New Times,
New Thinking.

  1. Politics
  2. Polling
5 November 2024

Trump or Harris, who will win? The New Statesman’s US election forecast

Our model predicts a very close contest.

By Ben Walker

Welcome to the New Statesman’s election forecast for the 2024 US election.

Our model has provided near daily projections of vote share estimates for each US state using a variety of poll and demographic data. It gives us a simple forecast: how likely is a Kamala Harris or Donald Trump victory at the polls?

For much of 2024, Joe Biden lagged behind Trump in the polls. Trump unusually found himself ahead of the incumbent - and the televised debate between the pair in June only hardened this dynamic. It put Trump in reach of states like Virginia, New Hampshire and Minnesota.

But Biden's decision to pull out upended it all. The Democrat base rallied behind Harris, though it was unclear how sustainable that current burst of energy was.

Her performance in the debate against Trump served as an introduction of sorts to the voters. And her numbers benefitted. But with no further debates, the novelty wore off. And it was Trump who spent much of October making the headlines and directing the political weather.

The election is likely to come down to the wire in seven states: Pennsylvania, Wisconsin, Michigan, Nevada, Arizona, Georgia, and the wild-card contest, North Carolina. Biden won six of these in 2020, but Trump has eked out leads in all of them for sustained periods this year.

The New Statesman's model is a probability model. Polls have error margins and in some of these key states pollsters have historically performed poorly. But, analysts have learnt from their mistakes too, namely the calamitous miscalculations during the 2016 Hillary Clinton campaign.

Give a gift subscription to the New Statesman this Christmas from just £49

A healthy model ought to factor for all of this. While pollsters claim they have improved their methodology, healthy scepticism wouldn't go amiss. After 2016, for instance, the industry claimed they would alter their samples and surveys. But in 2020 the Trump vote was once again underestimated in key swing states.

In 2020 our model forecast that Trump had a one-in-ten chance of clinging on to the White House. Had our model run in 2016 the numbers would have said Trump had a one-in-three chance. Pollsters claim they have corrected for underestimating the Trump base since these elections. But there's a risk they have also overcorrected.

Here are the probabilities of a Kamala Harris win, state-to-state.

[See also: Vladimir Putin’s enemy within]


Listen to the New Statesman podcast

Content from our partners
How the UK can lead the transition to net zero
We can eliminate cervical cancer
Leveraging Search AI to build a resilient future is mission-critical for the public sector