Long Read – Ballot Box Britain: GE 2017 Under PR

As something a bit different for a pre-Election long read, I take one last look back at the 2017 General Election. It’s a rare whole-UK moment for Ballot Box Scotland, as we take a trip through the alternate reality of having used Proportional Representation instead of First Past the Post the last time we elected the House of Commons. The start of the post is very word heavy, but you get pretty maps once you’re past that.

Basic Principles

When talking about electoral reform in the UK it’s usually in a relatively abstract sense, that “seats should match votes.” When demonstrating how PR might work, we then reach for over-simplifications like just taking the UK wide votes and giving every party their share. That’s fine as far as it goes because it isn’t necessary to come into everything with a detailed scheme – but nerds like me like to come up with schemes. Very few countries use such simple systems of PR, and certainly none as big and diverse as the UK, so a look at an actual concrete system might be useful.

Long-time followers of Ballot Box Scotland will know of my fondness for “Scandinavian” style electoral systems. Apart from countries like the Netherlands which simply vote as one huge bloc with no subdivisions, they are the most proportional electoral systems going. I’m so keen on them that I run an example version translated to the Scottish Parliament as a regular feature of my polling coverage. I’d never gotten round to doing a similar translation for Westminster, which is more complex than either Scotland or the countries I’m drawing my inspiration from. With the next GE just a week away and filled with the weird “do something, but not anything useful” energy that can come with ennui, I’ve finally done a quick version.

As with all my implementations of this type of system, it leans specifically towards Norway’s version, which means;

  • The country is divided into a number of districts* which each elect a number of MPs via proportional representation based on votes cast in that district
  • One seat in each district is held back to be allocated based on the overall share of the vote above a % threshold (“levelling seats”)
  • Parties (or independents) keep any seats they win in districts even if they don’t cross the overall % threshold
  • Lists are “open”, allowing voters to specifically vote for a candidate rather than just accept the list order presented by a party

* Note that in practice, we’d almost certainly still use the term constituency. I’m using “district” within this post to differentiate from our existing single-member constituencies.

The big difference versus the countries that originated this basic system is that the UK is much larger (circa 65m people versus the around 10m in Sweden and 5m in Denmark and Norway) and more diverse. So instead of using the vote across the whole UK to allocate the levelling seats, I’m treating Scotland, Wales and Northern Ireland individually, then doing the same in each of England’s 9 regions.

This ensures that already (or historically) represented parties like Plaid Cymru and the various Northern Irish parties aren’t accidentally thrown out of their respective countries for obvious lack of UK-wide support, as well as providing opportunities for new parties to break through in England by targeting specific regions.

It’s also worth highlighting the bullet that says it is possible to win seats without crossing the threshold if a party or candidate performs strongly enough in a district. This is in line with the Norwegian example, where for example the Green (3.2%) and Red (2.8%) parties have a seat each in parliament for Oslo, despite being below the 4% threshold.

This is quite substantially different from either STV or AMS, the two systems people most often suggest in the UK. Again if you’ve been following Ballot Box Scotland for a while, you’ll know I’m an STV-sceptic. It’s better than FPTP of course, but it isn’t actually very proportional and doesn’t minimise wasted votes or maximise voter choice between candidates to the extent commonly believed. (See here for more detail on that if you want it.)

AMS by contrast is likely to be more proportional than STV, but Scottish experience tells us two things. Firstly, that tiresome stuff about tactical voting doesn’t go away, it’s just redirected. When one party is likely to win so many constituencies they won’t get list seats, you get tactical list vote chat. The other major issue is that two “tiers” of representative confuses voters about who represents them, and leads to some people declaring one of those tiers (invariably the PR list one) is illegitimate. Again, I’d skip merrily to the polls to vote for AMS in a referendum if that was the only option, but it has issues.

Exact Process

Within each nation/region;

  • Allocate all but one seat in each district via Sainte-Laguë method
    • This divides each party’s vote share by one more than twice the number of seats they’ve won so far
    • E.g. after a party wins one seat, their vote share is divided by three. After two seats, it’s divided by five, and so on
    • Using this rather than D’Hondt given the relatively small size of districts
  • Once all directly elected seats are known, apply Sainte-Laguë again to the overall vote share for all parties above 3%, accounting for district seats already won
  • For each successive seat, allocate it to the district with a seat still remaining to fill where the party in question came closest to a(nother) seat via the simple quota
    • The simple quota is the total number votes cast in that district for above-threshold parties divided by the total number of seats in that district
    • E.g. In a district with 100,000 votes above threshold votes and 5 seats, the quota is 20,000 votes.
    • A party that has already won 1 seat in that district and has 32,000 votes therefore has 0.6 of a quota towards a second seat
  • Repeat process until all seats have been filled

I usually use a threshold of 4%, but a combination of recently discovering that the Parliamentary Assembly of the Council of Europe recommends 3% as a maximum and UK polling suggesting a couple of smaller parties might roughly achieve that has led me to settle on 3% here. For the purposes of illustrating how different things could be with PR, widening the net a bit to capture those smaller parties is a useful exercise.

One final process note doesn’t relate to seat apportionment but fits best here. I’m also calculating the Gallagher Index for each region, under both the real FPTP and projected PR results. This is a measure of how proportional a result is. The closer the number is to 0, the more proportional.

District Design

As you’d expect from Ballot Box Scotland, the most detailed work has went into Scotland. Balancing electorates between single member constituencies as required in First Past the Post creates some odd creatures, never mind that our boundaries date back to 2005 and the building blocks are ward boundaries dating to, I kid you not, 1999. So I built Scotland’s PR Westminster seats from scratch using whole council areas as building blocks, allocating an electorate appropriate number of seats to each district, and using 2017 council election results to work out a very rough way to divide votes between existing constituencies which are split between these proportional districts.

I don’t have anything like the knowledge of English, Welsh or Northern Irish geography necessary to feel confident doing the same thing there, so I just stuck existing constituencies together. In England’s case I actually did the sticking together as a project way back ahead of the 2015 election. It largely follows Ceremonial County lines so that might make it not entirely terrible. For Wales and NI though there might just be some awkward moments. If you have a particularly strong objection to a given district, do remember I’m not the Boundary Commission and I’m not going to impose this on you – it’s just a for-fun experiment.

Most districts have between 4 and 9 seats. The three exceptions are one Welsh district which has 3 because I didn’t feel like the component seats could be put anywhere else, and Scotland’s two islands constituencies remain as-is with only one seat each. All votes in the island constituencies still count towards allocating national seats however, so it’s not simply retaining first past the post there. This gives a total of 105 districts, down from 650 constituencies, with an average of 6.2 seats per district. In reality my preference would be for larger (for example I split Devon into two 6’s rather than a single 12) but I acknowledge even these small (by PR standards) districts would a challenge to sell to British voters.

As a final note, that’s why levelling seats are often the crucial bit for delivering proportionality here. Spain (mostly) uses similarly small districts, but since all MPs are elected directly from them with no levelling seats, the proportionality overall can be quite badly off – though using Sainte-Laguë rather than D’Hondt does take a lot of that edge off.

Caveats

The big caveat with any translation of UK General Election results from First Past the Post to any system of Proportional Representation is that it is absolutely guaranteed that voters would make very different decisions under PR. Apart from reducing the rates of tactical voting overall, smaller parties in particular would almost certainly do much better under PR. Scotland gives us credible evidence of this, as the UK Parliament is the only election out of four here that doesn’t involve PR.

The best party to demonstrate this are the Greens, who typically have the highest contest rate outside of the 4 Westminster parties. Looking at Glasgow, for example, across four elections in which every voter had the option to vote Green in recent years;

  • 2015 Westminster – 2.8% of votes, 1.7% of electorate
  • 2016 Holyrood List- 10.2% of votes, 4.7% of electorate
  • 2017 Council – 8.7% of votes, 3.3% of electorate
  • 2019 EU – 12.1% of votes, 4.3% of electorate

Even in the proportional election with the fewest Green voters as an overall share of the electorate, it was still twice as many as the First Past the Post election. This will almost certainly hold true UK-wide and for other parties, most prominently UKIP (in 2017) and the Brexit Party (2019).

Not only would they make different decisions, in many cases they would actually have the option to make decisions they couldn’t have under FPTP. That UKIP and the Greens each only won 0.2% of the vote in Scotland in GE17 might initially seem shocking. But when you remember that out of 59 seats, UKIP only contested 10 and the Greens 3, it goes some way to explaining how they were that low. Across the UK, Greens were present in 70% of seats and UKIP in 58%, whereas it’s hard to imagine a form of PR where they wouldn’t have been able to stand everywhere.

In the same vein, politicians and parties would differ greatly with PR. If we’d brought in PR decades ago, Jeremy Corbyn is far more likely to have been the leader of a dedicated Further Left party (a la the Socialist Left Party, to keep to our Norwegian parallels) with seats in parliament than Labour, which would have been a narrower grouping of mainstream social democrats. Similarly, plenty of people currently in the Conservatives might instead have joined UKIP were it electorally viable – Jacob Rees-Mogg comes to mind. And rather than go through dramatic periods of infighting and tiny, inconsequential splinters we’ve witnessed recently, PR also makes it safer for parties to split without the worry they’ll hand the next election to their opponents.

Nonetheless, there isn’t any credible way of estimating how voters would have voted under PR (either in 2017 or next week) or what different parties might have existed, so this post operates purely on the basis of the votes actually cast.

Scotland

Starting here in Scotland, our 59 seats are spread across 10 multi member districts plus those two awkward single member island districts. The SNP won the most votes in eight of the PR districts plus the Western Isles, whereas the Lib Dems have their only lead district in the whole UK in Orkney and Shetland. The Conservatives would have led in the other two districts, which cover some of their traditional strongholds. Votes and seats in total (vs actual 2017 seats);

  • SNP – 36.9%, 22 seats (-13)
  • Conservative – 28.6%, 17 seats (+4)
  • Labour – 27.1%, 16 seats (+9)
  • Lib Dem – 6.8%, 4 seats (nc)
  • Gallagher Index – 0.4 (19.7)

The SNP are substantially over-represented thanks to FPTP, so PR redistributes over a third of their seats to other parties. In 2017 Labour won half as many seats as the Conservatives despite only being 1.5% behind them, whereas they’d be a much more appropriate one seat behind in this system. They win at least one seat in each PR district, including areas like the Highlands and Tayside where they were quite far off winning any FPTP seats. They were also only 0.5% behind the SNP across Edinburgh, so that was nearly another splash of colour to the map.

Conservatives also pick up seats in every part of the country, thanks to a combination of a more proportional seat share and the ability to gain representation in urban areas that otherwise elude them. Notably, the Lib Dems won exactly their fair share of seats in 2017, and in this model they win them in the districts covering their real ones. Very low vote shares for any parties outside those 4 would also give this the lowest Gallagher Index anywhere in the UK.

Wales

Welsh electoral geography is the second most confusing after Northern Ireland, so I ended up with a lot of small districts and quite widely scattered Wales’ 40 seats into 9 districts. Labour come out on top in all but one of them, with the Conservatives leading in the biggest and most rural one. Votes and seats in total (vs actual 2017 seats);

  • Labour – 48.9%, 20 seats (-8)
  • Conservative – 33.6%, 14 seats (+6)
  • Plaid Cymru – 10.4%, 4 seats (nc)
  • Lib Dem – 4.5%, 2 seats (+2)
  • Gallagher Index – 2 (18.1)

Labour still rack up half of the seats in the country, winning more representation in the more rural parts of Wales at the expense of many of their seats in the urban south. The Conservatives, who are also represented in every district, unsurprisingly go in the opposite direction by picking up those urban seats whilst dropping a couple elsewhere. Plaid hit the correct number in 2017 anyway, but here they’d lose out on representation in Ceredigion and one of their seats in the Gwynedd area in exchange for seats in the southern valleys.

Finally, the Lib Dems would not have witnessed their Welsh wipeout had PR been in play. They’d win a seat in the Cardiff district as well as Ceredigion and Powys, which contains the Brecon and Radnorshire constituency they picked up in this year’s by-election.

Northern Ireland

The smallest bit of the UK by population, Northern Ireland’s 18 seats are divvied up between 4 districts. Three of them would have seen the DUP out in front, and the other would have leaned Sinn Féin. Votes and seats in total (vs actual 2017 seats);

  • DUP – 36.1%, 7 seats (-3)
  • Sinn Féin – 29.4%, 6 seats (-1)
  • SDLP – 11.8%, 2 seats (+2)
  • UUP – 10.3%, 2 seats (+2)
  • Alliance – 8.0%, 1 seat (+1)
  • Independent (Sylvia Hermon) – 2.0%, no seat (-1)
  • Gallagher Index – 4.3 (19.9)

PR would have resulted in greater balance between the Unionist and Nationalist sides, going from 11:7 for Unionists to 9:8, with the Alliance party’s single MP in their Belfast heartland not defining as either. The flip side is that Independent Unionist Sylvia Hermon wouldn’t have been elected under this system.

As the two largest parties, the DUP and Sinn Féin end up represented in every district. The SDLP and UUP who dropped out of Westminster (for the second time, in the latter’s case) in 2017 wouldn’t have with PR, both winning seats in the two border districts. Alongside Alliance, that would have given a much greater diversity to NI’s Westminster delegation, given Sinn Féin’s policy of abstentionism meant it was just the DUP and Hermon in reality.

North East England

For England’s regions, we’ll start with the smallest, the North East. I’ve bunched the 29 seats here into 5 districts. Labour have a comfortable lead in four of them, whilst the Conservatives came first in the only rural one. Votes and seats in total (vs actual 2017 seats);

  • Labour – 55.4%, 17 seats (-9)
  • Conservative – 34.4%, 10 seats (+7)
  • Lib Dem – 4.6%, 1 seat (+1)
  • UKIP – 3.9%, 1 seat (+1)
  • Gallagher Index – 2.6 (29.9)

You don’t need FPTP to be totally dominant somewhere, as Labour show here, easily winning more than half the vote and thus most of the seats. Rather than limit the Conservatives to the two rural seats in the north plus one of the Cleveland constituencies, this gives their voters everywhere representation, whilst the single Lib Dem seat is in the Northumberland district, which covers the Berwick Upon Tweed constituency they historically held.

The most interesting thing here is that it’s the only region where UKIP crossed the 3% threshold. Since it’s the smallest English region it’s only good for a single seat in the Sunderland district, but this won’t actually be the last we see of UKIP…

Under FPTP this was the least accurately represented region in the UK, so PR offers a vast improvement.

North West England

One of the largest regions, I’ve condensed the North West of England’s 75 seats into 10 districts. The two northern districts plus one in the south east tilt Conservative, and Labour win the other seven. Votes and seats in total (vs actual 2017 seats);

  • Labour – 54.7%, 42 seats (-12)
  • Conservative – 36.1%, 28 seats (+8)
  • Lib Dem – 5.3%, 4 seats (+3)
  • UKIP – 2.3%, 1 seat (+1)
  • Gallagher Index – 1.7 (14.3)

Again, with more than half the votes Labour are perfectly capable of retaining regional dominance under PR. This was more fairly balanced than the North East as it was though, so this is a less dramatic change, mostly the Conservatives winning a fairer share of seats in Manchester and Merseyside.

The Lib Dems naturally pick up one of the seats in Cumbria, where former leader Tim Farron’s seat is located, and the ghosts of MPs past also give them a directly elected seat in the south east of Manchester and the levelling seat in Merseyside North and East. They haven’t had any MPs in the area covered by Cheshire East recently, but it was their next strongest result in the region.

UKIP may not have crossed the threshold here, but Sainte-Laguë is favourable enough to small vote shares that they pick up the last directly elected seat in the Greater Manchester West district, where they won nearly 7%.

Yorkshire and the Humber

One of the few English regions that’s actually close to being natural, I’ve split Yorkshire and the Humber into 8 districts to elect its 54 seats. Labour come out on top in the five districts that make up West and South Yorkshire, whilst the Conservatives lead in the 3 covering the North, East and Humberside portions. Votes and seats in total (vs actual 2017 seats);

  • Labour – 49.0%, 28 seats (-9)
  • Conservative – 40.5%, 23 seats (+6)
  • Lib Dem – 5.0%, 3 seats (+3)
  • Gallagher Index – 3.5 (15.8)

We’re now starting a run of regions where only the classic big 3 parties win seats. As in the rest of the North, Labour lose out versus FPTP in the urban seats, but do far better in North Yorkshire, whereas the Conservative trade off is fewer rural seats for more urban.

The 3 Lib Dem seats all come from areas they have recent history in. The Sheffield district incorporates former leader Nick Clegg’s Sheffield Hallam seat which he lost in 2017. Leeds West Yorkshire covers the Leeds North West seat also held until 2017, whilst North Yorkshire includes Harrogate and Knaresborough which was Lib Dem from 1997 until 2010.

Although not shown on the map, this region had the highest vote share for “other” parties anywhere in the UK. That’s primarily the 0.8% for the Yorkshire Party – exactly the kind of regionalist party that might do well with PR, having only stood in 21 seats, but winning 2.1% across those. They won 4% in the EU elections here in May, for example.

East Midlands

Moving down to the Midlands, the Eastern end’s 46 seats are here squeezed into 8 districts. The Conservatives were the most popular party in the 6 outer ones, whereas Labour came top in the two districts right in the middle. Votes and seats in total (vs actual 2017 seats);

  • Conservative – 50.7%, 24 seats (-7)
  • Labour – 40.5%, 20 seats (+5)
  • Lib Dem – 4.3%, 2 seats (+2)
  • Gallagher Index – 3.1 (13.2)

We’ve switched from clear Labour leads to clear Conservative ones now, though the 10.2% between the two parties absolutely did not justify the 16 seat gap between them under FPTP. That narrows to 4 here, with a now-typical boost for Labour in the more rural areas.

Even at the height of Lib Dem success this wasn’t an area they did well in under FPTP, their only MP recently being in Chesterfield (here in Derbyshire North district) from 2001 to 2010. They barely registered there in 2017 though, with their strongest support now in the Bosworth constituency in Leicestershire. That’s where one of their two seats here goes, with the other in Lincolnshire just because as a 7 seater it was most friendly to lower vote shares.

West Midlands

Popping over now to the quite similar West Midlands, where I’ve put the 59 seats into 9 districts. Just like the Eastern portion, the Conservatives performed strongest in the 6 districts in a ring around a core urban area, this time of 3 districts, where Labour came top. Votes and seats in total (vs actual 2017 seats);

  • Conservative – 49.4%, 30 seats (-5)
  • Labour – 42.1%, 26 seats (+2)
  • Lib Dem – 4.5%, 3 seats (+3)
  • Gallagher Index – 2.5 (8)

A relatively modest change here versus FPTP, with that usual flow of rural seats from Conservative to Labour and urban seats from Labour to Conservative thanks to PR giving most votes real weight.

The Lib Dems have had marginally more success, and more recently, here under FPTP. The seat they win in Birmingham with this system is a reminder they held the Birmingham Yardley seat from 2005 to 2015, whilst they held a seat in Herefordshire from 1997 to 2010, so again that’s a sensible place for them to pick up a levelling seat. The Warwickshire seat is perhaps a holdover from a history of coming second in some seats there.

South West England

For the South West of England, I settled on 8 districts to spread the 55 seats across. All but one of these would have seen the Conservatives in first place – only Bristol and Avon would have went Labour’s way. Votes and seats in total (vs actual 2017 seats);

  • Conservative – 51.4%, 29 seats (-18)
  • Labour – 29.1%, 17 seats (+10)
  • Lib Dem – 14.9%, 9 seats (+8)
  • Gallagher Index – 2.7 (28.3)

There was a huge over-representation for the Conservatives thanks to FPTP so they’d be down by over a third with PR, dropping seats basically everywhere. They’d still have a majority of seats though, with Labour miles behind but represented in every district.

Even when they were reduced to a shadow of their former selves, this is the Lib Dems best region. In this system not only would they have won at least one seat in every district, all of them are directly elected seats – they don’t even need the levelling seats to ensure they get their fair share.

East England

In the East of England, I went for 8 districts to contain the 58 seats. This is a true Conservative heartland, as they easily came first in every district. Votes and seats in total (vs actual 2017 seats);

  • Conservative – 54.6%, 33 seats (-17)
  • Labour – 32.7%, 19 seats (+12)
  • Lib Dem – 7.9%, 5 seats (+4)
  • UKIP – 2.5%, 1 seat (+1)
  • Gallagher Index – 2.3 (27.1)

Going from FPTP to PR basically just means the Conservatives would lost out on seats everywhere for Labour to get their proportional share. There’s also the usual little boost for the Lib Dems, mostly in districts covering areas they recently held MPs in.

In Norfolk the party still hold North Norfolk and held one of the Norwich seats until their 2015 collapse. They also used to represent the Cambridge seat and another seat covered here by the Essex North and West district. They haven’t recently held seats in the Hertfordshire area, but did come second in the St Albans constituency that’s included in Hertfordshire West which makes that district their best % vote in the region.

Just like the North West, there’s a UKIP seat here despite them not meeting the threshold. Interestingly, although the only seat they’ve ever won at a general election (Clacton, here in Essex North and West) is in this region, that’s not where they pick up the seat. In 2017 they still managed a respectable 20% in the Thurrock seat, which is how they’d scrape the last directly elected seat in Essex South.

South East England

The largest regional (or national) subdivision of the UK, the South East’s mighty 84 seats slot into 12 districts. This is another region where the Conservatives make a clean sweep, winning the most votes in every district. Votes and seats in total (vs actual 2017 seats);

  • Conservative – 53.8%, 46 seats (-26)
  • Labour – 28.6%, 25 seats (+17)
  • Lib Dem – 10.5%, 9 seats (+7)
  • Green – 3.1%, 3 seats (+2)
  • Speaker – 0.7%, 1 seat (nc)
  • Gallagher Index – 2.1 (27)

As this is the largest region, it therefore has a correspondingly huge drop in seats for the Conservatives. They still comfortably have a majority of seats in the region, they just had to make way for everyone else to get a fair crack at the whip – Labour naturally taking the lion’s share of the PR spoils.

This is another strong region for the Lib Dems, even if they only won two MPs under FPTP (one in a constituency covered by East Sussex, the other in Oxfordshire). Their support holds up well enough here for them to win all 9 of their seats directly, without requiring any levelling seats.

For the Greens, this is what the North East was to UKIP – the only region they meet that 3% threshold requirement. They benefit from the size of the region as in addition to a direct seat in East Sussex (where Caroline Lucas’ Brighton Pavilion seat is) they also pick up two levelling seats. One is in the awkwardly named Southampton, New Forest and Isle of Wight district, where the Isle of Wight itself was their second best result in 2017. The Kent East one is nowhere special to the Greens, just being the district that had availability at that point in the process.

Finally for this busy region, it also includes what was then the Speaker’s seat of Buckingham. Based on 2017 votes John Bercow would have been re-elected, but this raises the question of how this would actually work in practice. With PR, voters might not be inclined to actually vote for the Speaker when they can cast a partisan vote.

Over in Ireland where they use PR (in the form of STV), the Ceann Comhairle is automatically deemed re-elected if they wish to stand again, and voters in their district therefore just elect one less TD to the Dáil. That might not sit well with folk here though. In the Scottish Parliament, convention has been established that a new Presiding Officer is elected by the chamber each term, and the 2007-11 session PO even successfully stood for re-election in 2011 on a party ticket. That might be a sensible option under PR, since we’re already ditching one parliamentary tradition, why not ditch the convention that the Speaker serves basically until they give it up?

London

Last but not least, London. The 73 seats here are across 12 districts, which should relatively neatly follow Borough boundaries, though that’s how they end up with big long names – shout out to Tower Hamlets, Kensington, Hammersmith and the Cities. Labour’s strength in the capital is such that they would have the most support in 11 of them, with just one favouring the Conservatives. Votes and seats in total (vs actual 2017 seats);

  • Labour – 54.5%, 41 seats (-8)
  • Conservative – 33.1%, 25 seats (+4)
  • Lib Dem – 8.8%, 7 seats (+4)
  • Gallagher Index – 2.2 (7.5)

A really modest difference here, just shuffling a few Labour seats out for the Conservatives and Lib Dems. The latter pick up a good bunch of their seats in the areas in the south west of the city where they actually have MPs at the moment, plus a couple of other districts that cover seats they previously held.

The Whole Picture

So summing up everything, across the 12 nations and regions of the UK, there are 105 districts electing 650 seats. The Conservatives won the most votes in 50 districts, Labour in 41, the SNP in 9, DUP 3, Lib Dems 1 and Sinn Féin 1. Votes and seats in total (vs actual 2017 seats);

  • Conservative – 42.3%, 279 seats (-38)
  • Labour – 40.0%, 271 seats (+9)
  • Lib Dem – 7.4%, 49 seats (+37)
  • SNP – 3.0%, 22 seats (-13)
  • UKIP – 1.8%, 3 seats (+3)
  • Green – 1.6%, 3 seats (+2)
  • DUP – 0.9%, 7 seats (-3)
  • Sinn Féin – 0.7%, 6 seats (-1)
  • Plaid Cymru – 0.5%, 4 seats (nc)
  • SDLP – 0.3%, 2 seats (+2)
  • UUP – 0.3%, 2 seats (+2)
  • Alliance – 0.2%, 1 seat (+1)
  • Speaker – 0.1%, 1 seat (nc)
  • Independent (Sylvia Hermon) – 0.05%, no seat (-1)
  • Gallagher Index – 1.9 (6.5)

Phew! That’d make for quite a busy parliament, with 12 different parties represented, up from the 8 that actually were. It’s also narrowly more than the 11 that won seats in the 2015 election. Crucially though it’s a parliament that much more accurately represents the diversity of views not just across the whole country, but within each part of it. With such a dramatically altered parliament, we might be in a very different political situation right now, and likely not even facing yet another snap election in a few days time.

If you were paying attention to all the Gallagher Index bits as we ran through each nation and region, you might be surprised to see that overall the real FPTP one was just 6.5 – lower than it was any individual area! Effectively what happened was that the massive disproportionality in Labour’s favour in the North of England was largely balanced out by the Conservative’s absolute dominance of the South.

Even if representation was just about the top line votes and seats, FPTP was still rubbish given it had poor representation of NI parties, too many SNP seats, and too few Lib Dem seats. But it also failed because it simply pretends that (most) rural Labour voters don’t exist, and vice versa for (most) urban Conservative voters. That ability to dismiss entire areas of the country as belonging to one party or another does a huge disservice to the many voters in them who back other parties.

Seat shares and proportionality indexes out the way, we can think about hypothetical governments. In all cases, I’m discounting the 7 seats in total that’d have been won by the Speaker (non-partisan) and Sinn Féin (abstentionist).

  • Conservative – DUP (286 vs 357)
    • The confidence and supply arrangement actually negotiated in 2017 would have been no good in this scenario, being miles off a majority.
  • Conservative – DUP – UKIP (289 vs 354)
    • Bringing UKIP in for the Maximum Leave coalition is barely any better, given UKIP’s small handful of seats.
  • Conservative – Lib Dem (328 vs 315)
    • There’d have been the numbers for a revival of the 2010-15 Coalition, but would the Lib Dems have been as willing having been punished for the last one and with Brexit such a pressing issue?
  • Labour – Lib Dem (320 vs 323)
    • Echoes of the 2010 general election here, as the Lib Dems weren’t solely capable of picking a governing party. They come up just short on the Labour side, assuming Labour were happy to work with them.
  • Labour – Lib Dem – Green (323 vs 320)
    • Adding the Greens to the mix for a Traffic Light Coalition would have given a narrow majority, but it’d be vulnerable.
  • Labour – Lib Dem – SDLP (322 vs 321)
    • Alternatively, Labour could have sought to bring in their Northern Irish sister party. That’d deliver an even narrower majority. They could also have pulled in both for 325 vs 318.
  • Labour – SNP – Plaid (297 vs 346)
    • Say Labour did feel able to make common cause with the pro-Independence parties, it wouldn’t have been anywhere near enough.
  • Labour – Lib Dem – SNP – Plaid – Green – SDLP – Alliance (352 vs 291)
    • This is less a governing option and more those parties that might have been willing to pursue a second EU Referendum. A comfortable majority, with hard line Labour Leavers likely balanced out by equally hard line Conservative Remainers.

So options there for either the Conservatives or Labour to have led the government if we had PR. Fans of Strong Governments Getting Things Done might look at that with alarm, but I’d remind you that FPTP hasn’t exactly been a dream in recent years in terms of stable government or policy delivery. If you’re going to have chaos (with or without Ed Miliband), far better to be because everyone’s views are (more) fairly represented than because a completely unpredictable electoral system has thrown up weirdness whilst a party with minority support tries to press an agenda they can’t even agree on amongst themselves. And as always with PR, for the “you’d never get anything done brigade!”, I remind you the rest of Europe has yet to slide into the sea as a result of using actually democratic voting systems.

And that’s it! The 2017 General Election re-packaged and re-calculated for Proportional Representation. Just in time to repeat the mammoth amounts of data entry and map making this required after polls close on Thursday! For folk who may have been hoping this post would be a projection for this election but PR, I may not have had time to do so now, but I will for the next one. Don’t worry, though, I’ll make sure I get through all the Scottish stuff before another spell as Ballot Box Britain.