Six months out – a look at the pendulum

Six months from today (yesterday, their time) the United States goes to the polls to elect a president: a contest that, barring something utterly unexpected, will pit incumbent Republican Donald Trump against Democrat challenger Joe Biden. Current polling shows Biden in the lead, but not overwhelmingly so.

As the 2016 election demonstrated, however, the result of a US election is determined not by the majority of the popular vote but by the vote of the electoral college. Voters may mark ballots with the names of the presidential candidates, but they are actually electing, in their several states, members of the electoral college.

To residents of other presidential states this is highly mysterious, but it should be familiar to those from countries, like Australia, with parliamentary systems. Trump’s achievement three and a half years ago of winning despite being outvoted was no different to John Howard’s in 1998: each managed to collect his smaller number of votes in just the right places to translate into a majority of seats.

As in Australia, votes gained in places that are already safe for one or other party are useless: it does the Democrats no good to pile up bigger majorities in California or New York, nor the Republicans in Kansas or Wyoming. Only the marginals – seats in Australia, states in the US – really matter.

So the presidential election can be modelled with a pendulum, just like an Australian election, but with the added refinement that the electorates (mostly states, but two states vote by congressional district as well as at large) have varying numbers of seats. A given swing, therefore, will move electoral college seats in blocks, not just one at a time.

What, then, does the pendulum tell us about November? In 2016 the Democrats won 51.1% of the two-party vote and carried 20 states, plus the District of Columbia and one of Maine’s two districts, for a total of 232 electoral college seats. The Republicans, with their 48.9%, carried 30 states, plus the other Maine district and all three of Nebraska’s districts, for a total of 306 seats.

Despite the media often giving the impression that the last election represented a dramatic movement, the swing involved was very small: just 0.9% to the Republicans, the smallest swing since the nineteenth century. But Trump was very fortunate (or skilful, if you prefer) in getting that swing in the states where it could do the most good.

To see this, consider just the marginal states – those requiring swings of less than 6% to change hands. There are seven such states on the Democrat side of the pendulum and ten on the Republican side. But the Republican ones are a lot bigger: they comprise a total of 181 seats as against only 49 for the Democrats.

Here are the Democrat marginals, with the number of seats in brackets and the margins*:

New Mexico (5) 4.7%
Virginia (13) 2.8%
Colorado (9) 2.7%
Maine at-large (2) 1.6%
Nevada (6) 1.3%
Minnesota (10) 0.8%
New Hampshire (4) 0.2%

Colorado, Virginia and (to a lesser extent) New Mexico have been shifting more towards the Democrats over the last few elections, so unless there’s a big swing against him Biden is unlikely to be troubled there. The other four will be targets for Trump, but they are too small to make much difference unless the election is very close.

The Republican side of the pendulum is much more interesting. Here are Trump’s marginals:

Maine 2nd District (1) 5.6%
Iowa (6) 5.1%
Texas (38) 4.7%
Ohio (18) 4.3%
Georgia (16) 2.7%
North Carolina (15) 1.9%
Arizona (11) 1.9%
Nebraska 2nd District (1) 1.2%
Florida (29) 0.6%
Wisconsin (10) 0.4%
Pennsylvania (20) 0.4%
Michigan (16) 0.1%

Biden needs an extra 38 seats for a majority, so to win it on a uniform swing he needs to win those last three very close states, Michigan, Pennsylvania and Wisconsin – a swing of just 0.4%. Those were the three states that most attention focused on in the 2016 post-mortems; all recorded above-average swings to Trump, and all are seen as traditional blue-collar industrial states.

But because there are so many Republican marginals, the Democrats have plenty of additional options. Texas alone would be enough for a majority, even without any of the three super-marginals; so would Florida plus Arizona, or Arizona plus North Carolina plus Georgia.

Hence the difference sometimes talked about between a sunbelt strategy and a midwestern strategy: either could provide a path to victory without any gains (or even with small losses, such as Minnesota or Nevada) in the other. But in reality Biden’s campaign is most unlikely to be a pure version of either. It will target seats wherever they seem available, forcing Trump to commit resources in different parts of the country.

Trump’s path to victory, on the other hand, is much narrower, as it was in 2016. He has to do two things: first, keep hold of at least one of the key three, Michigan, Pennsylvania and Wisconsin; and second, make sure there are no losses among the big states further up the pendulum.

It’s certainly possible – he did it last time, after all. But there’s not much of a margin for error. You can try it out yourself with the map at 270toWin. And demographic change seems to be working against him: while big states like Georgia and Texas are coming onto his opponents’ radar, there’s very little in the way of compensating gains the other way.

As in a parliamentary system, unless there’s a gross malapportionment (which there isn’t in the US, although there’s a small bias towards smaller states), targeting your vote can only get you so far. A candidate can be lucky and win with 49% or even 48% of the vote, but a big swing will inevitably bring the necessary seats with it – possibly in quite unexpected places.

That said, as partisan divisions become more rigid, landslides are getting more and more unusual. No candidate has won with more than 55% of the vote since Ronald Reagan in 1984. But with any luck, this time the result will at least be clear enough for the US not to embarrass itself by again anointing the loser as president.


* This is the way margins are expressed in Australia, as the amount of uniform swing required to change hands. Note that American practice is different, usually quoting the gap between the parties, so that, for example, a 53-47 result would be described as a margin of six points, whereas in Australia it would be 3%. Note also that I have factored out the minor party vote, whereas American figures usually quote the raw numbers.

10 thoughts on “Six months out – a look at the pendulum

  1. So the trick to converting the US EC to a Butler/ Mackerras-type pendulum would be to treat, say, New Mexico as if it were a block of five single-seat electorates that would all swing together as one if, but only if, its popular vote changes by 5.6% or more.
    Where there’s a will…
    I vaguely seem to recall, many years ago, someone coming up with a variant of the “House of Commons U-shape” that could handle three-way contests – might have been the late “The Bulletin”, about four decades ago when Social Credit was polling double figures in New Zealand and the Lib-Soc Dem alliance very nearly outpolled UK Labour – but can’t remember the details. Just that it was very clever…


  2. Putting the margins in “Australian” terms certainly makes the margins in Ohio and Iowa look a lot less daunting. Though it’s still very weird seeing them occupy the same place on the pendulum as Texas.


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.