14MAY

Wales Greens fall from 10 to 2

4 min read

20:05UTC

YouGov's final Senedd MRP, run 25 April to 4 May on more than 4,600 Welsh adults, projects the Wales Green Party at 2 seats tomorrow, both in Cardiff. The same model showed 10 three weeks ago.

← Back to Mutiny in week one Jump to analysis ↓

PoliticsAssessed

Key takeaway

YouGov's final Senedd MRP drops Wales Greens from 10 projected seats to 2 under D'Hondt allocation.

YouGov's final Senedd MRP projects the Wales Green Party at 2 Senedd seats tomorrow, both in Cardiff, on fieldwork run 25 April to 4 May 2026 across more than 4,600 Welsh adults. The same model showed The Greens at 10 seats three weeks earlier , enough for a Plaid-Green coalition to clear the 49-seat majority threshold. The 8-seat collapse has happened before any vote has been cast.

Wales votes tomorrow under closed-list proportional representation, the first British election to use it. Each of the 16 six-member constituencies allocates seats by the D'Hondt method, a divisor formula. A worked example shows the squeeze. In a constituency where Plaid Cymru wins 33% and The Greens 5%, D'Hondt awards seat one to Plaid (33 / 1 = 33), seat two to Plaid (33 / 2 = 16.5), seat three to Plaid (33 / 3 = 11), seat four to Plaid (33 / 4 = 8.25), seat five to Plaid (33 / 5 = 6.6), and seat six to whichever runner-up clears 5%. The Greens' 5% clears the entry threshold once, at most. Plaid hoovers up the consolidating left vote because each successive seat awarded shrinks Plaid's effective tally only marginally, while the Green tally is exhausted on the first allocation.

Plaid Cymru projects at 43 seats on 33% of the vote, Reform UK at 34 seats on 29%, and Welsh Labour at 12 seats on 12% ¹. Plaid plus Labour reaches a majority in 89% of YouGov's simulations. The Plaid-Green route to the threshold , which earlier briefings treated as the kingmaker pathway, now yields about 45 seats in the best scenarios and falls short. More in Common's final Senedd MRP ties Plaid and Reform at 34 each, with Labour third on 14 superseded) ². Both models reach the same destination: the Welsh left bloc has consolidated into Plaid. The Wales Governance Centre had already framed this in mid-April as 'consolidation, not conversion' , meaning Welsh voters did not migrate ideologically; they regrouped tactically inside the same bloc because the new system made it arithmetically rational.

The knock-on flips the coalition map. Rhun ap Iorwerth governs with Welsh Labour, or in a hung Senedd. Plaid plus Labour deliver a 55-seat working majority; Plaid plus Greens stop at about 45.

Deep Analysis

In plain English

Wales used a new voting system for the first time in this election. Instead of voting for a local candidate, voters chose a party, and seats were handed out based on each party's share of the vote in their area. This method is called D'Hondt and it works like a maths formula that progressively reduces the effective value of each seat a party wins. A party on 33% keeps winning seats because each new seat only slightly reduces its score. A party on 5% wins one seat but then falls behind almost everyone else in the queue for the remaining seats. The result: when two parties are aiming at the same voters, the bigger one hoovers up the seats and the smaller one ends up with almost nothing. Plaid picked up about four percentage points in the last fortnight of the campaign, mostly from people who had said they would vote Green. The Greens lost those four points. On the D'Hondt formula, that was enough to move eight seats from the Green column into Plaid's, and that happened before a single vote was cast, purely based on polling.

Deep Analysis

Root Causes

1. **Government of Wales Act 2006 and subsequent reform process.** The Senedd Cymru (Members and Elections) Act 2024, following the work of the Senedd Commission, replaced the 60-seat mixed-member system with 96 closed-list seats across 16 six-member constituencies. The reform removed the personal-vote element that had partially insulated smaller parties under the old system.

2. **D'Hondt divisor mechanics in small constituencies.** Six-member constituencies under D'Hondt produce effective thresholds near 12% per constituency (Wales Senedd Research Service, 2026 briefing). A party on 5% can win a single seat in a good constituency but zero in a bad one; the overall collapse from 10 seats to 2 reflects the variance in constituency-level performance once the divisor sequence works against a dispersed 5% share.

3. **Wales Green Party as a standalone party since 2025.** The Wales Green Party only separated formally from the England and Wales Green Party in 2025, meaning it contested its first Senedd election under a new brand with a new membership base and limited local infrastructure. The party lacked the ward-by-ward candidate networks that allow a party to concentrate vote in high-probability constituencies.

4. **Plaid's explicit final-fortnight vote-management messaging.** Carrie Harper's 'let Reform in' framing in April was a vote-management message targeted at the portion of Green support that placed beating Reform above expressing a Green preference. Under closed-list PR, that message was arithmetically rational for the voter to act on, which Plaid calculated and the Greens did not counter effectively.

What could happen next?

Consequence
Rhun ap Iorwerth's route to First Minister runs through Welsh Labour arithmetic; the Plaid-Green kingmaker pathway that earlier projections treated as primary has closed in 89% of YouGov simulations.
Immediate · 0.85
Precedent
Wales is the first British jurisdiction to run closed-list D'Hondt at devolved level; the Green collapse will be cited in any future English PR debate as evidence of what the system does to junior partners in ideological pairs.
Long term · 0.8
Risk
Anthony Slaughter's Wales Green Party faces a credibility question at its first ever Senedd election: 2 seats on 8% of the vote is a survivable result but not the kingmaker outcome the party publicly projected.
Short term · 0.75

Source Landscape

This story draws on neutral-leaning sources

Primary parallel: Spain's 1996 and 2000 general elections under D'Hondt in multi-member constituencies produced the same dynamic. Izquierda Unida (IU), the left coalition standing to the left of the PSOE (Spanish Socialist Workers' Party), won 10.6% of the vote in 1996 but only 21 of 350 seats, while PSOE's 37.5% yielded 141 seats.

The ratio of seats-to-votes was 2.5:1 for PSOE and 0.7:1 for IU, a structural penalty the D'Hondt divisor imposed on the smaller progressive party. Belgian electoral researchers at the Universite Libre de Bruxelles documented the same pattern across their six-party system: in multi-member constituencies below eight members, bloc-junior parties consistently win fewer seats per vote-share point than the bloc leader.

Counter-parallel: Scotland's Additional Member System (AMS) for Holyrood regional list seats softens this effect materially. The Scottish Greens have held regional list seats since 2003 precisely because the regional list top-up corrects for FPTP constituency disadvantage; the 2026 projection of 16 Holyrood seats for the Scottish Greens contrasts directly with the Welsh Greens' 2-seat collapse under closed-list PR.

AMS rewards parties that can mobilise a dispersed 5 to 8% across a wide region; closed-list D'Hondt in six-member seats rewards the strongest party in each local cluster.

Consensus view: The Wales Governance Centre at Cardiff University framed the final-fortnight shift in mid-April as 'consolidation, not conversion': Welsh voters did not change their policy preferences; they regrouped tactically inside the progressive bloc because D'Hondt made fragmentation arithmetically costly.

Cardiff University political scientist Richard Wyn Jones has written that closed-list PR in six-member seats produces a natural consolidation gravity, meaning a party on 5% cannot survive the divisor sequence once a rival on 33% has accumulated multiple seats in the same constituency.

Counter-view: The Welsh Senedd Research Service noted in its March 2026 briefing that a single-seat threshold under D'Hondt in six-member constituencies sits at roughly 12% of the constituency vote, not 5%, meaning The Greens were not categorically squeezed from the system but from competitive parity with Plaid.

Open-list systems, as used in Sweden's Riksdag and the Netherlands' Tweede Kamer, partially buffer this effect by letting voters redistribute preference within a party bloc rather than abandoning it entirely; closed-list PR removes that buffer.

Key tension: Whether the 8-seat Green collapse reflects a durable feature of D'Hondt under closed-list rules or a campaign-specific artefact that a better-resourced Wales Green Party could overcome in 2030.

First Reported In

Update #6 · 1 Days to Go: D'Hondt squeezes Welsh Greens

YouGov· 6 May 2026

Read original →

Causes and effects

Caused by

Wales Greens projected as Senedd kingmakers

YouGov's first Senedd MRP (Green 10 seats) was superseded by the final MRP projecting only 2 Green seats, an 8-seat collapse driven by D'Hondt squeeze

Occurred 11 Apr 2026

Read story →

YouGov second Welsh MRP flips Plaid lead

YouGov's second Senedd MRP (Reform 37, Plaid 36) was reversed by the final MRP (Plaid 43, Reform 34), flipping the projected coalition arithmetic

Occurred 21 Apr 2026

Read story →

This Event

Wales Greens fall from 10 to 2

An 8-seat Green collapse before any ballot is cast is the first electoral-system event of the 2026 night, and it shows how D'Hondt allocation rewards bloc consolidation over pluralism in the first British election to use closed-list PR.

Led to

Plaid Cymru forms Welsh minority government

YouGov final Senedd MRP's D'Hondt compression of Wales Greens to 2 seats was confirmed by actual result, enabling the Plaid minority government arithmetic

Occurred 7 May 2026

Read story →

Different Perspectives

Russell Findlay (Scottish Conservatives)

Findlay refused to resign as Scottish Conservative leader after the party fell to 12 Holyrood seats and lost all five constituency MSPs. He declined Swinney's post-election talks invitation, the only major-party leader to do so.

John Swinney (SNP)

Swinney committed on 14 May to a Holyrood Section 30 vote within a week despite winning seven seats fewer than his own trigger threshold, relying on a SNP-Green majority of 73. He tabled a meeting with Starmer; Downing Street disputed that any referendum discussion was agreed.

Rhun ap Iorwerth (Plaid Cymru)

Ap Iorwerth was sworn in as First Minister of Wales on 12 May, the first non-Labour head of the Welsh Government since 1999. He governs as a minority without a written Green confidence-and-supply agreement, his cabinet entirely Plaid.

Richard Tice (Reform UK)

Tice framed the Harborne £5 million gift as an unconditional personal security payment, citing milkshake incidents and the 2025 firebomb attack on Farage's home. Reform's position is that the Standards Commissioner investigation is politically motivated.

Wes Streeting

Streeting resigned as Health Secretary on 14 May, writing that Starmer would not lead Labour at the next election. He had not formally filed leadership nominations as of Thursday evening, making his departure a public verdict on the incumbent rather than a candidacy.

Green Party

Zack Polanski's campaign delivered the Hackney and Lewisham mayoralties and both councils, plus 543 English council seats, establishing the first Green governing base in outer London. The 153-seat MRP undershoot was attributed to FPTP tactical dynamics in marginal wards rather than a polling error in vote share.