No, Polymarket Whales Aren’t Evidence of Prediction Market Manipulation

It seems likely that a small group of traders (or possibly a lone individual) have pushed up former President Donald Trump’s odds on Polymarket, the crypto-based prediction market, by heavily betting on the Republican nominee.

There has been a good deal of handwringing about this, both from supporters of vice president and Democratic nominee Kamala Harris on X (formerly Twitter) who have an obvious incentive to call the markets manipulated, and from the mainstream press. “[T]he surge might be a mirage manufactured by a group of four Polymarket accounts that have collectively pumped about $30 million of crypto into bets that Trump will win,” The Wall Street Journal speculated.

Indeed, there may be a lone bettor who has gone very long on Trump. But that’s not necessarily evidence of something sinister.

Aubrey Strobel is a communications partner at Bitcoin app developer Trust Machines and the host of The Aubservation podcast.

First, the obvious point: it’s not just Polymarket where Trump is rallying. Trump “yes” shares are trading at around 59.9 cents on Polymarket right now, indicating the market sees a 59.9% probability he will win. (Each share pays out $1 if the prediction comes true, and zero if not.) Predictit – a U.S. platform with strict betting limits (which would in theory be immune to the Polymarket whale) – has Trump at 56%. Kalshi, a U.S.-based, regulated platform, has Trump at 58%. If you think this price action constitutes “manipulation,” you would have to wonder why Kalshi – a platform where manipulation would be reportable – is trading in line with Polymarket.

The online bookies also have Trump ahead: 59% at BetOnline, 55% at Betfair, and 60% at Bovada to name a few. And Trump is also trending up in the poll aggregates, so it’s not inconceivable that his true odds of winning are also increasing. Trump’s rally in the prediction markets coincides with Harris’ rough Fox News interview and new polls show Trump trudging upward.

Secondly, the fact that a single anonymous entity has made large bets is not by itself evidence of “manipulation,” as many Harris supporters are claiming. Bloomberg columnist Matt Levine flatly says: “this does not look like market manipulation: Fredi9999 is not buying sloppily with the effect of pushing up the price, but rather buying carefully, in a manner that seems designed to get him a lot of Trump contracts for his money.” The simplest possible explanation is that a trader simply thinks Trump is underpriced and is willing to bet heavily on it.

The mere existence of a large buyer does not imply manipulation. The entire premise of markets is that prices compress available information by rewarding those who take risk to express their views. The identity of the traders or the distribution of trades is irrelevant; in theory, everyone has a motive to extract information from markets by betting when their belief about the fair value of an asset diverges from the market’s value. Markets don’t need to be democratic in order to be reliable. They just need the most informed participants to financially express a view. Individual traders having concentrated ownership of an asset doesn’t in any way delegitimize the price. No one questions the price of Apple stock because Warren Buffett’s Berkshire Hathaway owns a lot of it.

This is a long-winded way of saying the market is never “wrong.” It simply reflects all available information. If you correctly disagree with the market, you can be rewarded for that belief, by betting yourself. U.S. users have alternatives to Polymarket, which is barred from serving them under a regulatory settlement. If you believe the Polymarket whale a) has meaningfully pushed up the price of the Trump contract, and b) is wrong, you can simply bet against him or her or them by going long on Harris. Even though it’s not risk-free – Harris still needs to win for your bet to pay off – if you thought her “real” odds were 55%, you would be buying something worth 55 cents for 40 cents today. Even if you might not be willing to do that, other market participants will. So if the Polymarket whale is indeed misinformed, now that we know there’s a (potentially misinformed) whale, you would expect the odds to decline as traders incorporate this new information. Unless of course, the prediction markets are generally reliable and the whale hasn’t influenced them much.

But let’s assume the trader really was spending heavily to “paint the tape” and make a Trump victory appear more likely.

First, it’s not even clear this would benefit Trump. It might cause complacency among his voters, reducing their turnout on Election Day (or even galvanize Harris supporters to show up at the polls). Second, you could simply spend that $30 million on ads in battleground states and meaningfully influence the race. Polymarket remains as of yet a fairly obscure platform so any PR dividend from moving the odds slightly would be extremely hard to quantify.

Plus, as Polymarket itself noted in a detailed blog post Monday rebutting the manipulation claims, “if any user has any reason to believe that trades are being made for anything other than financial reasons, it is easy to adjust the odds to account for this” by excluding the whales’ bets

Why pay attention to prediction markets? When it comes to elections, markets can deliver information faster than pundits, polls, or the press. This trend will continue through the November election, not just on the main presidential contract, but also the state-level contracts. Traders have a financial incentive to surface nonpublic information, so election watchers will be able to see victory probabilities updated in real time for all 50 states, rather than waiting for Fox or CNN to call them. By harnessing trader enthusiasm, markets can and do outperform designated experts.

Polymarket has been outpacing these trusted sources of electoral information for a while. Long before President Joe Biden’s disastrous debate performance, Polymarket traders were signaling nontrivial odds he would drop out while mainstream pundits dismissed the possibility.

Additionally, contracts like the election winner contract give us an arguably more useful informational output than mere poll aggregates. National polls showing Harris up 49% to Trump’s 47% don’t tell you that much about her actual odds of victory: what matters is of course the electoral vote tally and specifically what happens in the battleground states. The highly liquid Polymarket election winner contract compresses all of that information and gives us a simple estimate of each candidate’s odds of winning.

It’s also worth remembering that a 60% to 40% lead on a prediction market is not remotely comparable to a 60% to 40% polling lead. Prediction markets should be thought of as high-beta derivatives of the polls. If Trump were polling at 51% to Harris’ 49% in the national polls, his actual odds of victory would be 90% or more. In prediction markets, 40% to 60%, 45% to 55%, or 60% to 40% odds are all coin-flip territory. So Trump’s rally on Polymarket does not reflect a massive revaluation in his odds of victory. Polymarket whale or not, it’s still an extremely close race.

Note: The views expressed in this column are those of the author and do not necessarily reflect those of CoinDesk, Inc. or its owners and affiliates.