What would Liverpool’s local elections look like under proportional representation?

A little experiment to explore the Red Liverpool narrative

8 min readJan 4, 2021

Liverpool has long been a Labour city. Since 1997 it has returned only Labour MPs, and since 1983 just one of the city’s MPs was from a different party (David Alton, in Mossley Hill, for the Liberals/Liberal Democrats). Political pluralism isn’t much better on the local level: 72 of the city’s 90 councillors are Labour (down from a high of 80 in 2016).

It wasn’t always like this. The Conservatives used to run the city (there’s a political historian out there who’s written about this kind of stuff. I’ve heard he’s very clever, as well as funny and handsome, so you should give it a read).

I’ve always been interested in this idea of Liverpool as a Labour city when at best it is an exaggeration of reality and at worst erases huge swathes of Liverpool’s population — see the chart below.

In a similar vein is the constant talk of Liverpool as a Remain city, when more people voted leave than voted Labour in any local election since 2016! Leave voters are the great ignored in Liverpool’s politics.

So, I decided to embark on a little experiment. I’d ask people in Liverpool how they’d vote in under different electoral systems. Hypothetical surveys are not perfect by any means — people take time to unlearn habits formed under FPTP, and parties would campaign in different ways — but it’s a good starting point for understanding the city’s hidden electoral dynamics.

First Past the Post

So, let’s start with a sense of what Liverpool looks like now, using first past the post.

In 2019, Labour won 83% of the seats on 59% of the votes. The Liberal Democrats, who won around one third of Labour’s vote, won just one tenth of their amount of seats. The Conservatives also failed to win a seat, despite winning around 4% of the vote.

Nobody who knows a thing about first past the post should be surprised by these results. Labour is clearly popular within the city, but is it as popular as these figures make it out to be?(spoiler: no)

Before we go on, here’s a quick and dirty way we can check how well my survey respondents reflect the electorate:

Not too far off. Labour is pretty much spot on, Lib Dems underperform and the Conservatives over-perform. But generally, given the very unscientific sampling methods, it’s not bad going.

(Closed) Party List

The first proportional system I look at is the closed party system.

In this system, voters put a cross next to the party they support. Votes are then tallied, and seats are awarded proportionally to the percentage share each party wins.

In my model, I treat the whole city as one district. I’ve not included a minimum threshold for representation. (I’d like a 5% threshold to stop things getting too freaky but I’d never go as far as Turkey in 2002, who went full YOLO and threw up a 10% threshold — which, interestingly, made Turkey’s PR system more disproportionate than any UK general election!)

I asked respondents:

Now, imagine Liverpool used a proportional system whereby the percentage of a vote the party received was broadly equal to the number of seats it won. For example, if Party A wins 45% of the vote it would win 45% of the seats. The argument in favour of this system is that you can vote for the party you want to win, rather than worrying about keeping other parties out. If the local election was held tomorrow, using this system, which party would you vote for?

The results are presented below. Labour still storm the vote, but the Greens and the Conservatives do very well too. The lefter-than-Labour option, the Trade Union and Socialist Coalition also put up a good showing, with a solid 5%.

But how does this translate into seats?

Well, as we can see in the image below, unsurprisingly Labour haemorrhage seats but are still left with a majority of 16 seats out of 30, so they’d still call the shots. Perhaps they’d do a Saint Jacinda and go into a semi-coalition with the Greens anyway (maybe if Joe makes his brown envelopes compostable?)

The biggest winners are the Green Party. In an election with 30 seats at play, they’d increase their representation from 1 seat to 4. The Lib Dems also do better, winning an extra seat.

The Liberal Party are the real losers — their ultra-concentrated vote share in the Tuebrook ward isn’t enough to get them a seat in a city-wide election where the effective threshold for representation is about 3.3%.

Under this form of PR the Conservatives would enter City Hall, returning for the first time since 1998. The TUSC and UKIP would also take a seat.

The final diagram for this section shows how respondents moved from their first-past-the-post choice to their choice in the closed party list election.

Labour lost a significant chunk of their votes to the Greens, and a lesser amount to the TUSC. Most other parties held the bulk of their vote share, but interestingly those who wouldn’t vote in FPTP tended to move towards UKIP — perhaps not surprising given the lack of articulation of pro-Brexit voices in Liverpool’s body politic.

Additional Member System

The next voting system I looked at was the additional member system, the type used in Scottish and Welsh devolved elections, and in Germany and New Zealand’s general elections.

In this system, voters get two choices: one vote for the person they would like to represent their local ward/constituency, and a vote for their preferred party. The local votes are awarded as they would be under FPTP, whilst the party votes are allocated proportionately.

Some assumptions. I’ve kept the city’s 30 wards, and I’ve allocated 15 seats for the party-list election. This gives 45 seats, or 50% of the total council. This means elections could be held every 2 years.

Votes-wise, there is not actually much difference — Labour see a slight drop whilst most other parties see an increase.

But how would this translate into seats? Remember, the low response rate in some wards makes it hard to map out single ward-level results with much confidence, but we’re looking at something like this (accounting for ties):

And, how the council would look under the additional member system as a whole. Labour still dominate — that’s a benefit of getting more votes™️, but there are fewer wasted votes, and a much greater incentive for parties to both campaign across the city as well as focusing on specific wards.

Single Transferable Vote

Just like a successful Tinder date, we’ve reached the climax. STV. Widely hailed by many as the best electoral system, it’s recommended by the commies at the Electoral Reform Society. It’s use in Scottish local elections means that even places like Glasgow have Conservative representation on the council.

I asked respondents to

imagine each ward elected all three councillors in the same year. You must rank your party candidates in order of preference — you can mix and match parties based on your own views or how you might want to see the council balanced. (Note, it is rare for parties to stand full slates of candidates, because this can be electorally damaging). If the local elections were held tomorrow, using this system, how would you rank the following party candidates to represent your local ward?

All respondents had the choice of one Conservative, Liberal Party, TUSC, and UKIP candidate, and two Labour, Green, and Liberal Democrat candidates. Parties typically shy away from standing the maximum amount of candidates for fear of splitting their vote.

To process the STV election process, I used the R package STV (a prime example of Branding 101 if ever there was one).

I’ve made some assumptions here. I’ve merged Liverpool’s 30 wards down into 10, with each returning 3 councillors. This probably isn’t how it would be done if English local elections used STV, but we are where we are, and here’s where we are in map form:

New Ward 1 = Allerton and Hunts Cross, Belle Vale, and Speke-Garston
New Ward 2 = Childwall, Church, and Woolton
New Ward 3 = Cressington, Mossley Hill, and St Michael’s
New Ward 4 = Greenbank, Picton, and Wavertree
New Ward 5 = Central, Princes Park, and Riverside
New Ward 6 = Kensington and Fairfield, Old Swan, and Tuebrook and Stoneycroft
New Ward 7 = Anfield, Everton, and Kirkdale
New Ward 8 = Clubmoor, County, and Warbreck
New Ward 9 = Croxteth, Fazakerley, and Norris Green
New Ward 10 = Knotty Ash, West Derby, and Yew Tree

So, how does the election results look?

Or, in our trusty parliament image thing:

Once again, Labour lose seats under PR (although in some wards they might have risked more candidates and probably done quite well) but the Greens do better than they would under FPTP (obviously) and also under closed party list.

The Lib Dems do do worse than under closed party list, but then there’s not much you can do when nobody likes you. The Conservatives get 2 seats (amazingly they win the first seat in the combined ward of Central, Princes Park, and Riverside, but I think this is most likely due to the low response rate in this area, rather than a secret love for the Tories in L8…) The TUSC and UKIP both fail to make it in under STV.

Conclusion

So, if you scratch the surface of the “Liverpool is a Labour city” story, you find that… Liverpool is a Labour city, just not as much as it looks thanks to our first past the post system.

The Greens have a decent amount of support which they can’t tap into under the current system, and even the Conservatives would have a seat or two on the council. Under any of the proportional systems Labour would have an incentive to campaign in many more seats, as would other parties.

Anyway, as I said at the start, this was a bit of fun. Next up will be an analysis of constituency-level voting behaviour, and then I’ll be looking at how the Scouse identity interplays with English, British and European identities, as well as with holding populist values, and what this means for electoral behaviour in the city.

I’m on twitter as @DrDavidJeffery if you have any comments or thoughts, or you can email me at djeffery[@]liverpool.ac.uk.