If you’re a Democrat (or even just a democrat) in America you have to get used to mood swings. A few days ago there was deep depression when a New York Times opinion poll showed Donald Trump with a convincing lead over Joe Biden in critical states. But yesterday (Tuesday in the US), actual voters in several states gave a message rather more favorable to the Democrats.
In most states, the important elections are in even-numbered years, but a few go the other way and vote in the off years (you can read my preview here of the previous round, two years ago). Contests this year included three state legislatures, two governors, assorted mayors and two important referendum questions.*
Kentucky and Mississippi are both very solid Republican states, and both had governors seeking re-election. But Kentucky’s governor, Andy Beshear, was actually a Democrat, having won in an upset result four years ago with 49.2% of the vote (50.2% two-party). Yesterday he won a second term, with a still narrow but more comfortable 52.5% against Republican Daniel Cameron.
Mississippi also swung slightly towards the Democrats, but not enough to unseat governor Tate Reeves. With a few polling places still to report, Reeves has won with 51.6% of the vote (52.3% two-party), down from the 51.9% (52.6% two-party) he recorded in 2019. Republicans also retained control of both houses of the Mississippi legislature, increasing their majority slightly in the lower house and remaining unchanged in the state senate.
New Jersey these days is a fairly solid Democrat state, although governor Phil Murphy won only narrowly in 2019, with 51.2% of the vote (51.6% two-party). He wasn’t up for re-election this time, but Democrats comfortably held control of the legislature; in the lower house they increased their majority from 46-34 to 51-29, while the upper house was unchanged.
As usual the most interesting of the off-year states was Virginia, although there too the governor was not up for re-election. Republican Glenn Youngkin won the governor’s race last time with (you may be seeing a pattern here) a very narrow 50.6% (51.0% two-party); Republicans also won control of the lower house of the legislature, 52 seats to 48. The state senate, which had a 21-19 Democrat majority, has four-year terms and was not up for election then.
So Youngkin, who represents what these days passes for the mainstream in the Republican Party and has been toying with a run for its presidential nomination, had the opportunity to win undivided control of the legislature, while the Democrats had the chance of taking both houses. It was the latter whose hopes were realised: they kept their senate margin intact and narrowly seized back control of the lower house, 51-48 with one seat still undecided.
Finally there was Ohio, whose voters were deciding on two ballot measures: one to safeguard the right to abortion, and one to legalise the recreational use of marijuana. Both were carried by decisive margins – 56.6% and 57.0% respectively – continuing a trend of success for both issues across the country in recent years.
The salience of the abortion issue in particular, on which Republicans have painted themselves into a small and unpopular corner, is unquestionably part of the reason for the Democrats’ relatively strong performance. But they are not getting as big a benefit as they might have hoped: the clearest message from yesterday is that the country remains deeply polarised, and close elections are going to be the norm for the foreseeable future.
That in turn makes the opinion polls, especially those conducted well in advance, even less of a useful guide than usual. And with both major parties (albeit for different reasons) unhappy about their likely presidential nominees for next year, it could take a long time, despite the abundance of data, for the shape of the race to really become clear.
.
* That’s not counting Louisiana, which also has elections this year (for both governor and legislature) but not this week: the first round was last month, on 14 October, with runoffs next week, on 18 November.