Ballot Box Briefing #2 – Scandinavian Style Proportional Representation

Initially published April 2020, updated September 2021.


Folk who’ve been following Ballot Box Scotland’s output for a while will have noticed the regular appearance of a hypothetical “Scandinavian Style Electoral System”, most commonly as a projection for Scottish Parliament polls. Although the basics of how the system works are explained in the graphic and I’ve made a few attempts over the years to describe the workings via posts, there is quite a lot to it, both in terms of the mechanics and why it’s preferable to Scotland’s current electoral systems. Giving it the depth of explanation it deserves seemed an obvious topic for the second Ballot Box Briefing.

As the name implies, this system is modelled on how (parliamentary) elections work in the Scandinavian countries. Each of those countries (plus Iceland) has a unique implementation, but they all follow the same basic principles. Unlike other forms of PR which rely either entirely on subdivisions of the whole area (which negatively impacts proportionality) or give the national vote primacy over local votes (which negatively impacts local representation), the general Scandinavian model combines the direct election of most MPs within local districts with a top-up allocation of MPs to those districts via the national vote. This delivers a highly proportional result whilst still preserving something of a local connection.


The briefing looks at Scotland’s current electoral systems in order to emphasise why adopting a Scandinavian model might be better, whilst acknowledging there is no such thing as the perfect electoral system. It ranks each system (both existing and proposed) according to four criteria;

  • Proportionality
    • How accurately does it reflect the political diversity of the electorate?
  • Locality
    • How local is the electoral geography relative to the electorate?
  • Utility
    • How many votes count towards electing a representative?
  • Specificity
    • How much choice do voters have between candidates and parties?

For each of these criteria, the systems are scored from “Very Poor” to “Poor”, “Fair”, “Good” or “Very Good”, in a way that is definitely completely objective and won’t be hotly contested by fans of one system or another. Apart from one system on one of the criteria, none of the three systems Scotland currently uses (First Past the Post, the Additional Member System, and Single Transferable Vote) rank better than “fair” on any of these criteria.

The thrust of the briefing then focuses on applying a Scandinavian model to Scottish Parliament elections. The proposed system would see Scotland divided into 13 PR Districts, compared to the 73 FPTP constituencies in 8 PR Regions we currently have under AMS. These districts largely follow council and/or historic county boundaries, and would elect a variable number of MSPs each, from 5 in the smallest to 14 in the largest. Most seats in each district are elected based on the votes there, with one or two held back to be allocated based on the national vote for all parties winning at least 3%. These nationally determined MSPs are then allocated to the best fit district for their party at that stage in the process. A final count of personal votes for candidates within parties would then identify which specific candidates from each party take up the seats won.

In Practice

If we apply these principles to the 2021 election, we’d see a result much more closely aligned with the overall number of votes for each of the major parties. Rather than the SNP coming up one seat short of a majority, they instead fall 10 short. With 11 Green seats, there would still be a majority for the cooperative government formed after the election, but the constitutional balance of the chamber would be a tight 66:63, rather than the overinflated 72:57.

Further examples of adapting the system to UK Parliament and Local Council elections are also included, in recognition of the fact that using a different system for every election is confusing for voters and makes it harder to engage. Overall, this system would be more proportional, make more votes count, and allow greater voter choice than any of Scotland’s existing systems. Only on the criteria of locality does a Scandinavian style system fare worse than any of the existing systems, being substantially less local than FPTP and marginally less than STV.

Of course, it’s unlikely that this kind of reform would actually be pursued. There isn’t the consensus for it at Holyrood (which controls both its own and local council electoral systems), and the majority at Westminster remains resolutely opposed to bringing that institution into the mid-20th Century, never mind into 2020. But it’s an interesting topic to take a look at! This briefing is quite lengthy, so I’d normally say it’s definitely one just for the most hardcore electoral system nerds. But since we’re in the midst of a lockdown, you never know who might be prepared to lose themselves in a detailed description of a voting system!

You can read the full Scandinavian Style Proportional Representation briefing here.


  1. I strongly disagree that any political party only acheiving 3% support should have ant member elected. If needing to apply a minimum, it should be at least 10% of popular vote, if not more. I would point out here that 3% is lower than “lost deposit” minimum.

    • Deposits are an undemocratic nonsense as well. A 10% threshold would be on par with Turkey, which has become a basically authoritarian regime. No credible form of proportional representation anywhere else uses such a high threshold. Anyone suggesting such a high threshold is fundamentally at odds with genuine democracy.

  2. Appreciate your work. The counter argument i’d make is that true proportional representation systems can have unintended consequences.

    1. In our current model, having list & constituency MSPs clearly creates two classes of representative. I know this isn’t the case for the Scandi system, but….

    2. … People don’t associate with representatives not directly elected the way they do with the winner of the constituencies.

    3. The Scottish body politic hasn’t shown the maturity to accept the inevitable coalitions that PR brings. Particularly in the very multipolar Holyrood environment, the fact there are so few politically acceptable coalitions is a massive problem when it comes to putting together a viable coalition.

    4. Pure PR systems intend to motivate voters into feeling that their vote ‘counts’ – but can also have an unintended side effect that if not much changes parliament to parliament because only a few seats change hands at all, this can depress turnout and also exaggerate other problem factors in democracies – only certain groups of people seeing any value in voting.

    5. Member/list systems mean that Parties – not voters – have control over who gets to be a representative of the ‘people’. This effect is even worse with the Scandi system.

    6. Having no lower % limit to representation in Parliament means you’re at the whims of the effects of loony fringe parties.

    I’m no FPTP fan, i should say. I’m not sure what the answer is to the problem, but like you, i’m interested in working it out.

    • Rob,

      Without intending to be cheeky, I’m not sure you fully read the piece!

      In response to 5, it’s pretty clearly stated this is an open list system, i.e. one where voters are able to vote for individual candidates, not a closed list system where they are elected in pre-determined order. Also, it’s a complete myth of UK politics that single-member constituencies result in a personal mandate. Most people vote on party lines. The party picks the candidate. There is no voter power in having to accept that one candidate or vote for another party. I really cannot fathom how this myth persists.

      In response to 6, it says there’s a 3% threshold.

      I really don’t think there is any evidence that PR depresses turnout. In fact I’m pretty sure the evidence base is the opposite.

      I quite agree our political culture is not entirely prepared for this model. The aim should be to change that.

      • While I agree that people vote for parties first & MPs second, I don’t agree with the assumption that Constituency MPs or MSPs don’t get some sort of a personal mandate – name recognition of your local FPTP MP or constituency MSP is generally higher than of list MSPs for instance, and whether its archaic or not, its built into our political consciousness of how we relate to our elected officials, the idea that they represent an area.

        I’d also argue that as well as elections deciding who _does_ represent an area, they also have a role to play in deciding who _doesn’t_. Tactical voting to oust an unpopular incumbent and gaming the system accordingly is often the only viable tactic for an electorate, when we neither have a parliamentary standards agency with any teeth, or a recall mechanism with any value. Negative voting might not be savoury, but many people vote primarily to keep an individual or party out. The safety net of a list, controlled by a party, means that in most cases, an electorate en masse can’t really reject who they consider to be an unsuitable candidate. Open list is a start; but requires an informed or engaged electorate

        I recognise that its (probably) impossible to square the constituency link with achieving maximum proportionality in a system, but focusing purely on translating vote percentages into a near-perfect equivalence of seats risks losing focus on what you want out of a Democratic system – representation. Of course, Democracy should be more than a cross in a box every four-five years, but for many disengaged people, it is not.

Comments are closed.