Our mission to help you navigate the new normal is fueled by subscribers. To enjoy unlimited access to our journalism, subscribe today.
President Trump stands a minuscule chance of being reelected unless he can carry two states where he now trails in both the polls and the odds on political gaming sites: Pennsylvania and Arizona. “To prevail, he’d need to pull off not one but both,” says Paul Krishnamurty, a professional political gambler and election expert for the prominent U.K. betting venue Betfair. Indeed, closing the substantial gap with Joe Biden in either the Keystone or Grand Canyon state poses a big challenge in itself. More than any other factor, it’s the heroics needed to score a twofer that’s holding Trump’s odds for victory as low as one in five.
For this analysis, I’ll use the odds on PredictIt, America’s only political gaming site, hosting around 100,000 bettors. The PredictIt data is a better measure of where the race stands right now than the polls, because gamblers adjust the candidates’ “prices” instantaneously in response to real-time events. When Trump tested positive for COVID, for example, his odds dropped sharply within minutes. Polls typically reflect what voters were thinking around five days earlier. The bettors also have “skin in the game” because they’re not just expressing an opinion, they’re risking their own cash.
Trump has a lock on 125 electoral votes in 20 states where he enjoys wide leads both in the polls and in the political exchanges, including PredictIt. In both Michigan and Wisconsin, prizes Trump claimed in 2016, his PredictIt odds are 29%, and in Minnesota, he’s at 22%. So chances are overwhelming that all three go blue. According to the Cook Political Report, six states qualify as “toss-ups,” Texas, Florida, Georgia, Ohio, North Carolina, and Iowa, as well as Maine’s one-vote 2nd Congressional District. They fall in that basket because all are close in the RealClearPolitics average of polls. Trump leads by 2.6% in Texas and 0.6% in Ohio, and trails 0.5% in Florida, 0.7% in North Carolina, and 1.4% in Iowa, while Georgia is tied.
PredictIt’s gamblers, however, give Trump a comfortable advantage in five of those states. In both Texas and Ohio his odds are a strong 72%, and in Georgia, Florida, and Iowa, he stands between 56% and 58%, making him a moderate favorite in each. The classic toss-ups are Maine-2 and North Carolina. At noon on Oct. 29, Trump rated 51% in the Tar Heel State, and in Maine-2, he led by a hair at 52%. The roller-coaster shifts in North Carolina, a state Trump carried by almost four points in 2016, must be particularly vexing to his strategists. In mid-October, Biden opened a big lead. Then, Trump went on the march, advancing from 45% to 56% by Oct. 27. In the past two weeks, his momentum has stalled as his chances shrank to the current 50-50.
Although Trump is near even-odds for North Carolina and Maine-2, we’ll assign them to his column, since at least without the former, it’s almost impossible to see how he wins. So if the President captures those two true toss-ups and the battlegrounds where he’s well ahead on PredictIt, he’d add 123 votes to his bedrock 125, bringing the total to 248. That leaves him 22 short of the 270 needed to win. (Amy Walter, national editor of the Cook Political Report, also explores Trump’s possible route to victory if he takes all the “toss-ups,” among other scenarios in which he keeps the White House.)
Trump stands a reasonable chance of winning in only two other states––yes, the stars of our story, Pennsylvania and Arizona. Everywhere else, he’s at 30% or less. Pocketing the Keystone State’s 20 votes would leave him two shy of victory, so he needs Arizona’s 11 as well. He’s a moderate long shot in both. As recently as early October, Trump lagged by around 30% to 70% in Pennsylvania. Then, he went on a tear, hitting 45% on Oct. 27 as it seemed that Biden’s call for an end to oil and gas production was putting the President on track to a lead by Election Day. But the bump didn’t last: By Oct. 29, Trump had retreated to around 40%. In Arizona, the President’s looking better at 45%. Once again, he’s staged a strong comeback since languishing at 30% in early October. But after hitting 48% on Oct. 24, his sprint stalled in the homestretch.
Keep in mind that to win, Trump must also hold North Carolina, a state that’s in play big-time. It’s a good bet that what’s dictating Trump’s overall odds is his chance of taking both the Rust and Sunbelt stalwarts that are still within reach. Let’s recall some simple statistics. The chances of tossing a coin and getting heads is of course one in two or 0.5. The odds of flipping two heads in a row is 0.5 on the first multiplied by 0.5 for the second, or 0.25, meaning one in four. On PredictIt, Trump’s odds of taking Pennsylvania are 0.4, and chances in Arizona are 0.45. Hence, his odds of taking both stand at something like 0.4 multiplied by 0.45, or 0.18. That’s roughly one in five.
We need to add a caveat. “It’s hard to price multiple states,” says Krishnamurty. “If Trump wins Pennsylvania, his chances of winning Arizona would be much higher than what’s shown on the betting sites. It would show that the national polls are wrong.” In other words, Trump always had a much better chance of winning in both Pennsylvania and Arizona than the surveys were showing. Still, the one in five we get from handicapping the double-win are the same 0.2 betting odds that Krishnamurty assigns to a Trump win. It’s also the current figure on the forecasting site created by Northwestern University data scientist Thomas Miller. That’s down from one in four on Miller’s platform on Oct. 28.
The most probable reason for Trump’s pullback from Wednesday to Thursday: his flattening momentum in Pennsylvania. Once again, if he can’t win there, capturing Arizona won’t be enough. And the smart bettors know it.