IAMERICA If the presidential election were tomorrow, Donald Trump would choose the Oval Office curtains. The EconomistIt is poll average puts him 2.3 points ahead of Joe Biden nationally. And in the six key states expected to decide the election — Arizona, Georgia, Michigan, Nevada, Pennsylvania and Wisconsin — he leads by an average of 3.8 points. Betting markets view Mr. Trump as a clear favorite. Never in his last two campaigns have his general election polls been this strong. Is it time for the world to prepares for a second Trump presidency?
The elections are still nine months away. Historically, polls taken before the summer of an election year have been poor predictors of outcomes. But no former president has sought to return to power since the advent of modern polling. Views of the ubiquitous Mr. Trump are much firmer than of typical challenger candidates, who at this point in the race are generally still fighting to secure their party’s nomination. As a result, even though Mr. Trump is not yet the presumptive Republican nominee, the current polls between him and Mr. Biden could be particularly informative.
National surveys over the past month have varied widely, from an eight-point lead for Mr. Trump to a six-point lead for Mr. Biden. Polling averages, which dampen the effect of these outliers, suggest that Mr. Trump holds a clear lead. But the polls that establish these averages differ in their methods and degree of rigor. Democrats looking for a glimmer of hope can find comfort in a clear pattern: Pollsters with better accuracy scores show better results for Mr. Biden. On the other hand, their lower-quality counterparts give Mr. Trump the advantage.
Public confidence in polling has weakened following high-profile industry underestimates of support for Mr. Trump in 2016 and 2020 (even though polls taken before the 2018 and 2022 midterm elections were accurate). To reliably estimate pollsters’ accuracy — measured by the magnitude of their historical errors and whether they systematically exaggerate support for a particular party — requires a large sample of surveys across many elections. FiveThirtyEight, a data journalism company, recently updated its ratings of U.S. pollsters. It evaluates them on the combination of their files and their methodological transparency.
Some sounders are systematically more precise than the terrain. But there are many ways to judge quality. The EconomistThe general election polling average of . weights polls based solely on sample size and recency, so that larger, more recent polls contribute a larger share to the overall average. Using this methodology, Mr. Trump leads Mr. Biden by 2.3 points in national polls. That compares to a 0.2 point lead for Mr. Biden in an unweighted average that gives polls from six months ago the same weight as those from last week.
The size of Mr. Trump’s lead varies widely depending on the quality of pollsters, as assessed by FiveThirtyEight (see chart). This early in the election cycle, the highest quality pollsters have only conducted polls sporadically. (One exception is a weekly survey conducted by YouGov, an online pollster, to The Economist.) However, a total of 13 surveys were carried out in 2024 by companies in this group. On average, they show a near tie between Mr. Trump and Mr. Biden.
On the other hand, most of the polls published in January 2024 come from the middle class of pollsters: companies with good but not exceptional results. Polls in these levels (“good” and “decent”) give Mr. Trump a lead of 2.4 points and 1.7 points, respectively. Meanwhile, pollsters with a poor record or without previously released results show Mr. Trump having an average margin over Mr. Biden of about six points.
National polls reflect the general mood and correspond to the popular vote. But thanks to the electoral college system, winning the popular vote does not guarantee electoral victory. In 2000 and 2016, for example, Republican candidates won the presidency despite losing the popular vote.
Yet in recent decades, the Electoral College has generally benefited Republican candidates. If Mr Trump won the popular vote by a six-point margin, he would almost certainly win at least 358 Electoral College votes, giving him the biggest Republican victory since George HW Bush won 426 in 1988. That would put in-game even states that Mr. Biden won comfortably in 2020, as did Maine, Minnesota, New Hampshire, New Mexico and Virginia.
In the eyes of survey watchers who believe all polls are equal, Mr. Trump has opened up a modest but growing lead nationally. But for those who argue that historical pollster accuracy predicts future accuracy, MM. Trump and Biden are at an impasse.■
