Welcome to my latest weekly round-up of the political polls. You can sign up at the end to get future weekly updates via email.
Five national voting intention polls have come out in the last week, with a little more variation than previously. Even so, the picture is consistent one of a mid-single digits Labour lead, with the Conservatives on 31-35%, Labour 39%-40%, Lib Dems 10-12% and Greens 4-7%. Under the hood, the more formal education a voter has, the more likely they are to support Labour and the less likely to support the Conservatives.
These polls are all based on the assumption that each party stands everywhere (with nationalists standing only in their own country). But there was also a poll out this week looking at possible seat deals, from Best for Britain which promotes the idea of a mix of Labour, Greens and Liberal Democrat candidates standing down for each other. No surprise therefore – and no complaint from me – that their own write-up goes heavy on finding support for the idea of a seat deal in the poll. But dig into the details and a more complicated story is found.
First up, the poll rather buried the lede, which was that it found Labour would win an outright Parliamentary majority on its own, without any seat deals being required. But, as Best for Britain points out, Nigel Farage’s creation Reform might at the next election stand down in some seats in order to help the Conservatives, as they did in 2019. And if that happens… the poll suggests that the Conservatives would do better, but still fall short of winning (with 261, well short of the 325 winning post). Although Labour would not have a majority, they would be by a large margin the largest party in a hung Parliament and, given the shortage of plausible support from other parties for the Conservatives, there would still be a Labour PM. So really the story of the poll is that it shows Labour could get into power without any seat deals.
However, Best for Britain also polled what would happen if in response to a Conservative/Reform deal, there was then also a deal between Labour, Lib Dems and Greens. The poll shows that would boost the number of non-Conservative MPs and so Best for Britain argues that shows the merits of such deals.
Let’s dig into that further though. What the poll shows is very little variation in seat numbers for each of the non-Conservative Parties across the different scenarios. If there were Labour/Green/Lib Dem seat deals in response to Conservative/Reform deals, Labour would only be boosted by 15 seats according to this poll, the Lib Dems by 5 seats (and the Greens still on zero). Those are pretty small changes, and given the inherent uncertainties not only with polling, and then also with constituency modelling, not to mention adding on top of all that trying to accurately discover how people will act under hypothetical seat deals, those changes are not large enough to be sure that they’re robust to all those potential sources of error.
It’s true that the poll matches other polling, such as from YouGov this last week, which shows a greater willingness of Labour, Lib Dem and Green supporters to switching party support between themselves rather than to/from the Conservatives. But the poll doesn’t make the case that a seat deal is required for Keir Starmer to become Prime Minister or even that a seat deal would make that much of a difference for sure – both because the effect it finds is small and because other factors outside the poll’s ambit would also impact the outcome, as I explained when looking at the previous Best for Britain polling before Christmas.
One of those other factors is the political risk of formal candidate deals and the ammunition that offers up to opponents as well as the dissent and demoralisation they can cause internally, all of which then hinder winning.
A second is the opportunity cost. Time spent trying to organise seat deals is time not spent on campaigning against the Conservatives. Time and money spent researching the impact of seat deals is time and money not spent researching how to win over Conservatives. (And remember that every vote won over from the Conservatives counts double compared to a vote shuffled between Labour/Lib Dems/Greens as a vote won from the Conservatives is both one of their total and one on the non-Conservative total.)
A third is that tactical voting can happen – and does, spectacularly at times, as we saw in last year’s by-elections. So the true like-for-like comparison isn’t between a ‘how would you vote tomorrow?’ question versus a ‘how would you vote tomorrow with a seat deal?’ but by comparing the latter with how people would vote after several weeks of a regular diet of bar charts and the like.
Measuring potential tactical voting is fearsomely difficult to do with polls (as I explain, ahem, in Polling UnPacked) and so this last point isn’t a criticism of the poll. Rather it’s a reminder of the limitations of what polls can tell us.
If that all sounds rather negative about the scope for cross-party cooperation to defeat the Conservatives, it’s worth remembering that cooperation can take many forms other than such formal seat deals, and there’s evidence other than polling to look at. Andrew Rawnsley’s latest column is excellent on this topic as is Duncan Brack’s new essay for Compass.
It looks like there will be a poll from another organisation on the subject of seat deals out this coming week, so I’ll return to this topic when it is.
Before that, here are some highlights from what else there was in this week’s political polls.
- Voters in the Red Wall both back public ownership of key utilities (67%-80% support depending on the sector) and a four day working week (66% support). Though as ever remember the caveat that a single poll on a policy issue only rarely captures all the angles to it, with different questions and a developing debate on the topic not infrequently resulting in different public takes.
- Polls continue to bring promising political news for Labour on the economy with only a quarter saying the Conservatives are best on the cost of living, Labour more trusted than the Conservatives on the topic (though scoring only 37% trust) and a shocking quarter of Britons saying they’ve skipped meals due to rising prices.
- For party leader ratings, the picture continues to be one of Starmer outscoring Johnson on best Prime Minister and on overall ratings with, when he’s included, Ed Davey doing better than either in the aftermath of good news for the Lib Dems, albeit it with much higher ‘don’t knows’. Leader ratings on specific attributes also broadly continues to show Starmer outperforming Johnson, save on questions around being good in a crisis or having lots of personality. Add these new findings to those voting intention figures at the start and Beergate doesn’t seem to have been the game changer that some Conservatives wish, tell others (and perhaps even believe themselves?) that it has.
- Returning to immigration after last week’s figures showing falling concerns over immigration, this time we have a new YouGov poll showing a near equal split on its impact over the last decade. Asked, “Do you think the level of immigration into Britain over the last ten years has been good or bad for the country?” people split 29% saying it had been mostly good, 29% again saying it had been mostly bad and 30% saying it had been both good and bad.
- And finally, a focus group in Wakefield brought bad news for the Conservatives ahead of the by-election there.
For up to date polling news during the week, do check out my dedicated Twitter account, @PollingUnPacked or sign up for the emails below. Both ways will also get you a reminder when next week’s round up is out.
For the full story, and the fun story, about political polling check out Polling Unpacked: the history, uses and abuses of political opinion polls. According to the Sunday Times, it is “Essential reading for anyone seeking to understand modern politics … comprehensive yet surprisingly fun”.