I’m in Crikey today with a preview of the US mid-term elections. Voting will mostly take place overnight, Australian time, and results will start appearing from 10am tomorrow – FiveThirtyEight has a schedule.
The subeditor’s tagline says that my article has “everything you need to know,” but I’m afraid that’s not true. There’s a lot that I didn’t have room for, and I’ll try to cover some of it here.
The striking fact that requires explanation about this election is that the Democrats seem to be doing so much better in the House of Representatives than in the Senate. They’re hot favorites to win back control of the House (88%, according to Nate Silver’s classic model), but the Republicans have almost as big a chance of holding their majority in the Senate (80.5%, from the same source).
I mention some of the reasons for that: the fact that only a third of the Senate is elected each time, that 2012 was a particularly strong Democrat year, and that a 50-50 tie counts as a Republican majority because Mike Pence has a casting vote.
But of course another reason is the simple fact that, while the House is gerrymandered, the Senate is malapportioned. Very big states get the same representation as very small states, and, as in most places, smaller states tend to vote conservative.
So it happens that ten of the Democrat incumbents seeking re-election are in states that Donald Trump carried two years ago – including the four that are at serious risk of losing (Florida, Indiana, Missouri and North Dakota).
But there’s another puzzling thing as well about the contrast between the two houses. While the forecasting models say that the Democrats (as I just mentioned) are slightly more likely to win in the Senate than they are to lose in the House, you’d never guess that from the media coverage.
Most pundits have been focusing on the House, treating it as very much a live contest, and dismissing the Democrats’ chances in the Senate, often in quite cursory fashion.
I take the opposite view. I think for the Republicans to retain their majority in the House would be a major upset, basically for the reasons Silver gives in his final summary:
At a macro level — based on national indicators and the historical tendency of the president’s party to lose seats at the midterm elections — the situation looks bad for Republicans. But at the local level — when you evaluate factors one district at a time, as our model does — it looks worse. The polling is bad for Republicans, the fundraising numbers are awful, and the slate of potential Democratic pickups runs deep into Republican territory. The data is uncertain, because it contains a margin-of-real-world-error. But I don’t think the data is ambiguous. It says Democrats are over the threshold they’d need to win the House.
On the other hand, I have a nagging feeling that a Democrat win in the Senate is a real possibility. Perhaps it’s just my natural optimism; certainly I can’t claim I’ve been immersed in the data the way some people have. (I’ve probably given more attention to New Caledonia over the last week than I have to the US.)
In the Senate, inevitably, we look at the races individually; no-one thinks that nationwide polling is much use. House seats can be taken as to some degree interchangeable; a swing of x% will deliver about y seats, and you don’t feel you have to specify which ones. It’s like the House of Representatives in Australia: if the swing is there, the seats – approximately – will fall.
Senate contests, however, being bigger and fewer, are more high-profile; it’s about the individual contest more than the national trend. But I worry that this way of looking at it might lead us to misestimate the chances.
If you think of several contests as independent events, you will miss possible connections between them. On a racetrack, for example, knowing who won the first race doesn’t tell you anything much about who will win the next race.* But knowing who won one seat in an election is valuable information when looking at other seats, because if a party is picking up votes somewhere it may well be doing it across the board.
Now, I’m sure Silver’s model allows for this to some extent. But I have a suspicion it might not be enough, and that if the Democrats are on a roll, then the Senate might come out better than expected. Unquestionably, however, the odds are against them.
One Senate race in particular is worth special mention: the by-election in Mississippi, where the Democrats need a swing of 11.3% (two-party) to take the seat. That’s very unlikely, but under Mississippi’s idiosyncratic electoral law the election is just a first round; if no candidate has more than 50% of the vote, there will be a runoff three weeks later.
There are three serious candidates; two Republicans and a Democrat. It’s very likely that the Democrat will lead tomorrow, with the other two splitting the Republican vote. If the Democrats have picked up (net) a seat elsewhere, then the Mississippi runoff would decide control of the Senate, with all the national attention that would entail.
In the preview I also talk a bit about the Governors’ races, which are often neglected:
Not surprisingly, gubernatorial elections are more likely to be decided on state rather than national issues. But it would a mistake to think that their outcome is only of interest to residents of those states.
The key thing this year is that these newly-elected governors will be in office when the next census is taken in 2020. That gives Democrats the opportunity to have a say — possibly a decisive say — in the redistribution of electoral boundaries that will take place then.
Given how badly the country needs some fair elections, that would be a big step forward.
There are some other very good summary previews around. Here’s Brendon O’Connor and Dan Dixon in the Conversation; here’s Lesley Russell at Inside Story. And of course the American media are overflowing with information: this piece by Alexander Burns and Jonathan Martin in the New York Times would be a good place to start.
There’ll also be liveblogging tomorrow from everybody and their dog. I may join in if I feel I have anything particularly useful to say.
* It might tell you something; particular track conditions might suit some runners more than others consistently across the day. But for most purposes it’s OK to treat horse races as independent events.