Are the models that have the Lib Dems gaining a bunch of seats while getting less than their vote share in 2019 robust? It feels like these models were built primarily using data from a string of Conservative victories and they get warped once the Conservatives fall below a certain point.
I think south of the Severn-Wash line the LD would get lots of seats from tactical voting like in 1997.
North of that line it's Labour, with a bit of Reform along the UKIP coast, that would be the main beneficiaries.
Here is an example of generic Tory safe midlands seat
Spoiler alert! Click Show to show the content.
2019
CON 60
LAB 15
LD 10
Now according to polls it ought to be:
CON 30
LAB 20
REF 20
LD 15
Generic safe Tory seat in the rural south:
Spoiler alert! Click Show to show the content.
2019
CON 60
LAB 15
LD 15
Now according to polls it ought to be:
CON 30
LAB 25
LD 25
REF 10
As you see it wont take much effort to completely wipeout the Conservatives, but the wipeout would benefit different parties in different cultural regions, and tactical voting could play a role to suppress the Conservative seats numbers even more.
I'm assuming national polls can't capture movement from tactical voting then? It seems like the Lib Dems would have to rise in vote share to win those seats, given that they are projected to majorly benefit from tactical voting in 10-20% of the electorate, but it would also make sense if national polls could not pick up on that. Basically the difference between the party vote and the constituency vote in MMP.