US midterms: House of Reps

The United States goes to the polls on Wednesday (Tuesday there) for its regular two-yearly congressional election, choosing all 435 members of the House of Representatives and one third of the Senate. We’ll have a look at the Senate tomorrow; today, the House.

The first thing to say about House of Representatives elections is that if (like this one) there is no presidential election at the same time, the swing is almost always against the party of the sitting president. Voters effectively treat it as a giant by-election and give the party in power a kick, as they used to do with separate half-Senate elections in Australia.

Only once since 1934 has that rule not held: in 2002, a year after the 11 September attacks, the Republicans scored a very small swing in their favor. On one other occasion, in 1998, the party in the White House managed a small net gain of seats, although in conformity with the rule the two-party swing went (very slightly) against it.

The second thing to note, as you might expect from the “by-election” idea, is that the predictive power of mid-term elections is not very good. In 1994, for example, the Republicans scored a big swing to take control of the House, but went backwards in the presidential election two years later. Twelve years earlier, much the same had happened to the Democrats. And in the last three presidential elections, although the swings have been in the same direction as the preceding mid-terms they have been much smaller.

That said, the Democrat position is already precarious as to both presidency and congress. Due to the craziness of the electoral college the presidency could be lost in two years time with just a 0.3% uniform swing, even though Joe Biden won it by more than seven million votes. And the party’s majority in the House is only nine seats, 222-213: almost any swing at all is likely to deliver the Republicans the extra five that they need.

The third general point is that the relationship between votes cast and seats won is even more arbitrary than is usual in systems with single-member districts. The problem is that the drawing of boundaries is mostly in the hands of the politicians, so gerrymandering is rife on both sides (although the Democrats can at least claim to have made some efforts to end the practice). That leads not just to direct distortion, but also to a huge number of uncompetitive seats, which make the overall voting figures unreliable.

In 2020 the Democrats won 51.6% of the House two-party vote, suggesting a small bias against them in the boundaries – most of which have since been redrawn following the decennial census. But that figure should be taken with a large grain of salt, since there were so many seats in which sitting members were unopposed or faced only token opposition.

Conventional wisdom has it that a Republican win this week is a near certainty. In the middle of the year there was a revival of Democrat optimism, especially after the supreme court overturned abortion rights. But the favorable trend in the polls reversed itself a month or two ago, and Republicans now hold a clear if relatively narrow lead. Only Nate Silver’s partisan alter ego seems to give the Democrats much of a chance.

As of today FiveThirtyEight’s model gives the Republicans a 74% chance of winning a House majority; back in September that number dipped below 60%. That’s the polls-only or “Lite” model, which I regard as more accurate (it had a better record in 2020); the “Deluxe” version, which includes Silver’s bells and whistles, comes out at 83%. And at Sportsbet you can get the improbable odds of 15-2 against a Democrat win.

Assuming there is no upset and the Republicans take control of the House, the remaining question is how much difference it will make to 2024 and to the fate of American democracy more generally. The existing margins are so tight that Biden has not had an easy time in congress at any rate; he may even think that the increased opportunity for the Republicans to make fools of themselves (much as they did in 2011-12) will work to his advantage.

Perhaps more importantly, it will also put the divisions within the Republican Party more into the spotlight. As long as they were only in opposition, a certain degree of unity was forced on them; with control of one or even both houses, it will be harder to avoid taking some difficult stands. In particular, its leaders will have to decide what to do about a likely presidential bid by Donald Trump, which has the potential to split the party.

A Senate majority as well, of course, would add greatly to both Republican power and responsibility. Will have a look at the prospects for that tomorrow.