To interpret the early voting data in the US election – which can be viewed here – we have to work through the electorate’s habits in recent years. In 2020 there was an overwhelming number of early votes cast by registered Democrats. Meanwhile, Donald Trump actively dissuaded his base from early voting.
This time Trump is not discrediting the validity of early votes. But at the moment they are still leaning Democrat, as they did in 2012 and 2016. You can see on the table below how early voting is a consistently Democrat habit.
Take a look at Nevada – where Democrats were over-represented in early votes by 6 points in both 2012 and 2016. In other states, such as Pennsylvania in 2020, registered Democrats ended up over-represented in early votes by a significant 25 points. The independents are those not registered as either a Republican or a Democrat. A lot depends on how their votes get split up between the two major parties.
Let’s do some modelling. Let’s assume the early votes are going to skew expectations as much as they did in 2012 and 2016, though not as much as they did in 2020.
Currently, the early vote in Nevada looks like it’s at 35 per cent registered Democrats, 40 per cent Republicans and 25 per cent independents. If we weight that to account for previous “conventional error” – ie, how the election panned out in 2012 and 2016 – we might expect a result that looks like a narrow win for Kamala Harris. But if early voting is as unusually unrepresentative as it was in 2020 then Trump could win the state by 52 per cent.
In Arizona, the numbers suggest Harris will lose the state by a decimal point, 49.1 per cent to 49.2 per cent. But if the early votes follow a similar pattern as in 2020, Harris will lose by 7 points. In North Carolina, Harris should lose the state to Trump by 46.1 per cent to 51.9 per cent under a 2012/2016 framework. That’s worse than Joe Biden’s showing of 48.6 per cent to 49.9 per cent.
These numbers are based on partial returns and two big assumptions: that the competition doesn’t change drastically, and that the skew in early votes looks like 2012 more than 2020.
If the early votes are as unrepresentative as they were in 2020, then Donald Trump will be walking this election – winning Pennsylvania, perhaps, with more than 50 per cent of the vote. However, Trump is no longer dissuading early voting, and therefore we should be sceptical of a 2020-style skew.
It is worth noting that registered Democrats’ early-vote turnout is a long way off what it was in 2020 – it’s worth noting because it points to the idea that Democrats with a history of voting early are still yet to turn out. It wouldn’t be a reach to say the composition of early voters could get more Democratic closer to election day, not less.
This relatively low turnout could also be read as an indicator of low enthusiasm from registered Democrats, which would point to a clear Trump victory. On the other hand, if it just represents a recalibration in how registered Democrats choose to vote then the narrow losses for Harris in Arizona and Nevada (per my own modelling) might just become narrow wins. And in the rust belt, amid news from some pollsters that she is improving on Biden among white voters, then she could still win.
[See also: Joe Rogan could decide the US election]
Content from our partners
Related
