Two months to go

Just two months now until the United States presidential election, and with both parties’ conventions over it’s a good time to have another look at how the contest seems to be placed.

First, the polls. Democrat challenger Joe Biden holds a consistent but far from overwhelming lead over Republican president Donald Trump. FiveThirtyEight’s poll average today puts him 7.3 points ahead. If reflected in a uniform swing, that would win him an additional seven states and a majority of 162 in the electoral college.

The biggest shift in the polls came about three months ago. Until that point, Biden for a long time had mostly been leading by between five and six points. But by the middle of June that was up to 9.3 points. Since then it has wobbled around, but never fallen below seven points.

The difference might not sound like much, but it’s important to remember that Hillary Clinton led by about four points in the polls on the eve of the 2016 election. In fact she won by only two points, which was sufficiently close for Trump to narrowly snatch a majority in the electoral college. So if Biden’s lead gets back down to around five points, he’s in danger territory. But if he’s still leading by more than seven points going into the election, he will almost certainly win.

The picture from expert forecasts is much the same. FiveThirtyEight’s model gives Biden a 70% chance of winning. That’s also quite stable – it’s varied between 67% and 76% over the last two months. RealClearPolitics’ state-by-state forecast has Biden with a majority of 166 in the electoral college, the same as the last time we looked.

One indicator, however, has shifted very noticeably: the betting market. A month ago, going by Maxim Lott and John Stossel’s aggregation, it put Biden’s chance of victory at 59.2%; that figure is now down to 51.8%. Sportsbet this morning actually has Trump (very narrowly) as favorite.

So it looks as if the punters are relying on something other than the polls. The obvious culprit is the recent rioting, particularly in Oregon and Wisconsin, which gives Trump the opportunity to capitalise on a law and order campaign. So far, that’s not showing up in the polls, but the punters seem to expect that it will.

A month ago I said that “there’s certainly still time in which [Biden’s lead] could change. But something will have to happen to change it.” The riots and an associated white backlash could be that something, but at this point there’s very little evidence for it.

It’s also interesting to note that the shift in the betting market doesn’t show up much at state level. Sportsbet still has Biden as favorite in the Trump-held states of Michigan, Pennsylvania, Wisconsin, Florida and Arizona. The last two are extremely close, but the other three – which would be enough for victory on their own – are very clear. Lott and Stossel say the same, but with a stronger Democrat margin in Arizona.

It’s easy to get distracted by the different permutations of possible support in the electoral college, but there are two key things to keep in mind. First, if the overall margin is big enough, the college won’t matter: a margin of seven or eight points would render worries about individual states irrelevant. We might not know exactly where the wins will come from, but they’ll come from somewhere.

Second, in a close election, Trump has only a narrow path to victory. He has to hold on to at least one of his three closest states – Michigan, Pennsylvania and Wisconsin – plus the big states above them on the pendulum, including Florida and Texas. If he wins more than one of the close three, or if he picks up Minnesota from the Democrats, he can potentially drop another medium-sized state, such as Arizona or North Carolina. But his margin for error is small.

Biden, by contrast, has more options. That won’t help him much if the polling turns seriously bad, but it means Trump has a lot more territory to defend. It can’t simply be assumed that the electoral college will always work in the Republicans’ favor. Trump, for example, could get big swings in his favor in the midwest, holding all three of the marginals and picking up Minnesota, but he will still lose the election if he loses Florida, Arizona and North Carolina.

It seems to me as if punters have been spooked by the 2016 result and are assuming that Trump will be able to pull a rabbit out of the hat instead of actually looking at the mathematics of the electoral college. But with two months to go there’s still time for them to be proven right.

4 thoughts on “Two months to go

  1. > It can’t simply be assumed that the electoral college will always work in the Republicans’ favor.

    I think the Republican advantage is pretty much baked in as far as this election is concerned. The situation you outline is a bit far-fetched and would probably involve Biden winning the popular vote handsomely anyway.

    That advantage seems to be around 2 to 3 percentage points. In a recent tweet, Nate Silver gave Biden a 6% chance of winning the EC with a 0-1 point PV win and a 22% chance with 1-2 point PV win. (He didn’t say what Biden’s chances were in the event of a PV loss, but they’re obviously pretty minute.)

    Thankfully Biden’s present polling lead is some way above that 2-3 point zone.


    1. Yes, I saw those figures from Silver; much as I respect his work, I find some of them (particularly that 6% number for a 0-1 point lead) hard to believe, Just as I suggested last time that his state polling models might be exaggerating the uniformity of swing, I think here he might be exaggerating the extent to which the pattern will resemble 2016. But we’ll see in due course.


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 )

Google photo

You are commenting using your Google 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.