbreviations to refer to the state of Pennsylvania, all of which were incorrect or unusual. And they seemed to give up a surprising amount of personal information when contacted by another dealer.

Overall, her tone and reasoning seem strangely indifferent to someone with millions of dollars at stake in a prediction market.

Is FrediGroup one or more wealthy believers taking advantage of what they see as a bargain, or is there something more complicated going on?

One possibility, which is being widely discussed in some parts of the internet, is that FrediGroup is deliberately manipulating the market to create the impression of increasing momentum for Trump.

This theory is made even more compelling by the fact that numerous conservative commentators have been posting Polymarket screenshots all over the place in the weeks since FrediGroup's buying spree began

In other words, market action may be contributing to the perception of a large Trump lead – at least in some parts of the internet and media. This could be beneficial for a pro-Trump supporter or perhaps someone investing in neighboring markets that would be influenced by the perception of a likely Trump presidency.

Perhaps there is a motivated partisan behind FrediGroup, either directly or through the provision of funds behind the scenes. However, with the public information available, it is nearly impossible to prove this.

Another possibility raised by Barnard economics professor Rajiv Sethi, who I spoke to last week, is that FrediGroup may be trading on inside information. You may know something particularly devastating about Kamala Harris or her campaign.

If that were the case, it might be in their best interest to behave exactly as they do: playing the role of a confident, potentially irrational trader, they would keep Trump stock prices lower and improve their margins .

Insider trading is not impossible, but likely? Voting is already well underway in many states. After one of the most chaotic six months in American electoral history, what are the chances that an explosive secret will be kept secret that will swing the election in Trump's favor?

Is it possible that FrediGroup is trying to make a tidy profit using a sophisticated “pump and dump” system? After all, the 41 million shares the four accounts owned were about $2 million more expensive on Friday than when they were purchased.

Given the size of their holdings and the market's order book, such a plan is highly unlikely to work.

All trading on Polymarket is public (that is). ~on the blockchain~finally). This would likely make it harder for FrediGroup accounts to slowly start selling shares without causing a stir.

For weeks, with occasional interruptions, they did nothing but buy Trump stocks. Once they publicly reverse course, the price could fall quickly and they won't have much room left. With an average price per share of about $0.55, it wouldn't cost much to put them underwater.

If selling stocks slowly is a problem, could they do so quickly? No, not even close!

As of Friday, there were offers for a total of 3.5 million shares above its breakeven point of $0.55. If FrediGroup could fulfill all of these bids at once before the price collapsed, it would make about $145,000.

That left them with about 38 million shares with a break-even value of about $21 million. In fact, each of the suspect group's four accounts holds a larger number of shares than the total number of shares on offer at over 1 cent.

So why aren't traders gobbling up all of Kamala Harris' potentially undervalued stocks? There may be a real fear that FrediGroup is trading on inside information.

It is also possible that the value-at-risk of supporting Harris remains unattractive given the overall low market liquidity and the fact that most pundits view the election result as a coin toss. Surely it won't do much good to hurry. As long as FrediGroup's large limit orders continue to come in, Harris shares should be cheaper tomorrow than they are today.

And there are other structural features of some of these sites that may affect their predictive power. PredictIt, for example, charges fees of just over 5 percent on withdrawals, which distorts the incentive to trade with low perceived margins. This may not be as important for large or habitual traders who roll profits into follow-up bets, but it does deter a new bettor from making a deposit for a one-off bet with limited winning potential, especially since the site sets the maximum limit for traders in each individual market set at $850.

Maybe it's the Trump trades authentic. And the dealers could end up being right! Viewed purely as a momentum trade, they have been right so far. But a dealer could act for different reasons to varying degrees.

All of this complicates the idea that prediction markets are “more accurate than surveys,” as Musk tweeted in early October, when he was either unaware of or ignored all of the research testing the hypothesis.

It's easy to imagine Musk and other influential Trump supporters creating a feedback loop that directs followers to prediction markets and then uses them as a measure of campaign momentum. Of course, as an actual measure of choice, it would be like holding a thermometer over a radiator and then holding it up to show how hot it is outside.

So when can we trust these markets if we have reason to believe that prices may be distorted or overly susceptible to interference?

They can be most useful in the midst of large uncertain events, when there is a possibility that the underlying event could unravel at any time. This dramatically increases the risk of playing with the order book as there is always the possibility that one side of the bet will suddenly be wiped out.

A recent example of this was Joe Biden's resignation from the presidential election campaign. Prediction markets were particularly useful for quantifying fluctuations in the general consensus in a way for which there was no obvious substitute.

It was also important that they had the opportunity to find a solution at any time. Any attempt at market manipulation carries a greater risk of ending up in nothingness.

This will of course also be the case on election night. When the results from the key swing states come in, there will be a moment when the candidates' chances will reset to either 0 percent or 100 percent. (It's probably best not to expect a repeat of the 2020 election, when bookmakers like Predictit left the “next president” bet open after Joe Biden was declared the winner.) The returns for a bettor with a strong Hunch will be the same (and in a much shorter time), but there will no longer be any reason to bet against their true belief.

At this point we can probably start taking the opportunities in these markets a little more seriously.