Our biggest success of the May 2026 elections was undoubtedly in Scotland, where the Scottish Liberal Democrats played a blinder to reverse years of challenging Scottish Parliament elections. The Scottish Party won 10 MSPs, up from four in 2021, an outcome that is, surprisingly, our first net gain at any Holyrood election.
However, there’s a broader electoral issue that needs to be addressed. And that’s the disproportionality of the Scottish Parliament’s supposedly proportional system.
The Additional Member System (AMS) used to elect MSPs at Holyrood gives voters two ballots. One elects their local First Past the Post (FPTP) MSP, with 73 single-member constituencies across Scotland. The other allows voters to back a party list of candidates in one of eight regions. These regional seats are allocated from the list vote totals while taking into account constituency seats won by each party in each region.
This has delivered broadly proportional results since 1999. It’s never been perfect, but overall the system has been far superior to Westminster’s utterly unrepresentative First Past the Post system.
Now though, this election has exposed a significant flaw through a notable mismatch between seats won and votes cast. The SNP only won 38.2% of constituency votes and 27.2% of list votes, yet were rewarded with 58 seats (amounting to 45% of all seats available). To take this one step further, the SNP and their most likely government allies (in some form), the Scottish Greens, together won 73 seats (57.6%) on just 41% of the vote.
On one comparative measure of proportionality, the Gallagher Index, 2026 was the least representative election in Holyrood’s history by some distance.
The structural reason for this mismatch between seats and votes is largely down to the imbalance of constituency MSPs to list MSPs: 73 to 56, meaning that First Past the Post is in the driving seat, helped by the number of list seats being fixed and therefore only able to go so far to correct any disproportionality.
This flaw was exposed by the SNP dominating the constituency seats with good but not overwhelming vote shares, helped by the lack of Green candidates in most constituencies and the unionist vote being split between four main parties now including Reform UK, as well as split-ticket voting across the two ballots. This has allowed John Swinney’s party to be consistently first in the vast majority of constituencies while falling well short of a vote share that reflects that success.
Put another way, the SNP won 18 more seats than they would be entitled to if seats were allocated in each region on a purely proportional basis (the ideal outcome AMS is intended to produce), more than at any other Scottish election, according to Ballot Box Scotland.
The Liberal Democrat answer to this is, of course, replacing the Additional Member System with the Single Transferable Vote (STV), a system Scottish Lib Dem MSPs introduced in government for local authority elections in 2007. Our Scottish manifesto said this explicitly, calling on the next Parliament to
Change the voting system for the Scottish Parliament to Single Transferable Vote, mirroring the change we made for local government elections.
There are other alternatives to consider (there may be more political will for modifying the existing system by giving list seats more weight and opening them up to improve voter choice), but STV empowers voters with the ability to rank individual candidates across different parties. It ensures strong local links with multi-member constituencies. And it delivers proportional outcomes if designed effectively.
The route to replacing AMS with STV or another system is a two-thirds majority vote of MSPs (86 of the 129 MSPs). That’s a huge challenge, but not impossible. At the most recent Scottish budget, our MSPs extracted gains from the SNP; we now have an opportunity to build a cross-party consensus on electoral reform, and build on what we did with Scottish Labour when we secured STV for local government.
The scale of disproportionality at the 2026 election is coming into conflict with the principle of proportionality underpinning the Additional Member System. Liberal Democrats in Scotland must ensure that MSPs come together and reform Holyrood.
* Richard Wood is a Liberal Democrat member from Scotland who has been involved in the Liberal Democrats for Electoral Reform since 2021. He sat on the boards of Unlock Democracy (2024 - 2026) and the Electoral Reform Society (2022 - 2023).
Follow @libdemvoice.org on Bluesky
Like us on Facebook
Subscribe to our feed
Sign-up for our daily email digest
A journey along the Welsh Border in 1963
Recipe for a real change of direction
Day 9304: So Farewell Then, Kier Starmer
Dennis
Chloe
Slamdac
Nonconformistradical
Mick Taylor
14 Comments
Ultimately we should be using STV across the whole of the UK. It is the fairest electoral system which has already been stress tested across the democratic world.
You ignore the most obvious point that AMS (or, as they call it in NZ, mixed member proportional) is 100% proportional in NZ because they don’t ignore the overhang. To be 100% proportional you also need to make the party vote countrywide. Put frankly, the Scottish system is nuts.
It would be good if people leading the party actually understood the electoral systems. For a party committed to PR, the LIb Dems don’t half defend first past the post. The fiasco of the Welsh Assembly elections, where the party could have had an assembly member in half the constituencies but chose instead to have 1 in the entire country.
The party said it was going to win in Hampshire, Sussex, London, it didn’t happen. The spirit of the Liberal SDP Alliance needs to be reborn as a coalition to save this country from Farage.
After 27 years of Holyrood, we need a Royal Commission to look at:
1. The workings of Parliament and its committees to see if it has lived up to the promises made for the new body (it hasn’t) and how those processes could be improved;
2. The electoral system – the Welsh has a fresh look;
3. The number of MSPs – do we have the right number for Holyrood to do its job probably or should we have more or fewer;
4. The length of the term – it was originally four years – it’s now five years. Is that too long?
5. Who can stand for election and vote.
It should not look at the powers of Parliament (something that has had regular review) or open the Indy question – we can’t let everything in Scottish life be subservient to the divisions in Indy.
This article is absolutely correct in its overall conclusion on the failings of the PR system adopted for Holyrood and generally correct in its assessment of the causes of the problem. However, one area where it is a bit off the mark is in its unwritten belief (held by most Lib Dems) in the superiority of Proportional Representation as key to the cure to our country’s problems, and the inevitable corollary that there is an electoral PR system of voting and counting (usually STV) that will deliver what our country needs. I am not sure that this is anything like as certain as most of us believe, mainly because we rarely consider and never assess and debate the key question – What can go wrong?
In the case of Scotland’s D’Hondt system the answer is a heck of a lot because it is being gamed by the SNP and the Greens. What we have in Scotland is two parties in partial symbiotic relationship but one is also quite parasitic on the other. The symbiosis works pretty well except when it is knifed in the back by the SNP leader.
However imagine a different situation where the second party is called Alba, but the leaders are not split by a mutual visceral hatred. One says “Vote for us for the constituency” and the other says “Vote for us for the Region” and both say “… and what he says as well”. You might even call the other party SNP Reserves.
@Russell The NZ system of countrywide backup would not have made the result any more proportional.
Like you, I thought it might, but doing the calculations found that it would have actually have made very little difference: the pro-independence parties would have got the same numbers of seats, except that the SNP would not have got their one list seat, so 55.8% of seats on 41.2% of the votes for the SNP and Greens together. In terms of entitlement it is even slightly worse: the SNP should have got only 37 seats, so would have had 20 overhangs, not 18.
The principal difference making the AMS system work so differently in Scotland as against New Zealand is that we now have six parties with substantial support, where NZ support is concentrated on just two parties. I wrote an explanation of why the AMS system is likely to break down when there are many parties for LDV last December.
I agree with Richard that we should press to replace it with STV. The SNP are in principle in favour of STV, and might recognise that their current benefitting from it may not be repeated. STV is particularly good in being proportional as between two “strong coalitions”, such as pro and anti independence in the present case.
PS – Germany has had a similar increase in its number of parties, leading to similar disproportion in its last three elections. Germany’s first answer was to add extra list seats, but found that over 100 were needed in its 2017 and 2021 elections; so in its most recent election it instead disallowed 23 FPTP seats, picking on those with lowest % vote. Neither solution is satisfactory, and AMS has quite a few other problems.
@ Stephen Harte – “The electoral system – the Welsh has a fresh look”
Wales had two successive commissions/panels that both suggested STV.
Their new closed-list system is the result of a last minute hijack – I think
it was Welsh Labour’s dinosaurs.
@ David Evans – I agree that proportionality is not the only criterion.
Both the first Welsh commission and Make Votes Matter have come up with
long lists of criteria, which can be boiled down to fairness (aka proportionality),
voter choice and local connection.
I’ve written a long paper
on this, which generally strengthens the case for STV.
STV is particularly strong on voter choice / power, as opposed to party control,
which is probably the main reason those Welsh dinosaurs hate it.
I voted for the Lib Dem in Edinburgh Northern, and we’ve got a Lib Dem MSP here. However, I’m not sure if the numbers actually make sense in the article. Everyone knows the Greens didn’t stand in most constituencies. Everyone knows a regional vote for the SNP is likely a waste. The sensible way to calculate their vote share is therefore to add the SNP’s constituency share with the Greens regional share. That would be 52.2% of the votes for 56.6% of the seats. Not fully proportional, but I couldn’t agree with the stats in the article. There could be a better argument than doing mental gymnastics. A true proportional system empowers people to vote following their heart, and there would be no need for tactical voting/standing.
@Ando Du The regional vote is what is meant to determine the proportionality: SNP 27 + Green 14 = 41 %
In the constituencies SNP 38.2 + Green 2.3 = 40.5 %, which fits neatly with the simple explanation that 11 % or so of those voting Green on the list voted SNP in the constituencies. Of course there will be various voters making surprising choices, but your estimate of 52 % is double counting. The hard fact is that pro-independence parties got a maximum of 43 % support in this election, whether mesured by regional or constituency vote.
[I’m adding 2% to allow for pro-independence votes among those who voted for independents and smaller parties.]
@Denis Mollison In a pure PR system, that would indeed determine the proportionality. But, I have to respectively disagree here.
The Holyrood and its AMS system are unique in a way that a vote for the SNP on the regional ballot sadly don’t make any sense. It wouldn’t secure you an extra seat most of the time, and it wouldn’t change the seat allocation for other parties.
We have seen the trend for quite some people to split their vote not only in the SNP1 Green2 way. A very large number of voters split their votes in any given way. About 2% probably split it SNP1 Reform 2, for example. Why? You would expect the Greens to command a higher if not much higher vote share in constituencies if they stand everywhere. If you look at the polls, they’d have at least 6% in the constituencies. The fact that SNP + Green have a similar combined constituency and regional vote share can only mean one thing. A not insignificant portion of their voters split the ticket outside the camp, in combination with Lib Dem and other progressive voters choosing the Greens on the regional list to lock out Reform in urban areas. (part1)
@Denis Mollison I think there must be some double counting in my estimate, but equally the article doesn’t take account of any such split. A realistic number would be in between, but I think my number is closer taking into account all the factors. At the end of the day, it would be hard to know the result under a STV or another form of PR system. AMS makes it possible for you to split votes tactically and essentially vote in two MSPs you like. But a STV/PR poll means only one MSP. One Green and one Lib Dem is arguably better than one Green or one Lib Dem (plus one Reform) for some people including Lib Dem supporters.
Suppose we adopt a STV/PR system, the SNP and the Greens will likely have higher support than the number in the article. PR systems generally put a small penalty on smaller and especially the smallest parties. Voters take that into account when voting, because a vote still counts slightly more for larger party supporters. If I have to choose between them for the last seat on the list, I’d probably vote Green rather than Lib Dem. It’s simply the more reliable way to lock out Reform. It’s apart from my support for the Lib Dem.
That doesn’t mean STV/PR is a bad idea though. (part2 & end)
“a vote still counts slightly more for larger party supporters”
Not under STV. Small party voters may be less likely to get their first preference elected but they are actually slightly more likely to influence the final outcome. If you vote first preference for a candidate who goes over quota, you increase the value of the transfer for every voter who voted for them including some who may have given second preference to Reform. If you put small parties first and transfer at 100% as they are eliminated, you are more likely to influence the final round.
@Ando Du
“The sensible way to calculate their vote share is therefore to add the SNP’s constituency share with the Greens regional share.”
It really is not! The pro-independence vote was 41.2 % on the list and 40.5 % in the constituencies, plus around 1-2 % for minor pro-indep candidates/parties.
The discrepancies between list and constituency votes are indeed partly due to voters expecting the SNP to get ‘overhangs’ and therefore no list seats in most areas (all except H&I thi time).
But this problem is not unique to Scotland; I wrote about it on LDV last December –
https://www.libdemvoice.org/the-additional-member-system-and-its-overhang-problem-78820.html
Basically, AMS/MMP doesn’t work proportionally when you have a large number of competitive parties.
Wales has changed to a different system (though not ideal), and Scotland needs to consider it.