Government to Ban All US Election Betting

The ban would greatly slow, but not end, global progress on prediction markets

This afternoon, the government bureaucrats at the CFTC announced that they plan to ban all election betting (aka “prediction markets on elections”, aka “event contracts”) in the United States. They will also ban trading on events in general — for example, on who will win an Oscar.

I’m close to this issue, having reported on prediction markets since 2007, and having created ElectionBettingOdds.com.

I am strongly in favor of allowing prediction markets, because markets have a great track record of predicting the future, which is helpful for decision-making. There are smart modelers out there (Nate Silver being the best) but supplementing his estimates with bettors’ predictions still increases forecast accuracy.

The bureaucrats’ decision, if finalized and upheld, would have a big effect on the election betting space, as it would force the major exchange for Americans, PredictIt, to shut down.¹ Depending on the details of the regulation, it might also impact Polymarket and Insight Prediction, which are based in the US, but which already block people with US IP addresses from trading.

Democratic regulators support the ban, citing “the sanctity of elections,” while Republicans oppose it

Among the general public (and also among intellectuals and politicians) nobody on either side of the aisle seems to be clamoring for a ban on prediction markets.

But among regulators, it is a partisan question.

The WSJ reports:

CFTC commissioners voted 3-2 to release the proposed regulation for public review, with the agency’s three Democratic commissioners voting in favor and its two Republicans dissenting. The proposal won’t take effect until commissioners approve a final version, a vote that is likely many months away.

The Democrats stressed the risk that political-event contracts could pose to election integrity, which has become a top concern as the U.S. awaits a rematch between President Biden and Donald Trump later this year.

“Never before has the sanctity of elections been so critical or so under threat,” Christy Goldsmith Romero, a Democratic commissioner, said ahead of the vote. “The CFTC should not allow products in our markets with an unacceptable risk of unchecked abuse and manipulation that could threaten the sanctity of elections, thereby threatening democracy and national security.”

Summer Mersinger, a Republican commissioner, called the proposal “grossly overbroad” and an overreach of the CFTC’s authority.

Is it possible that the regulators have a point?

Why the “election interference” case is weak

I suppose it’s conceivable that someone could put a million dollars on Candidate X, and then … do what? Gather and forge 5,000 mail-in ballots to try to tip the election? Or that an election counter has $1,000 on Candidate Y, and lets a ballot for the opponent slip into the trash can. Or one could imagine a TV commentator who bets $50k on a candidate and then promotes that candidate, even though he knows the candidate is terrible, just to protect his bet.

Those all seem pretty far-fetched — the thing with election manipulation is that even the most powerful individuals are rarely in a position to tip an election. It’s much harder to flip an election than a sports match, because of the number of people involved.

And regarding the hypothetical TV commentator above, it makes little sense, because in an open market, you can sell and reverse your position in an instant (sites like PredictIt don’t use bookies; instead they’re like stock markets, where you buy and sell shares in candidates winning.)

It seems far less biasing than things like ideological extremism, or expected patronage (someone expecting a government job, or a business/industry that expects favorable treatment.)

So I don’t see the “sanctity of the ballot” point. Seems like the regulators just don’t want to have to bother with innovation.

Why the “we can’t be a gambling cop” excuse is weak

The other reason the bureaucrats give for banning prediction markets is that they don’t have the resources to bother with being a “gambling cop.” That seems a dumb excuse, since they don’t need to police gambling on elections if they just make it legal, and hold political trading sites to the same standards they hold stock exchanges to.

Why the regulators’ “election manipulation” excuse is weak

As far as market manipulation, market manipulators essentially subsidize smart traders. For example, imagine you have a billionaire who wants to inflate RKF’s odds to 20%. He can do that momentarily, but the longer he tries to hold it there, the more politically-savvy traders will see it and take free money from him. Let’s say he’s willing to sink a billion into it. That just makes it worth the while of whales — other billionaires, investment funds, etc — to take his money, too. Paradoxically, the harder one fights to manipulate a market, the more others are incentivized to make the market accurate. Countries face the same difficulty when trying to artificially manipulate their foreign exchange rates, by the way, and often resort to the force of law to get anywhere on it.

There’s no evidence of manipulation in hundreds of millions traded so far. If there were a lot of manipulation, we wouldn’t see the kind of accuracy we do.

Additionally, even if it worked, the social costs of market manipulation seems low. Should government also ban people from commissioning biased polls for the purpose of manipulating public expectations of an election? That happens, but it’s not exactly destroying our society.

Will election betting actually be banned?

Left to their own devices, the current CFTC will almost certainly ban it

Now that regulators have announced their proposed regulation, there will be a 60-day public comment period.

The comments are unlikely to sway the commissioners, as this is an issue that the CFTC has been considering for a long time, and which has already received lots of media coverage (by regulatory standards.) The commissioners should know where they stand on the issue by now.

The 2024 election result might change things

It’s unclear how fast the CFTC bureaucrats will move this forward. If Trump were to win the election, he would nominate a replacement for a Democratic CFTC commissioner in April 2025 (I won’t go into the details of the complex CFTC appointment schedule, but the important thing is that the seat would open up in April 2025.)

The ban may already be in place by then. To get a sense of how fast CFTC regulations get finalized, I looked to the time it took the CFTC to finalize other recent rules after they proposed them:

• “Swap Confirmation Requirements for Swap Execution Facilities”

  Proposed 8/25/2023, finalized 5/1/2024 (9 months)

• “17 CFR Parts 39 and 140 Reporting and Information Requirements for Derivatives Clearing Organizations”

  Proposed 12/15/2022, finalized 8/8/2023 (8 months)

• “17 CFR Part 39 Governance Requirements for Derivatives Clearing Organizations”

  Proposed 8/11/2022, finalized 7/13/2023 (11 months)

Based on those, the betting ban would be enacted around February 2025. It’s basically up to the regulators whether it’ll faster or slower than that.

It could then be rolled back by a future CFTC, although reverting the ban could take another year.

The regulation may be overturned in court

Another possibility for preserving election betting in the US may be via a legal challenge. The CFTC plans to ban election betting by defining election and event trading as “gaming”, which seems a bit questionable on its face, and which the Republican CFTC commissioners call “overreach.”

The WSJ reports:

CFTC regulations implemented after the 2010 Dodd-Frank Act prohibit event contracts that involve terrorism, assassination, war, gaming or activities banned by federal or state law.

But the regulations didn’t define “gaming,” which has led to disputes over whether the existing prohibition applies to contracts based on sports matches or political elections. Friday’s proposal spells out a definition of gaming that covers wagers on elections, sports or awards contests, meaning that such contracts would be barred.

If courts put a hold on the regulation and ultimately strike it down, that could prevent the ban.

If Trump were to win, and if there is a tough fight in court, then his administration might choose not to fight the case, which could increase the odds of the regulation being struck down.

But one would have to talk to very experienced lawyers to get a sense of how courts would rule on this; clearly the government lawyers think they’re on solid ground.

Election betting still has a chance in the US

Based on all the above, will a ban on election betting become the long-run norm in the US? Let’s imagine that there’s a 50% chance that the regulation gets upheld in court, and separately a 60% chance that either Trump loses, or that his CFTC appointment fails to overturn or block the regulation.

If those odds are right, that would mean the odds that the ban will go into effect, and stick around for years, would be around 30%.

Obviously that’s an extremely rough estimate. Thanks to election betting markets, I feel pretty confident in the Trump odds, but what the courts will rule is anyone’s guess.

Play-money trading shows promise

If it is banned, it’s worth noting that sites like Manifold and Metaculus have shown that play-money markets can also be pretty accurate. It reminds me of how Wikipedia works — just as some people write and edit articles as an unpaid hobby, and to help society and to feel smart, others do the same for predictions. I think there are long-term and scaling reasons to prefer real-money markets² but the play-money markets are certainly better than nothing, and still use the “wisdom of the crowd” — just without monetary incentives.

I’ve posted some play money markets on what will happen with the betting regulation. The “trading” is thin, but here are the estimates right now:

If you feel you have insights on this, feel free to click on one of the graph headers to predict.

Even if trading is banned in the US, it will continue in places like the UK.

A ban on American trading would be bad for the growth of prediction markets, and set it back several years at least, but international trading will continue on sports-focused sites in the UK like Betfair and Smarkets.

Sites like those mean that ElectionBettingOdds.com will always have something to track.

“Government to Ban All US Election Betting,” 2024, by ChatGPT-4 and Maxim Lott

1. The CFTC perhaps hinted at their new ban decision 1.5 years ago, when it ordered PredictIt to shut down, revoking a “no action” letter it had issued years earlier, which had said it didn’t plan to take regulatory action against PredictIt (which was billed as an academic project, as it had the backing of a university in New Zealand.) PredictIt is still operating, because it’s suing the CFTC, saying the agency can’t just revoke their letter — but it seems this regulation would make that point moot.

2. Play-money markets are more dependent on social dynamics; if they ever became truly influential, and cited everywhere in the news, they would attract mediocre, ideologically-motivated traders — who would reduce the quality of predictions. I suspect that real-money markets are more immune to that, because with real money, there becomes an incentive for smart, less-biased people to put even more money in and correct the odds.