Boris Johnson has failed his first electoral test as British prime minister, with the Liberal Democrats winning, as expected, a by-election overnight in the Welsh seat of Brecon and Radnorshire. The Lib Dems held the seat for 18 years until the Conservatives won it in 2015; now their candidate, Jane Dodds, has taken it back.
It was less than a crushing win: Dodds had 43.5% of the vote, as against 39.0% for Conservative incumbent Christopher Davies and 10.5% for the Brexit Party. If the latter two had combined against her, they would have won.
Davies had been forced by petition to a recall election after being convicted of filing false expense claims. In the circumstances, the fact that his vote was only down 9.6% on 2017 is not a bad result.
The Lib Dems were up 14.4% – slightly more than the drop in the Labour vote, which went from 17.7% in 2017 to just 5.3%. [See note below] They also benefited from the absence of the Welsh Nationalists, Plaid Cymru, who had 3.1% last time but stood aside in the interests of “remain” unity.
“Leave” candidates in aggregate had 50.2% of the vote, or 51.3% if the Monster Raving Loony vote is included (the Loony position on Brexit is unclear). That’s very similar to the “leave” vote for the constituency in the 2016 referendum, estimated by political scientist Chris Hanretty at 51.9%.
It does suggest, however, that there’s been some movement since the last electoral test, the European parliament election held on 23 May. Hanretty has done estimates for that as well, and his figures for Brecon and Radnorshire are as follows:
Brexit Party 35.0%
Liberal Democrats 26.7%
Plaid Cymru 9.4%
Total “remain” (inc. Labour) 52.9%
Total “leave” (inc. Conservatives) 47.1%
So despite the Lib Dem victory, there’s been a shift back towards “leave”. But it’s dwarfed by two other effects: the collapse of the Brexit Party vote, and the success of the Lib Dems in capturing most of the “remain” vote.
Either of those effects on its own is hugely important for British politics. But if both occur to a similar extent, as they did in Brecon and Radnorshire, they partly cancel each other out.
To understand this, consider the European parliament election as a whole (or at least for Great Britain as a whole; I’m leaving out Northern Ireland because its party system is quite different). Here are the votes and seats actually won:
But that tells us very little about how a British general election – widely expected before the end of the year – might look, because the voting system is completely different. As you can tell from the close correspondence between votes and seats, the European parliament is elected on a proportional system.
But the British parliament is first-past-the-post in single-member constituencies. Hanretty’s extrapolation onto those boundaries gives the Brexit Party an extraordinary 414 seats, against 74 Lib Dems, 67 Labour, 64 Scottish and Welsh Nationalists, ten Greens and just one Conservative.
That, of course (as Hanretty himself stresses), will not happen. But the extent to which the result differs from it will depend primarily on two things: how the balance shifts between the Brexit Party and the Conservatives, and how successfully the “remain” vote can avoid being split among multiple parties in each seat.
To test what might happen, I re-ran Hanretty’s figures with some additional assumptions:
- A uniform swing of 20 points from the Brexit Party (including UKIP) to the Conservatives.
- A uniform swing of two points from Labour to the Conservatives.
- Consolidation of the Lib Dem, Green, CHUK and Nationalist vote in each seat into what I label the “Remain coalition” vote.
With the usual caveat that swings are never uniform, the first two seem to me to be in the right ballpark. The third is unrealistic, in that consolidation will not be complete, but it will be offset by the fact that there will also be some consolidation via tactical voting between the Remain coalition and Labour, which I haven’t allowed for.
Anyway, on that basis I got the following: Remain coalition 404, Conservatives 115, Brexit Party 65, Labour 48.
That seems an unrealistically large “remain” victory. So I added an extra assumption:
- Tactical voting by Brexit and Conservative voters, to the extent that in each seat a third of the support for whichever is running behind will migrate to the other.
That delivers an extra seat to the Brexit Party (from the Remain coalition) and an extra 107 to the Conservatives – 97 from the Remain coalition and ten from Labour. The new totals are 306 Remain coalition, 222 Conservatives, 66 Brexit Party and 38 Labour. (It should be noted that because of the tactical voting mentioned earlier, some of what I’m calling Remain coalition wins would probably go to Labour.)
Add in Northern Ireland, which will deliver between six and ten extra seats to the “leave” side, and you’ve got a very interesting looking House of Commons.
I don’t for a moment think it will be exactly like that. And it could be very different: if the Tories cannibalise almost all of the Brexit Party vote, but the “remain” parties fight among themselves, there could easily be a large Conservative majority.
Conversely, if Farage’s party holds onto a reasonable chunk of its May vote and his supporters resist any urge to vote tactically, while the other side displays the sort of unity that they did in Brecon and Radnorshire, it will be a big triumph for “remain”.
And if either side wins a clear majority on a minority of the vote, which is also very possible, it will bring further discredit to the already battered project of British democracy.
Note: I’ve re-written the fourth paragraph above because in the original version I omitted the Labour vote from 2017 – being in a hurry I had neglected to check Wikipedia’s figures, which for some reason had left off the Labour candidate (it’s since been fixed). Adam Carr has the correct figures.