I would strongly suggest that the LD second place in both Cities of London & Westminster and Finnchley & Golders Green were aberrations, and that in reality Labour will be the main challenger to the Tories in those seats in 2024. Historically that has been the case. Wimbledon also could reasonably be viewed as a three-way contest.

Expand full comment

This seems rather disingenuous. As usual, professional pundits overestimate or overstate the likely effect of tactical voting. Primarily because they overlook the extent to which the 'base' figures (based on uniform swing) already include a significant amount of tactical voting at the previous election. The purported effects of tactical voting are really projections of an INCREASED level of tactical voting over and above that already included in the base figures. And there must be limits to the plausible extent of increased levels of tactical voting at both the upper and lower bounds. The 'upper bound' being the limit at which voters in a particular constituency are unlikely to vote tactically because they don't see any realistic likelihood of a vote for their second-choice party being sufficient to displace the incumbent. The 'lower bound'being where the 3rd or 4th placed party vote has already been squeezed so far that the only remaining voters for that party (from the previous election) are those who will insist on voting for a 'lost cause' and refuse to vote tactically for a 2nd-preference candidate under almost any circumstances.

In simple terms, for Labour to achieve an effective overall majority, they would need to gain approximately 100 seats from Conservatives and a further 15 from SNP. Or, to put it another way, roughly speaking, all the seats which they lost in 2019, plus any others where in 2019 they came 2nd and within 10,000 votes of the incumbent MP. They might or might not achieve the necessary swing to achieve that goal (based on current voting intention polling it would be too close to call). But, in any event, its extremely unlikely that any plausible increased level of tactical voting over and above that already in effect in the 2019 election would tip the balance.

Of the top 12 seats listed as most likely to be susceptible to increased LD > Lab tactical voting, I would suggest that 10 would likely fall to Labour anyway. And the other two (Bromley and NE Somerset) would likely remain Conservative. 2 or 3 of the next few seats on that list (Basingstoke, Banbury & Beckenham) would be the only constituencies which I would estimate likely to be susceptible to Labour gains due to increased levels of LD > Labour tactical voting.

In addition, before we leave the Con/Lab marginals, 2 seats further down the suggested likely list of those susceptible to increased tactical voting (Blackpool North and Great Grimsby), I would also estimate as likely Labour gains, irrespective of the actions of 2019 Lib Dem voters. These two seats are extreme examples of where the LD vote has already been squeezed so hard that there is highly unlikely to be anything substantial enough left in 3rd party votes to make a difference to the 2024 result. Albeit those Blackpool North numbers possibly moot, due to significant boundary changes.

The list of possible additional Lib Dem gains due to increased Labour > LD tactical voting seems rather more plausible. That is despite the methodology used to arrive at that conclusion being worse than dubious. There is no rational foundation at all to suppose that a (reasonable overall estimate) 10% Con > Lab swing should convert to a 5% swing to LD in constituencies where they currently place second. That might plausibly happen as a result of Lab > LD tactical voting -- but to present that as a 'starting point' for likely FURTHER tactical voting is rather like having your cake and eating it. It smacks of the scientist deliberately modifying the rules of the experiment in order to 'prove' his pet theory.

A better formula / methodology might be,: Instead of assuming a 'flat' uniform swing of 10% Con > Lab (regardless of existing/previous vote shares) -- to hypothesize a 'uniform' (base starting point) reduction of 25% in the Conservative 2019 vote and a uniform increase of 30% in the 2019 Labour vote. With no change, at this point, to the 2019 LD vote.

[Using this formula of Con -25% and Lab +30% rather than simple 10% swing is also how I got the additional 10 likely Lab gains not subject to increased tactical voting.]

That could result in 2 further seats (Hitchin & Harpenden and Wokingham) falling to Lib Dems, solely due to the Con > Lab swing. Against that, 3 'central' London seats (Finchley, Westminster and Chelsea) have to be considered as 3-way marginals. The slightly more sophisticated suggested version of the Con > Lab swing would be sufficient to push Labour ahead of Lib Dems into 2nd place in each of these constituencies. And any attempt by voters to be 'clever' by voting tactically there would be at least as likely to be counter-productive as to achieve the intended result.

The remaining 13 seats suggested are all then more or less plausible LD gains due to increased Lab > LD tactical voting to different degrees. Although I would group and order them rather differently in terms of likelihood, thus:

Highly plausible:

1. Wells

2. SW Surrey

3. Taunton Deane

4. Brecon

5. Harrogate


6. Chippenham

7. Romsey

8. Sutton & Cheam

9. Wantage

10. SE Cambs

11. Woking


12/13 Totnes & Witney

The reason for the different ordering of likelihood/feasibility is the proposition that its not simply the additional percentage of 3rd-place votes necessary to achieve an unseating that matters. Its easy enough to envisage 50% of a 3rd-place party's votes tactically transferring to the second-place party in the right circumstances. At least as important as that is the perception in a potential tactical voter's mind as to how likely such a switch of vote would be to make the desired difference. That might reasonably be measured in terms of the ratio of projected 2nd-place votes to projected 3rd-place votes. I opted for an arbitrary cut-off value of 2.5 to 1 (after taking account of revised 'uniform' swing) as the point beyond which enough additional natural Labour supporters might consider tactically voting Lib Dem in distinguishing 'likely' vs 'unlikely' tactical gains.

Expand full comment

Is there any way of getting this analysis into the heads of Keir Starmer, Labour MPs, councillors, campaign leaders and activists?

Expand full comment

Rob, avid reader of your work. I notice Wimbledon isn't included, is there a particular reason for this?

Expand full comment