This is a longer and more detailed version of a post that was originally published by Prospect here on Friday. The opinion polls published over the weekend do not collectively show any substantial change from the figures in the table below.
By Stephen Fisher, 9th September 2019.
The government has lost its majority in parliament, so it is unlikely to be long before there is a general election. Already at the last election, the Conservatives attracted most Leave voters while Labour was most popular among Remain voters. So what has changed in the party preferences of Leave and Remain voters, and with what implications for the next election?
The Conservatives lost many of their Leave voters to the Brexit Party at the European elections earlier this year largely because of the government’s failure to deliver Brexit on schedule at the end of March. At the same time, frustrated with the complexity and ambiguity in Labour’s position on Brexit, many Labour Remain voters switched to the Liberal Democrats or Greens.
Since May (the month and the leader) the challenger parties have waned and the two main parties have recovered somewhat, but Westminster-vote intentions are still closer to the multi-party competition we saw in the Euro elections than they are to the two-party dominance of the 2017 election.
The table shows how Leave and Remain voters from 2016 voted in 2017, and how they intend to vote now. The changes show that the rise of the Liberal Democrats and Greens has been largely confined to Remain voters, while the Brexit Party has, unsurprisingly, only really attracted those who voted Leave in the referendum.
Table: 2017 vote and current vote intention, by 2016 vote
Note: 2017 figures come from the British Election Study Internet panel. 2019 figures are the average of the most recent poll from each of Deltapoll, Opinium, Survation and YouGov taken between 21 August and 3 September 2019. The UKIP/Brexit row shows figures for UKIP from 2017 and the Brexit Party in 2019, and the change between the two.
What is perhaps more surprising is just how both main parties have suffered setbacks among both Leave and Remain voters. Labour have fallen back almost as much among Leave voters as they have among Remain voters. While the largest outflow from Labour in absolute numbers is that among Remain, in relative terms, it is the other way round. Labour have lost just under a third of their 2017 Remain voters, but as much as half of their former Leave voters. As a result, Labour supporters are even more predominately on the Remain side than they were in 2017.
Meanwhile, far from simply losing Leave voters to the Brexit Party, the Conservatives have lost almost as many Remain voters, mainly to the Liberal Democrats. Again, although there are similar absolute numbers of Leavers and Remainers who have deserted the Tories since 2017, in relative terms there is a big difference. The Conservatives have lost just one in six of their Leave voters, but more than a third of their Remain voters since 2016.
Now, among Labour vote intenders Remain voters outnumber Leave voters by roughly three to one, conversely those who say they will vote Conservative are roughly three to one on the Leave side.
From votes to seats
When it comes to thinking about what might happen at a general election in different constituencies with different levels of support for Brexit, the increasing correlation between major-party support and Brexit position at the individual-level makes surprisingly little difference for seat outcomes.
Since the rate of decline of the Labour vote is only slightly greater among Remain than among Leave voters, on the figures in the table we would expect the Labour vote to drop only slightly more in places which voted Remain than in places which voted Leave. Similarly, little geographical variation should be expected in the change in the Conservative vote as a result of the Brexit divide.
The correlation across constituencies between the 2016 referendum vote and the 2017-19 swing suggested by recent polls is so weak that the outcome of a seats projection which takes differing Leave and Remain voting patterns into account is practically the same as one based on traditional uniform change.
This remark is perhaps most surprising with respect to the Liberal Democrats, who are expected to do much better in more Remain voting areas. As it turns out, on current levels of overall support, concentrated growth among Remain voters does not yield them any more seats than they would have achieved with the same rise in all constituencies.
In part this is because the vast majority of constituencies are fairly evenly divided between Leave and Remain. No seat has been estimated to have fewer than 20% support for the local losing side at the referendum, and the local losers make up at least a third of the constituency voters in over 90% of seats. So called “Leave seats” still have many Remain voters, and vice versa.
For the Liberal Democrats, this means that having their growth almost entirely concentrated among Remain voters does not actually concentrate their strength geographically very much more. Also, given their poor performance in 2017, there are barely any seats where the party are starting from a strong enough point and where there are enough Remain voters to help them win when they otherwise would not win on a uniform 10-point swing from the Conservatives or 12-point swing from Labour.
I imagine anyone who has read this far is keen to know how many seats each party would be projected to win. I hesitate to provide the figures because they are not a forecast! Much could change in overall levels of support for the parties between now and the election, as it did spectacularly in the 2017 campaign. Also the polls could be wrong again. The purpose of this blog is primarily to elucidate the nature of the electoral geography given current polling patterns, not to forecast the outcome.
That said, the projections give the Conservatives 320 seats, a similar tally to the 318 they won last time, and enough for them to govern with the Democratic Unionist Party.
The Conservatives’ increased lead over Labour in votes (up from 3 points in 2017 to 7 points in recent polls) might not yield a majority, but it would hurt Labour. The party is projected to lose over 30 seats, mainly to the Conservatives. Those Tory gains, would be offset by losses to the Liberal Democrats and SNP, who are projected to increase their tallies to 31 and 48 respectively.
To secure a majority the Conservatives are trying to win over Brexit Party supporters. If they succeed in winning them all, Boris Johnson could be returned with a majority of over a hundred.
If that happens, it will be difficult for Remain voters to coordinate to stop them. Tactically avoiding the flow of Labour votes to the Greens and Liberal Democrats in the Labour-Conservative marginals is not enough. Nor would additional Labour to Liberal Democrat tactical voting in Conservative held Lib Dem targets; the Labour vote was already squeezed in those seats in 2017. If Brexit Party supporters rally behind the Prime Minister, it would take unprecedented levels of tactical voting by Labour and Liberal Democrat voters in favour of the other, and all perfectly coordinated across constituencies, to prevent a Conservative majority.
On the other hand, if the opposition supporting Remain voters do succeed in tactically coordinating in this way and the Brexit Party support holds up, the Conservatives would be reduced to fewer than 230 seats and Labour would win a majority, albeit a modest one of 30.
Tactical coordination among Leave voters would be more politically effective than it would on the Remain side because of the more even geographical distribution of the Leave vote. While the Remain vote is more concentrated in Scotland, London and some big cities, there are small Leave majorities in most seats. Overall, while Leave won 52% of the vote they are estimated to have won 64% of constituencies. So long as those who did not (or could not) vote in 2016 continue to turn out in general elections at low rates, Leave voters can win big majorities under first-past-the-post if they can coordinate on a single party to represent them.
Not only is coordination among Leave voters more effective given the electoral geography, but politically it looks easier. Polls consistently show that Brexit Party supporters are much better disposed to Boris Johnson than Liberal Democrat supporters are to Jeremy Corbyn, or for that matter Labour party supporters to Jo Swinson. In Scotland, among supporters of Labour, the Liberal Democrats and the SNP there is perhaps still greater antipathy for each other’s parties.
The leadership of the opposition parties in Westminster might currently be able to coordinate on how their MPs should vote in parliament on the timing of the election. Whether those parties can coordinate on how tactically their supporters should vote at the general election is another matter.
The danger for Remain voters, and others concerned about the possibility of a no-deal Brexit, is that the parties they support might attract only Remain voters. In particular, for Labour the route to majority government depends on winning back Leavers as well as Remainers.
Acknowledgements: The analysis for this blog uses estimates of 2016 referendum results by constituency which were compiled by Chris Hanretty.