Wales has long been a Labour fiefdom. Keir Hardie was first elected in 1900 in Merthyr Tydfil as Independent Labour and with three others became the party’s first official MPs in 1908. Many Valleys constituencies, such as Caerphilly, have an unbroken record of Labour winning at Westminster elections. As someone with family in this part of Wales, it became ingrained that these were Labour areas (for example, I know many people didn’t actually want the Senedd but voted for it as Blair wanted it). Even in 2019, which saw former coal mining areas such as Bassetlaw desert Labour, Wales and the Valleys remained strong for them.

In the Senedd, Labour has governed alone or in coalition since 1999. Last year’s Caerphilly Senedd by-election was therefore a canary in the coalmine for Labour in Wales. It wasn’t just that they lost but that they collapsed from 35% to 11%

Recent polling has backed up the danger for Labour with a December YouGov Senedd poll showing Labour on only 10% and a recent Find Out Now poll putting them on 12%.

This article will look at the worst case for Labour.

The new voting system

When Labour set up the Assembly in 1999, they picked a voting system that was to their advantage. This had 40 constituencies based on the Westminster seats and 20 top-up seats. Crucially, the top-up seats were split into 5 regions, meaning that Labour could still win seats in their weakest regions like Mid and West Wales. This meant that in 2016, for example, Labour won 48% of the seats on 32% of the regional vote.

For the 2026 election, Labour and Plaid changed the voting system to something that could be described as “sort of PR but not really”.

In a typical PR system, seats are tied closely to the votes cast for each party. For example, in the Netherlands, there are 150 seats meaning a party only needs 0.67% of the vote at the national level to win a seat. The downside of this is that 15 parties won seats in the 2025 election, requiring a multiway coalition. To counter this, many countries have a minimum threshold to win seats, e.g. 4% in Sweden.

In the new Welsh system, the 32 Westminster seats are paired to give 16 constituencies with 6 Senedd members each. The Boundary Commission decided to use Welsh only names while avoiding words like “and” and “North” as English speakers might not understand them.

	Westminster Constituency 1	Westminster Constituency 2
Afan Ogwr Rhondda	Aberafon Maesteg	Rhondda and Ogmore
Bangor Conwy Môn	Bangor Aberconwy	Ynys Mon
Blaenau Gwent Caerffili Rhymni	Blaenau Gwent and Rhymney	Caerphilly
Brycheiniog Tawe Nedd	Brecon, Radnor and Cwm Tawe	Neath and Swansea East
Caerdydd Penarth	Cardiff West	Cardiff South and Penarth
Caerdydd Ffynnon Taf	Cardiff North	Cardiff East
Casnewydd Islwyn	Newport East	Newport West and Islwyn
Ceredigion Penfro	Ceredigion Preseli	Mid and South Pembrokeshire
Clwyd	Clwyd East	Clwyd North
Fflint Wrecsam	Alyn and Deeside	Wrexham
Gwynedd Maldwyn	Dwyfor Meirionnydd	Montgomeryshire and Glynd?r
G?yr Abertawe	Gower	Swansea West
Pen-y-bont Bro Morgannwg	Bridgend	Vale of Glamorgan
Pontypridd Cynon Merthyr	Merthyr Tydfil and Aberdare	Pontypridd
Sir Fynwy Torfaen	Monmouthshire	Torfaen
Sir Gaerfyrddin	Caerfyrddin	Llanelli

With 6 seats per constituency, the theoretical minimum vote percentage to get a seat is 17%, however, it is likely to be lower than that due to wasted votes.

Seats are allocated by d’Hondt PR using the formula Seats=Vote share/(existing seats+1) e.g.

I’ve done some experiments, and with 6 parties, the threshold to get a seat varies between 9% and 14%. The key thing is that it is not just about a party’s vote share but also opposition parties’ vote share. For example:

• Party A 40.9%, Party B 27.2%, Party C 13.5%, Party D 13.4%, Party E 3.0%, Party F 2.0% (A 3 seats, B 2 seats, C 1 seat)
• Party A 27.3%, Party B 26.6%, Party C, 18.2%, Party D 9.4%, Party E 9.3%, Party F 9.2% (A 2 seats, B 2 seats, C 1 seat, D 1 seat)

So, in the examples, Party D’s result also depends on other parties’ votes – finishing just ahead or just behind another party can make all the difference.

So what does this mean for Labour?

The YouGov poll had Labour on 10% level with the Conservatives and just 1 point ahead of the Greens. Jac Larner has a calculator tool for the overall result and each seat:

Jac Larner has a calculator tool for the overall result and each seat here

Based on UNS, the YouGov poll would give Labour 8 seats, with their highest vote share of 18.8% being in Afan Ogwr Rhondda. However, there is an issue with using UNS.  Due to the large swings taking place from 2021, Labour would end up getting 0% in some of their weaker constituencies such as Gwynedd Maldwyn using UNS.

If instead we switch to using proportional swing, then Labour gets 6 seats with a high of 13.9% in both Clwyd and Casnewydd and Islwyn.

How low would Labour have to go for 0 seats? Using the calculator, I took the YouGov results and then gradually took vote share away from Labour and gave to the Greens. The result I got was that Labour get 0 seats on a 7.5% vote share.

Could it really get that bad?

Labour are still dropping like a stone in UK polls, while the Greens look set to pass them.  It could be even worse by May.

In addition, Labour have also had their problems in the Senedd with 3 first ministers this term and unpopular policies like the 20mph speed limit.  And of course, they have been running Wales since 1999, so we are well overdue for a change.

A benefit of the old system was that it allowed Labour to squeeze parties like the Greens in the constituencies (last time Labour got 40% in the constituency vote and 36% in the regional vote). In a PR system, a message like “Vote Labour to stop Reform” doesn’t work any more as any non-Reform vote stops Reform (as long as the party is above the seat threshold). Labour are going to have to rely on a purely positive message as to why voters should re-elect them.

Finally, Labour should be worried about turnout. At the 2021 Senedd election, turnout was 46.6% compared to 56.0% for the 2024 UK General Election in Wales. Differential turnout could really punish Labour if non-Labour supporters are much more motivated to vote.

Conclusion

It seems unlikely that Labour will actually get 0 seats, but even if they ended up with, say, 5 seats, that would still be an unmitigated disaster for the party. The danger is that Labour get squeezed between an anti-immigration Reform on the right and the more left-wing options of Plaid and the Greens.

So long a stronghold in even Labour’s worst years, the Red Dragon could torch what’s left of Sir Keir Starmer’s premiership.

Gareth of the Vale 2