Los Angeles Mayor Karen Bass spent last week touting a poll released on Thursday from an unknown company called Median Strategies showing her leading challenger Nithya Raman in her re-election campaign by double digits and winning 85 percent of third-place finisher Spencer Pratt’s voters. The poll was a rare bit of good news for Bass after a tough summer that featured a statistical rise in homelessness and an adviser going to work for a reviled logistics firm responsible for a disastrous warehouse fire in Boyle Heights.
The only problem with the poll is that it was fake.
Persistent Los Angeles Times reporter Noah Goldberg got Median Strategies to admit that the Bass-Raman poll was bogus and “should not be treated as genuine polling data.” Indeed, all three Median Strategies polls, including one showing Wisconsin gubernatorial candidate Francesca Hong up in her primary by 23 points (she lost) and another showing Nevada Gov. Joe Lombardo up in his general-election campaign by five, were not backed up by any numbers.
On its website, Median Strategies would only say that it was engaged in a “short-term social experiment” to see how polling data could “spread through the political information ecosystem.” But the financial disclosure that Median added was much more interesting. It insisted that nobody involved with the project received any “financial benefit” from the scheme, and it explicitly cited prediction markets.
“No individual involved held or traded any position in an election prediction market, betting market, or other financial instrument tied to the races covered by Median Strategies. The project was not undertaken for financial gain or to influence the value of any financial or prediction-market position,” the note reads. Median Strategies could not be reached for comment.
Of course, there’s no reason to believe an entity whose only public action involves lying. Because there is literally no other motivation to introduce fake polling into the political bloodstream other than to game a prediction market. And while this one may not have made any money, the point may simply have been to offer a road map to any political actor using invented information to make money from elections.
In fact, the strategy was not even particularly innovative. It has been used for decades, primarily in thinly traded “penny stock” markets, where information of any kind is highly sought. The term of art is a “pump-and-dump” scheme.
Here’s the short version: A promoter touts some inside information about the stock, either by directly cold-calling prospective investors, sending voluminous email spam, dropping it into internet forums or social media, or publishing a fake press release or misleading “investment research.” The information isn’t true, but if it works it generates enough interest to get movement on the stock, which the promoter has already purchased. Having bought low, the promoter sells high and makes off with the profits, and as the scheme collapses when the truth is revealed, all the investors who piled in lose their shirts.
An election prediction market has the same character as a stock. The markets are too new and too numerous to be heavily traded, and in head-to-head matchups, new developments can cause the market to tilt in favor of one candidate. The poll takes the place of the inside information, though you could imagine other things, like fake news about fundraising data or outside spending, erroneous reports about a misstep on the campaign trail, or whatever, can move numbers.
The Bass poll worked perfectly for this purpose. The California Post, Rupert Murdoch’s attempt to penetrate the Golden State with right-leaning news, enthusiastically ran with the poll because of the inference that Pratt voters were kingmakers. (That web URL is still live on the California Post website, but it’s been gutted and replaced with a story about Bass demanding legal action.)
In a classic pump-and-dump, one story like that is all it takes for a market to move. The perpetrators already have their investment and can sit back, watch it rise, and sell when they want. There are resemblances between that and what happened in L.A. mayoral markets around the time of the poll release.
The Post story was published on August 13 at 10:53 p.m. Eastern time. On prediction market Kalshi, a contract on Bass to win the mayor’s race was trading at 64.3 cents on the dollar before the poll drop and 65.9 cents shortly thereafter, peaking at 67.2 cents before falling once the polling fiasco was revealed. The equivalent contract on Polymarket, a Kalshi competitor, was even more pronounced. The market was hovering at 61 cents the day before the poll release. You can see some action around that time, including one $6,000 bet on Bass. The market for Bass peaked on August 17 at 5:30 a.m., before the L.A. Times exposure, at 70 cents.
The Associated Press cited 20 separate Bass trades on Polymarket within 15 minutes of Median Strategies releasing their fake poll. These can be viewed at the market site.
A smoking gun of a whale trade on Bass followed by a sell-off is hard to discern, but on prediction markets, where everything is anonymous and unlimited new traders can be conjured up, that can be hard to uncover. If nothing else, Median Strategies proved the concept that a pump-and-dump scheme is easily transferrable to election prediction markets, which have the same mix of thin trading, high emotion, and easily led traders as penny stocks. That’s valuable information for would-be scam artists.
“I think this falls into the category of hard to prove but very hard to think it’s anything else,” said Corey Frayer, a former financial regulatory staffer in Congress now with the Consumer Federation of America.
Pump-and-dump schemes, incidentally, are securities fraud, and those caught setting them up can get prison time. Jordan Belfort, the famed Wolf of Wall Street, was convicted for running pump-and-dump schemes. But those cases, rooted in investor protection, are typically enforced out of the Securities and Exchange Commission. Prediction markets have been adopted by the Commodity Futures Trading Commission, which has a market integrity mandate.
Certainly, manipulating markets represents a threat to market integrity. But that slight change in focus, the generally outgunned nature of the CFTC, and its current role as an evangelist and indeed the rescuer of prediction markets from state regulation put it in a bad position to uncover and refer fraud of this nature for prosecution. Nearly all the prediction market manipulation we have heard about is a result of self-regulation from the companies, which could be selective.
What the Median Strategies saga reveals is the ease with which pump-and-dump schemes can be carried out in election markets. That is supposed to be an acceptable price to pay for learning the future. But it’s becoming very clear that, as applied to elections, prediction markets don’t know a damn thing and are of no utility. Last night’s Florida Senate race is a good example; Democrat Angie Nixon never rose above 11 cents on Kalshi until she won and paid out $1. One reason? There weren’t any polls.
