Sunday, 27 August 2017

A Fairer Voting System?

My previous post was some time ago and I ended by saying that I'd be posting on the situation post-Brexit referendum.

This proved to be more difficult as time went on as the UK Government's position, and the position of the UK opposition, had no relationship with reality. Every statement could simply be met with a response along the lines of 'that isn't how it works'.

It has become increasingly clear that no one expected the UK to be a position where it was seeking to make a deal with the EU which would be self damaging but where none of our politicians would have the bravery to say 'Stop, this is wrong'.

So the Government blindly triggered Article 50 and then, with the clock running called an opportunistic and politically motivated General Election. The polls gave May a huge lead, one she was going to blow completely as she ran a non-campaign, meeting only party workers and eschewing the general public and television debates.

This gave Jeremy Corbyn, with  a series of large public rallies the chance to unite his party behind him and silence his critics in the Labour party.  It gave Theresa May the same, of course, but he succeeded where she failed.

As we all know now, Labour's campaign was a great success and they almost pulled off a surprise victory. Some pundits have even reckoned that had the election been a week later Corbyn would be PM by now.

From May's perspective the result was a shocking defeat as she lost her majority and had to make a deal with the DUP to stay in power.

This re-ignited my interest in voter engagement and voting systems. You can see from posts back in 2011 that I was unconvinced by the Alternative Vote system on which we had a referendum. AV is not PR, it is just a different way of electing one representative. To have PR, by any method, you need to have multi-member constituencies elected in such a way as each vote increases the likelihood of your favoured candidate being elected, and so fewer votes are wasted.

I therefore started examining different voting systems.

Voting Systems and Brexit

This was done with the backdrop of Brexit, with online Leave supporters merrily chanting the mantra that the EU was undemocratic. This is a claim which withstands no scrutiny, but Leavers are now a religious cult with Brexit their god, so reason, analysis and facts do not alter their views one iota. The result of this was of course an examination of the way Members of the European Parliament are elected, with emphasis for comparison with the Westminster, in the UK.

I have to declare some history on this subject.  Before the UK moved to electing MEPs regionally by the d'Hondt system in 1999 we used FPTP and did so by combining a geographical block of Westminster Constituencies together, and it was in a contest for one of these large single member constituencies that I had my first experience of being part of a successful election campaign. That was 1989.

So the idea I had was that, instead of each constituency electing one member, we combined a number together, kept the number of seats the same and used the d'Hondt system to allocate seats.

Why not use national voting figures?

One of the familiar arguments about PR is the old '10% of votes should equal 10% of seats' line. It's a strong argument but it ignores the regional variations of the UK, in particular the fact that none of the Northern Ireland parties stand anywhere else in the UK, that the main GB Parties, Conservative, Labour and Lib Dems do not normally stand in Northern Ireland but have sister parties there who do, that the Scottish National Party and Plaid Cymru only stand in Scotland and Wales respectively, and that the England and Wales Green Party, the Scottish Green Party and the Green Party of Northern Ireland are all legally separate parties even if they share a general philosophy.

From this it is clear that treating the UK as a single homogenous entity would lead to a false suppression of minority regional voices.

Another argument, and oddly, one that is used in support of FPTP, is that the link of MP to community through representation being determined by a geographical boundary strengthens representation and this would be lost if a MP simply represented their party and not their constituency. It's a fair point.

I therefore sought to find a way to keep this constituency link, and a multi-member constituency based on traditional boundaries appeared to be the most practical way to do this.


I decided to test the system by building theoretical multi-member constituencies based on County boundaries. An obvious place to start being Cornwall as it is in the South West corner of mainland Britain.

Cornwall has 6 Westminster constituencies and in the 2017 election all 6 returned Conservative MPs. Is it therefore safe to assume that a majority of the Cornwall voters voted Conservative? The figures show otherwise.

When combined the votes of the 6 Cornwall Constituencies are:-

  • Conservative - 152,428 (48%)
  • Labour - 83,968 (27%)
  • LibDem - 73,865 (23%)
  • Green - 3,218 (1%)
  • UKIP - 897 (0.3%)
  • Others - 323 (0.1%)

When the d'Hondt system is applied to these numbers instead of the Conservatives having 6 MPs they receive 3, Labour 2 and the LibDems 1.

This is not pure proportionality, but appeared to me to be a step in that direction. To ensure I wasn't being swayed by any anti-Conservative bias I thought it best to test the system again, but in a predominately Labour county. So the next area I looked at was South Yorkshire.

South Yorkshire

In the 2017 General Election Labour won every one of the 14 seats in the South Yorkshire area, including for the first time ever, the Sheffield Hallam seat of one time LibDem leader and Deputy Prime Minister Nick Clegg.

So I again combined all the votes for, in this case 14 seats, and this was the result :-

  • Labour - 356,899 (57%)
  • Conservative - 186,515 (30%)
  • LibDems - 37,177 (6%)
  • UKIP - 29,228 (5%)
  • Green - 7,792 (1%)
  • Others - 9,312 (1%)

Again applying the d'Hondt system to these number Labour instead of having 14 MPs now have 9 and the Conservatives 5. This again appeared to be a more representative result.

I pressed on.

The National Result.

Once I had completed the calculations for each county and metropolitan area of the UK, which involved in some cases splitting large counties and metropolitan areas into smaller areas, combining London Boroughs, and in the case of Herefordshire and Worcestershire combining two counties together, the result was this :-

  • Conservatives - 294
  • Labour - 288
  • LibDems - 23
  • SNP - 25
  • PC - 2
  • UKIP - 0
  • Greens - 0
  • DUP - 7
  • SF - 6
  • UUP - 2
  • SDLP - 2
  • Alliance - 1

This seemed to indicate that I had found a more representative system than FPTP.  Three questions remained.

  1. Would the system be flexible enough to enable the electorate to vote out unpopular MPs or parties or were the large constituencies forming safe seats?
  2. Was it correct to assume that people would vote the same way using FPTP as they would using a more proportional system?
  3. Could a way be found to retain voters' say over the individual candidates elected rather than just numbers of seats allocated to each party?
Question one could be numerically tested.  As a result of the 2017 General Election being called early, it used the same constituency boundaries as the 2015 General Election and the 2010 General Election before than. This is unusual in UK politics as the Boundary Commission proposes changes to parliamentary boundaries quite frequently so as to reflect population movements.

Using the results of the 2010 and 2015 elections I repeated the exercise for the whole country using the same seats and the results were :

2010 - A good election for the LibDems

  • Conservative -   255
  • Labour - 212
  • LibDems - 150
  • SNP - 12
  • PC - 3
  • UKIP - 0
  • Greens - 0
  • DUP - 6
  • SF - 5
  • UUP - 3
  • SDLP - 3
  • Alliance - 1

2015 - A good election for UKIP and the SNP

  • Conservative - 267
  • Labour - 226
  • LibDems - 29
  • SNP - 34
  • PC - 3
  • UKIP - 72
  • Green - 1
  • DUP - 5
  • SF - 5
  • UUP - 3
  • SDLP - 3
  • Alliance - 2

This answered question 1.  The system was flexible enough to enable voters to take seats from one party and give them to another as a party's popularity waxed or waned.

So, do people vote the same way in every election, or do they use their votes in the most effective way given prevailing conditions. The simple answer is that people will vote differently if they think a different vote will deliver a preferred result. Tactical voting under FPTP is real and huge and it is skewing our democracy. In the 2017 election there was more information around about how to vote tactically than I can remember before. Sites like SwapMyVote enabled people to swap a wasted, normally locally third party vote with another voter in a different constituency so as to maximise voter effectiveness. Under the multi-member d'Hondt system this is unnecessary as every vote increases either the chance of a candidate of a voter's preferred party being elected or increases the number of MPs that party will receive.  There is no reason to vote tactically.  This should improve the performance of minority parties across the country.

The third question was how can voters determine the allocation of seats to candidates within a party. This can be done using the Open List System. This is how it works. Voters are given two ballot papers. On the first they are asked to cast one vote for the party or independent candidate of their choice.  These votes, once added up, would determine the allocation of seats to each party. The second ballot paper lists all the individual candidates standing by party and the voter is given the same number of votes as there are seats up for election. So in the examples given above voters in Cornwall would cast up to 6 votes whilst voters in South Yorkshire would cast up to 14 votes. Voters would be able to cast these voters across party lines. This would enable the MPs elected to truly represent the community and not just their party.


I believe the use of an Open List System of multi-member constituencies using the d'Hondt method would deliver as more representative Parliament for the UK.

Please contribute you views and opinions below. I can also answer questions about your particular area if you ask them.


No comments:

Post a Comment