Splits and squeezes part 1: local stakes, local choices
Polls are national, but elections are local
A general election is not a single national contest. It is 650-odd local contests fought on the same day. Opinion polling gives us a sense of the national landscape of opinion, glossing over all the differences between one seat and the next. This is pretty good for getting a sense of the overall lay of the land, but in some circumstances local conditions can matter a lot. One example is tactical voting. The stakes facing voters vary from one seat to the next, with different parties in and out of the running. The degree to which voters recognise and respond to this can tip the balance in some seats and hence affect how overall vote shares translate into Westminster seats.
One big example of this came in the 1997 general election, when many voters were so eager to eject the Conservative government that they switched in large numbers to whichever party was best placed to defeat their local Conservative MP. The result was a a big boost for both Labour and Liberal Democrat challengers, sufficient to deliver 25-35 extra Tory seats to the opposition parties.1
A more recent example comes from Nigel Farage’s decision to stand Brexit party candidates down in Conservative held seats in 2019, while continuing to stand candidates in Labour held seats. Though Farage presented this as an initiative to assist Boris Johnson, it had the opposite effect. The absence of Brexit party candidates padded out Conservative incumbents’ 2019 margins of victory, but most didn’t need the help in an election where Labour were floundering. But Farage severely damaged Tory challengers’ prospects in Labour Leave seats, and by splitting the Leave vote, Farage’s Brexit Party probably saved upwards of 20 Labour MPs from defeat.2
Local context shapes the choices voters make in both these cases. The first is an example of tactical squeeze - supporters of smaller parties out of the running opting to coalesce behind the local front runner, who they prefer to the local incumbent. Lib Dem voters backed Labour in Con vs Lab seats and Labour voters back Lib Dems in Con vs Lib Dem seats as both groups of voters saw that tactical switches served the common goal of defeating the Conservatives.
The second example is the opposite dynamic - instead of voters coming together locally to ensure a national outcome they all want, voters fragment locally and undermine the national outcome they all want. This kind of split voting can enable a party to win locally even when a majority of local voters would prefer someone else in a two candidate contest. Hence a strong local showing for the Brexit party in 2019 paradoxically ended up helping many Labour MPs hold on, even though a majority of voters in these seats would have preferred the Conservatives to Labour if those were the only choices on offer.
Squeeze options - red, gold and unionist
There are signs the next election may resemble 1997 in one key respect: the differences between Starmer’s Labour and Ed Davey’s Liberal Democrats are modest, and voters’ desire to oust an unpopular Conservative government is strong. This creates favourable conditions for a “red squeeze” on third placed Lib Dem voters in Con-Lab contests, and a “gold squeeze” on third placed Labour voters in Con-LD contests. The same forces may also encourage third placed Plaid supporters in Welsh contests, and to third placed Green voters nationwide in many seas.
There is another tactical squeeze option in 2024 which wasn’t relevant in the Blair era - unionist squeeze in Scotland. Just as English opposition voters may be motivated to coalesce whoever is best placed to oust the local Conservative, so pro-union Scottish supporters of local third placed parties may be motivated to switch to the strongest unionist pchallenger to the SNP.3
There has been some evidence of substantial unionist tactical voting in the 2021 Scottish Parliament election, though with three unionist parties, there are several different possibilities, not all equally palatable. The easiest switches are Lib Dems voting for Labour in SNP-Labour marginals, or Labour voting for Lib-Dems in an SNP-Lib Dem marginals. Tactical co-ordination against the SNP should be an attractive option for unionist voters of both parties, for the same reason tactical co-ordination against the Conservatives is attractive for both parties in England and Wales.
It may also be possible for Labour and the Lib Dems to attract unionist tactical votes from Scottish Conservative supporters. If the Conservatives are struggling nationally, and out of the running locally, then the often strongly held unionism of Scottish Conservatives may take precedence over their traditional partisan antipathies. Kicking out a local SNP MP may be a consoliation prize sufficiently attractive to justify a tactical vote for a Lib Dem or even a Labour challenger.
Perhaps the least plausible kind of unionist tactical squeeze is Labour/Lib Dem voting for a better placed local Conservative candidate. Such candidates’ chances will be poor anyway when the Conservatives are struggling nationally, and the same resentments that are driving Conservative national decline in the polls will also serve as a barrier to tactical switching.
The Brexit split
The next election may feature unusually large effects from split voting as well as squeeze voting, with one party in particular holding the power to cause substantial disruption in Conservative seats.4 Brexit party successor ReformUK could cause a Brexit split in most Conservative seats if they follow through on threats to stand candidates nationwide in 2024. Most Tory MPs did not face a Brexit Party competitor locally in 2019 as the party’s then-leader Nigel Farage opted to stand candidates down in Conservative held seats. If current party leader Richard Tice follows through on his pledge to reverse that decision, most Tory MPs will have a new local opponent next time. Even if ReformUK perform poorly in terms of national votes, their capacity to split off strong Leave supporters who would otherwise vote Conservative could enable them to disrupt the balance of power in many local contests.
How much difference can squeezes and splits make?
Answering this question is not an easy task. The impact of local tactical movements depends on the size of the voter pool who might consider such shifts, the rate at which tactical voting actually happens in this group, the size of the local majority, and the patterns of national swing.
The first question - the size of the local squeezable or splittable vote - is fairly straightforward to answer. For the squeeze options, I simply aggregate together the 2019 vote shares for all the third placed parties who might be open to each kind of tactical switch. For the Brexit split, ReformUK are currently polling roughly half the vote share achieved by their predecessor UKIP in 2015, so I take half the 2015 UKIP vote as an indication of potential local sympathy for a Brexit splinter party. This is restricted only to the seats held by the Conservatives in 2017, where Farage opted not to stand candidates in 2019.
Doing this sum reveals there are 80 Con vs Lab seats where the squeezable local vote from Labour friendly third placed parties in 2019 was 15 per cent or more, and another 74 where a squeezable vote of 10-15 per cent is available.
There are 44 Con vs Lib Dem contests where the vote for squeezable third place parties is above 15 per cent, and another 18 where the squeezable vote is 10-15 per cent.
The union squeeze vote is even larger - there are 16 SNP vs Lab and 2 SNP vs Lib Dem seats with a squeezable third placed unionist vote above 20 per cent, and another 10 Lab vs SNP contests where it is above 10 per cent.
Finally, there are half a dozen Conservative held seats where a new ReformUK candidate could plausibly win above 15 per cent of the local vote, and another 27 seats where a new ReformUK candidate could split off 10-15 per cent locally.
These are hefty figures - big tactical squeezes or splits look theoretically possible in hundreds of seats. However, many of these will be very safe seats where the local Conservative or SNP vote is large enough to weather even big squeezes or a big ReformUK splinter vote. Conversely even small splits and squeezes can matter where the local race is close. A large local shift is not necessarily decisive, while a small shift can have big implications.
How many seats could squeezes and splits have flipped in 2019?
We can start to get a handle on where tactical voting could, at least theoretically, be decisive by working out the maximum impact from tactical voting on a local result would be be with everything else unchanged. To do that, I carry out four simple reallocations:
Maximum red squeeze - reallocate all of the third placed or lower Lib Dem, Green and Plaid Cymru votes in Conservative-Labour contests, and see how many seats Labour gain
Maximum gold squeeze - reallocate all of the third placed or lower Labour, Green and Plaid Cymru votes in Conservative-Lib Dem contests, and see how many seats the Lib Dems gain
Maximum unionist squeeze - reallocate all of the third placed or lower Conservative, Labour and Lib Dem votes in SNP-Labour and SNP-Lib Dem contests and see how many seats Labour and the Lib Dems gain
Maximum Brexit split - deduct half of the 2015 UKIP vote share from the 2019 Conservative vote share in seats where the Brexit party did not stand in 2019 and see how many seats the Conservatives then lose to Labour and the Lib Dems
Here’s what happens if we apply a maximal version of each split or squeeze, and change nothing else at all:
Seats changing hands in 2019 in each maximum squeeze or split scenario
If every third placed Lib Dem, Green or Plaid Cymru voter switched their support to Labour in 2019, the party would be 43 seats better off than it was (I’ve posted full lists of the seats changing hands in each scenario at the end of the article). While many of these are defeats averted in ultra-marginal battlegrounds, the party also picks up a number of seats in harder to reach places. A perfect red squeeze would deliver Chipping Barnet, Truro and Falmouth, and Wycombe, none of which have elected Labour MPs in 70 years or more. Winning over all sympathetic smaller party voters would also have delivered the seats of Iain Duncan Smith (Chingford and Woodford Green, Con since 1966) and Sir Graham Brady (Altringham and Sale West, never Labour previously). Rushcliffe, held by Tory veteran Ken Clarke for nearly 50 years, would have flipped to Labour on his retirement.
A perfect gold squeeze would have delivered an extra 17 Conservative seats to the Liberal Democrats, more than doubling their 2019 Commons intake with wins in a mix of familiar and new battlegrounds. Nearly half the gains would restore the Lib Dems in seats they held prior to the post-Coalition collapse of 2015. The other half would come in strongly Remain seats that have turned against the Tories since Brexit, including Wimbledon, South Cambridgeshire and Woking. Several of the high profile defectors running in Lib Dem colours in 2019 would have won with a perfect gold squeeze, including Chuka Umunna (Cities of London and Westminster), Luciana Berger (Finchley and Golder’s Green) and Philip Lee (Wokingham).
A perfect squeeze on third place pro-union votes would have delivered 19 SNP seats to Labour, and saved Lib Dem leader Jo Swinson from defeat in East Dunbartonshire. While some of these are marginal seats Labour had already managed to reclaim briefly in 2017, many are not. In particular, a strong unionist squeeze would offer Labour much brighter prospects of recovery in many central belt strongholds where the party had towering majorities prior to the SNP tidal wave of 2015.
Finally, if ReformUK were to stand new candidates and split off half the votes its predecessor UKIP claimed in 2015, then 9 Conservative seats would fall - six to the Lib Dems, two to Labour and one to the SNP. Paradoxically, the biggest impact from a Brexit split comes in Remain seats such as Winchester and Wimbledon where UKIP performed poorly in 2015, but where even a small Brexit split would be enough to tip the local balance.
From 2019 to 2024
There is a lot of theoretical potential, then, for squeezes and splits to influence outcomes in many seats. There are many seats where the squeezable or splittable 2019 vote is large, and many seats where squeezes or splits could have changed the outcome. But neither of the measures I have tried out can give us a good handle on the question that really matters - where will the impact of squeezes and splits be felt next time, and how much could it change the outcome? Answering this requires grappling with several further issues.
We need, first of all, to price in the broader shifts in the electoral landscape. Tactical voting won’t change the outcome in many of the seats where tactical voting could have been decisive last time, as they are ultra-marginal contests that would fall anyway due to national swings. If Labour are on course to form the next government, they won’t need tactical votes to Burnley and Bridgend. We need to know where tactical voting could tip the balance after accounting for national swing.
We also need a more realstic measure of how much tactical voting is possible. Third party support is never pushed to zero. Many smaller party supporters are too committed to their first choice, too hostile to the local second placed party, or not hostile enough to the local winner, to want to switch. Others might be open to switching but unaware of the local stakes. A more realistic estimate of tactical squeeze needs to account for this.
Finally, we need to combine squeezes and splits. Many of the Conservative seats where tactical squeeze could matter are also seats where a new ReformUK candidate could produce a Brexit split. The combined effect of squeeze and split could be decisive in seats where each effect in isolation is not.
In part 2 of this analysis, I will build a simple model to estimate the combined effects of splits and squeezes when assuming a national Conservative to Labour sufficient to put the opposition close to majority territory, and realistic rates of vote shifting in response to local circumstances. This will give us a better sense of the potential for squeezes and splits to influence the general election outcome next time.
Lists of seats which would have changed hands in 2019 if we apply “perfect squeeze” or “Brexit split” scenarios
Perfect red squeeze Labour gains in 2019
Seats not won by Labour in 2017 in bold
2. Perfect gold squeeze Lib Dem gains in 2019
Predicted gains in seats not held by LDs pre-coalition in bold
3. Perfect Unionist squeeze Labour and Lib Dem gains in 2019
Predicted gains in seats not won by an opposition party in 2017 in bold
4. Brexit split Labour and Lib Dem gains in 2019
See for example Evans, Curtice and Norris (1997) “New Labour, New Tactical Voting?” - https://www.tandfonline.com/doi/pdf/10.1080/13689889808413005
For a detailed explanation of how this figure was arrived at, see this analysis from John Curtice, Stephen Fisher and Patrick English: https://electionsetc.com/2021/12/10/what-impact-did-the-brexit-party-have-in-the-2019-general-election/
Analysis by John Curtice found evidence of substantial unionist tactical voting in the 2021 Scottish Parliament election. https://whatscotlandthinks.org/2021/05/the-constitutional-question-dominates-how-scotland-voted-in-2021/
There are also many seats where a local Remain split is possible next time, as the Unite to Remain agreement of 2019 unwinds. The Unite to Remain alliance of 2019 made few waves, and is now largely forgotten, but it resulted in the Liberal Democrats, Greens and Plaid Cymru co-ordinating candidates across 43 seats. Many local contests will therefore see one or more generally centre-left and pro-Remain parties standing a candidate next time having sat out the last contest, which could also result in votes being split off from other local contenders (including Labour, who did not participate in the Unite to Remain alliance). However, in most cases the parties which stood down had little or no local strength, so the disruption capacity of Remain splits is much lower.
I'm curious how many 'North Shropshire's there might be - seats where Labour came 2nd in 2019, with labour votes previously lent to Lib Dems returning 'home' after the Coalition, but Labour believe (with some reason) their 2017 or 2019 results were as good as they got, and the only chance of a non-Tory would be a Lib Dem vote. This would probably include suggesting activists work in more likely areas. There's clearly a belief in both Lib Dems and Greens (principally for Council seats) that they can attract wider coalitions in such areas.