Into the final month

With just one month left to go to the United States presidential election, it’s time for another look at where the contest stands. You can read earlier previews in this series here, here and here.

It’s been an eventful week, with the chaotic first debate between the two candidates and then the hospitalisation of president Donald Trump with Covid-19. No-one knows what might happen next. But that makes it all the more important to get a good grasp on the basics of what the polls are saying and how the rivals are currently positioned.

Last month I said that Democrat Joe Biden enjoyed a lead in the polls that since the middle of June “has wobbled around, but never fallen below seven points.” Over the following four weeks it showed even more consistency, staying between 6.6 and 7.7 points (I’m using FiveThirtyEight‘s aggregator). Then last week, after the debate, it ticked upwards to 8.2; as of this morning it’s sitting on an even eight points.

Over that time, the movement towards Trump that we noted in the betting market also reversed itself – first gradually then, after the debate, very sharply. Many agencies have suspended presidential betting since Trump’s diagnosis, but for what it’s worth Maxim Lott and John Stossel’s aggregation of the odds now has Biden as a 58.1% chance to Trump’s 30.8%.

As of today, FiveThirtyEight’s forecasting model gives Biden 81 chances in 100 of winning, as against Trump’s 19. His rise over the last month has been steady, from a low of 67 at the end of August.

The state-by-state story is a bit more complicated. Two months ago I presented a table of the polling in marginal states; here’s the current version, again using FiveThirtyEight’s poll aggregates:

State2016 winnerMarginNow leadingMarginSwing to Dem
Virginia (13)DEM2.8%DEM6.0%3.2%
Colorado (9)DEM2.7%DEM6.3%3.6%
Nevada (6)DEM1.3%DEM3.2%1.9%
Minnesota (10)DEM0.8%DEM4.9%4.1%
New Hampshire (4)DEM0.2%DEM4.9%4.7%
Michigan (16)REP0.1%DEM3.7%3.8%
Pennsylvania (20)REP0.4%DEM3.2%3.6%
Wisconsin (10)REP0.4%DEM3.5%3.9%
Florida (29)REP0.6%DEM1.6%2.2%
Arizona (11)REP1.9%DEM1.9%3.8%
North Carolina (15)REP1.9%DEM0.7%2.6%
Georgia (16)REP2.7%DEM0.3%3.0%
Ohio (18)REP4.3%DEM0.3%4.6%
Texas (38)REP4.7%REP1.2%3.5%
Iowa (6)REP5.1%REP0.7%4.4%
South Carolina (9)REP7.5%REP3.0%4.5%

(Note that these are margins expressed in the Australian fashion, showing required swing, so that for example a six point lead is a 3% margin. Numbers in brackets are votes in the electoral college.)

There’s been very little change. Two months ago the Democrats were narrowly ahead in Texas; they’ve now lost that lead, but Ohio has shifted the other way. The range of swing is a bit larger than it was, but it’s still very narrow. The median swing among the 16 is 3.7%, down fractionally from August’s 4.05%. There’s also nothing much of a regional pattern, although over the last couple of months the Democrat vote has generally been holding up slightly better in the midwest than in the south.

But that consistency is enormously more significant now than it was two months ago. Then, there was plenty of time for things to happen, and a lot of people who could be expected to be still making up their minds. With only a month to go the situation is very different. Polling that is this stable suggests an electorate that is pretty much set in its views.

Some at this point may recall last year’s federal election in Australia, where an unusual degree of stability in the last few weeks of polling was a sign that something was wrong: that the pollsters were “herding”, or co-ordinating their results, and therefore multiplying their likelihood of error. As it turned out, they were overstating Labor’s vote by about three points. Could the consistency of the American polls be a similar warning sign?

It’s not impossible, but the cases are very different. The stability I’m describing here is of polling averages, not individual polls: the polls themselves seem to be still bouncing around about as much as you would expect, although less so as time passes. The polling industry in the US is vastly bigger than in Australia; by this point there are dozens of polls being released every week, not the two or three we might get here.

So the polls deserve to be taken seriously. That doesn’t mean Biden’s lead is unassailable – as I’ve said before, anything less than about a five point lead in the polls puts him in danger territory. And the last week has thrown more variables into the mix. It’s not impossible that Trump’s illness will win him a sympathy vote that could swing some of the states that he needs.

But time is fast running out for the president’s campaign. He needs to make up a lot of ground, and the road to victory is looking steeper and narrower every day.

3 thoughts on “Into the final month

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.