State of Play

Assessing the US election with two months to go

I’ve hung back from writing about the race for President until now because of the extraordinary amount of flux over the summer.

Joe Biden’s career-ending debate performance was only 73 days ago. Since then we’ve seen: an assassination attempt against Donald Trump; Biden replaced by Kamala Harris; the selection of JD Vance and Tim Walz as VP candidates; both conventions; and Robert Kennedy Junior (who was polling at 15% two months ago) dropping out of the race and endorsing Trump. That’s a lot of moving parts.

With two months to go things feel like they’re finally settling down, though of course there is still plenty of opportunity for drama, both at planned events, like Tuesday’s debate, or of the unplanned “Hillary’s emails” type.

The broad shape of the race is largely undisputed by analysts. At the point Biden dropped out Trump had become the strong favourite, leading by around 4 points in national polls and by more in the key swing states. When Harris took over there was an immediate transformation in the race and she has continued to improve her position since, reaching a national lead of 3 points after her convention which has, so far, been sustained.

Despite this big improvement, and the Democrats’ understandably buoyant mood, all this shift has done is take us back to a toss-up race. The various election models, of which more later, are mostly hovering around the 55/45 mark (though not all in the same direction). This is because the electoral college system now has a strong bias towards the Republicans. Indeed since 1988 – 36 years ago – the GOP have only won the popular vote once (in 2004) but have won three elections (2000, 2004 and 2016).

Harris probably needs to be somewhere between 2-4% ahead nationally to win a bare majority of electoral college votes. And the key swing states are largely the same as last time. Due to demographic shifts Florida has likely slipped out of the Democrats reach now, even though they won it as recently as 2012. And Ohio also seems securely in the GOP column, at least when it comes to Presidential elections.

But otherwise the list – with associated electoral college votes – looks very familiar: Arizona (11), Nevada (6), Georgia (16), North Carolina (16), Michigan (15), Wisconsin (10) and Pennsylvania (19).

In each of these states the polls are within the margin of error, with Trump slightly ahead in North Carolina and Arizona, Georgia a dead heat, and Harris slightly ahead in the others. Simply put: Harris needs 44 votes from that list to win. Wisconsin, Michigan and Pennsylvania would be (just about) enough. The latter is critical, being the most likely tipping point state, which is why many thought Harris would pick governor Josh Shapiro as her running mate. The last few polls in that state have it extremely close.

It's hard for Brits to fathom how it can be so close. It made some sense for Biden to be struggling, given his obvious infirmities, even against a candidate as flawed as Trump. But Harris is (relatively) youthful and articulate. Even if one doesn’t agree with her policies they’re well within the mainstream and she represents no threat to America’s international reputation. Trump meanwhile is, well, Trump. A man who still regularly implies his opponents are trying to rig the ballot; who encouraged an attack on the Capitol to which seven deaths have been linked; whose commitment to NATO is shaky at best; who seems openly dismissive of the rule of law and regularly “jokes” about dictatorship.

But as I wrote back in 2022 – party identification in the US works very differently to the UK, tying in numerous other aspects of people’s identity including religion, race and geography. I quoted Lilliana Mason from her book “Uncivil Agreement: How Politics became our Identity”:

“The American electorate has sorted itself into two increasingly homogenous parties, with a variety of social, economic, geographic, and ideological cleavages falling in line with the partisan divide. This  creates two megaparties, with each party representing not only policy positions but also an increasing list of other social cleavages. Parties…draw convenient battle lines between an array of social groups….A single group identity can have powerful effects, but multiple identities all playing for the same team can lead to a very deep social and even cultural divide.”

Which is not to say there’s no voter churn in the US – there is – but it’s a lot lower than here. Throughout the dramas of the summer, though the Democrats have gone down and up as they changed candidates, Trump’s numbers have stayed remarkably static. His supporters simply aren’t going to shift, though equally he’s unlikely to win over many who aren’t already on board.

So what will decide such a close election? To try and answer that I’m going to dig deeper into some of the models and the polling than underpins them. Then I’ll finish with a look at the senate and house races, which will also prove critical to the effectiveness of whoever wins the Presidency.