James Johnson is co-founder of JL Partners. He was the Senior Opinion Research and Strategy Adviser to Theresa May as Prime Minister, 2016-2019.
Sometimes this campaign has felt like the longest: two candidates locked in with no serious primary process for either, epoch-defining events like the attempted assassinations of Donald Trump making hours of news feel like days.
On the other hand, it has sped by: we have had a new candidate at the eleventh hour leave commentators and the public struggling to catch up. The unchanged themes of an angry campaign have made the weeks feel like one amorphous blob, like blended watercolours on a Rorschach test of a nation gone wrong.
There is electoral stasis too. There are forty-one days to go, and the race is still on a knife edge. Our election model at J.L. Partners has a 50.9 percent chance of a Trump win, and a 48.8 percent chance of a Harris win.
You are right to ask yourself in those circumstances: what is the point of the pollsters? All we can appear to tell you is that the race is a coin toss.
Trump, though, does appear to have more states to play with. Our central estimate is that Trump will win the electoral college with a 287 to 251 margin, because he picks up the states of Georgia, Arizona, Nevada and Pennsylvania from the Democrats, and holds North Carolina.
A New York Times/Siena College poll released on Monday showed Trump performing well in three of the ‘sunbelt’ states that make up part of this winning combination. Harris is favoured in Michigan and Wisconsin.
Just because otherwise my job would be rendered useless, I’ll say I think Trump slightly has the edge to win.
Our most recent poll had Harris moving to a one-point lead nationally, the first time she had led in our tracking series with the Daily Mail.
But that is not enough for her to win: because of the distribution of electoral college votes in the states (think the same logic as ‘first past the post’ in the UK), Trump would be in the White House on those numbers. Forty per cent of Trump’s victories in our model are won despite him losing the popular vote.
Some polls have shown Harris with stratospheric leads. Ipsos had a poll out yesterday showing a seven-point lead. But these are often conducted online, a methodology that works well in the UK but boosts Democrats in America where there is a much sharper partisan split on usage and trust of internet surveys.
I don’t think anyone is winning this race by seven points.
It is true that Harris has had a bump from the debate. Every week we ask voters what has changed their minds, and her – and Trump’s – debate performance have dominated for the last fortnight.
This is mostly down to the cacophony over allegations by Trump that Haitian immigrants were “eating the cats, eating the dogs” in Springfield, Ohio. The story – so untrue that even Vivek Ramaswamy had to disown it this week – has not put the focus back on immigration but has reminded voters of the worst parts of the Trump presidency.
A 41-year-old receptionist in Arizona who didn’t vote in 2020 told me one of the main reasons she had changed her mind from voting Trump was because “he’s talking about people eating cats and dogs”. A 40-year-old ticket taker usher who voted Trump in 2020 is not doing so this time “because he is talking crazy now with his accusations”. A 54-year-old Michigan man said he’s not opting for Trump “because he just sounds like a lunatic”.
The irony is that before the debate, a period where Trump enjoyed a boost in the polls, these same voters were warming up to the former president because they felt he had dropped his offensive tone. A skilled labourer from Nevada: “Trump isn’t really all out of turn at this election”. A female 30-year-old freelance writer from Pennsylvania: “He seems to be becoming a kinder person who owns his mistakes”.
The comments since underscore how Donald Trump just cannot seem to help himself. In those circumstances he is probably right to say no to another debate.
He is still doing better with non-white voters than he did in 2016 or 2020. He has continued to make inroads with Hispanic voters angry with the unfairness of the situation at the southern border. Kamala Harris has not got black voters as excited as they were about Barack Obama or even Joe Biden, despite being a Black woman herself. Last week the Teamsters union released a poll showing their voters back Trump by a margin of almost two to one.
If those numbers are echoed amongst blue-collar communities in the Midwest, then Harris is in trouble.
But Harris has done well to shore up the female vote. Her ads blitz on abortion is working, conjuring the dark omens of ‘Project 2025’, a conservative think tank book the Democrats have successfully presented as a dystopian plan for female oppression. She has steadied the ship with younger white women particularly, a group Biden risked losing over his support for Israel’s war in Gaza.
There are still unknowns that warrant further investigation. Around one in fifteen voters are undecided. That generates nothing like the uncertainty prior to British elections, but with the ultra-tight margins those voters could make all the difference. More analysis of that group is needed to see where – and why – they might lean.
We have countless national polls but are lacking well-designed (and expensive) surveys of individual groups such as over-65s and African American voters.
There are still developments to unfold. After the tumults of American politics recently, it would be a foolish man who predicts a quiet October. More prosaic events will matter too. The feel-good factor from the Federal Reserve’s interest rate cut last week might lend Harris a hand, or a pending dockworkers’ strike due to start on October 1st could cause economic chaos under her watch.
The election does not end on election day either: the margins are so close there is a 30 percent chance of a full state recount.
Perhaps the pollsters will all be surprised. After saying how impossibly close it is to call, perhaps one of the candidates will sweep it. Harris because people are just fundamentally sick at the idea of another term of Trump drama, Trump because people vote with their wallets that have felt prices go up dramatically over the last four years.
But I suspect the America of 2024 is not a nation of big, sweeping changes. It is one locked into camps, with only a handful of voters changing their minds in the middle. To be a swing voter is the norm in Britain, in America it is a no man’s land.
Sorry from the pollsters: that essential fact means this thing is just too close to call.