How Britain’s archaic voting system is warping its politics
The UK political system offers voters many choices, but it’s dated method of counting votes, First Past The Post (FPTP) means that in many cases voters only get the illusion of choice, and can actually end up inadvertently voting for something they don’t want. What makes matters worse is that FPTP disproportionately gives one party — the Conservative Party — an advantage over all others. In this article we’ll explore this effect, and in particular how it affected the most recent UK General Election in 2019.
Have your cake and eat none
Imagine, if you will, that the country is a cake.
Each party wants some cake (who doesn’t). The electorate decide between themselves how much each party gets. Whoever gets the most cake wins. This is FPTP.
Once upon a time, there were only two parties to choose from. Our electorate could therefore give more cake to the person that most of them liked. The people who supported the losing party aren’t particularly happy, but at least you can argue that the system is fair.
At the moment, more people like Lucy on the left than Tom on the right, so she gets more cake, making her the winner. If we play the game on another day, some people might change their minds meaning Tom wins the bigger slice instead.
Over time, more people become dissatisfied with the options on offer. There are many people who would never want to vote for Tom, but Lucy isn’t really motivating them to vote for her. Some people choose not to bother voting at all, and some people start voting for other candidates. Many people feel the reverse: they would never vote for Lucy but aren’t that excited about Tom either. However Tom is a ruthless populist; he’s willing to say and do anything to improve his chances and will back any policy that his natural competitors might also claim, leaving them very little opportunity to get compete. This means those voters who would never vote for Lucy have only two choices: Tom or abstention.
This is what’s happened in the UK, and election results now look something like the picture above. We can see that most people (who bother to vote) don’t want Tom to win, but since the opposition to Tom is divided, Tom has a bigger slice than any other individual so he still wins.
This is the issue with First Past The Post voting. Even though a majority of people don’t want Tom to win, they can’t all agree on who the best alternative is, so they all end up losing instead. The opposition won’t cooperate with each other so their power is diminished.
The problems with First Past The Post
The UK is broken up into constituencies, each of which gets to nominate a political candidate to represent them in the House of Commons. Currently, there are 650 constituencies in the UK. Around 80% of them are in England, which makes sense as a similar proportion of the UK’s population live in England.
In a General Election, which happens at least once every five years, candidates are chosen in each constituency using the First Past The Post system. FPTP is very simple (arguably its only benefit); each voter chooses their preferred candidate, and the candidate with the most votes gets elected. FPTP works well if the following two assumptions hold true:
- Voter turnout is high.
- The winner gets more than half of all the votes.
The first assumption has been slowly coming undone for a long time. As can be seen below, voter turnout has been falling for years. It famously took a big hit during the New Labour years and never really recovered: before Blair, turnout was never below 70%. After Blair, it was never above that:
The second assumption of FPTP has not held true for a long time. The UK has at least three or four parties that gain significant numbers of votes in most constituencies. This means the votes can get split in more ways — Sheffield Hallam for instance was won by Labour in 2019 with only 34% of the vote, meaning two-thirds of those who voted didn’t get what they want (and that’s not even including the one-in-five who didn’t vote).
Other parties have also been created over the years; many of them never gained enough votes to have a serious impact on politics, but a few have and they are also all Left-of-Centre, such as the Green Party and the Welsh and Scottish nationalist parties (Plaid Cymru and the SNP).
No such significant split happened in Right-of-Centre politics, allowing the Conservatives to dominate this space (there was a significant split in 2019, but the Conservatives mitigated against this; see below).
As we will see, these factors have increasingly led to political outcomes that differ wildly from the popular vote.
The 2019 General Election
The toy example above is very similar to what happened in the 2019 UK General Election. It was touted as a huge win for the Conservative Party, and vindication of their controversial policies and approach to governing (yes, I’m talking about things like Brexit, the unlawful proroguing of Parliament, a senior team consisting of politicians with questionable histories and unelected advisers being given huge amounts of unchecked influence). Not only did they win the election, but they won by a huge margin despite a campaign based on an ill-defined slogan and a party leader who spent more time hiding in fridges than engaging with voters.
None of this sounds like a recipe for success, and yet the Conservatives won a huge (80 seat) majority. Why? Were they really that popular?
No.
Thanks to FPTP, demographic imbalances, a weary and apathetic electorate and a plethora of political parties, no one really needs to win an election any more — they just need to lose less than everyone else. The Conservatives seemed to understand this and took full advantage of it; the other parties didn’t. Let’s take a look at the election results:
The Conservatives did indeed “win” the popular vote, but not by as much as the outcome would have you believe. With 43.6% of the popular vote, they were awarded 56.2% of the seats in Parliament — an outright majority, despite not having majority popular support. However that isn’t the worst thing. Parties like the Liberal Democrats won almost 12% of the popular vote, but got fewer than 2% of seats. The Brexit Party got 2% of the popular vote and didn’t win anything. Unless you are one of the top two parties in a constituency, the system is heavily rigged against you thanks to FPTP.
In the UK, the Right is united whereas the Left is split across three major parties (Labour, Lib Dem and SNP) and many other smaller ones. This division seriously diminishes the voting power of the Left. Interestingly 2019 almost threatened the Conservatives with a taste of this disparity. For the first time, a major political force on the Right, the Brexit Party (BXP) was going to compete against the Conservatives. All signs were pointing to a disaster for the Conservatives, until (for reasons unknown) BXP stood down their candidates in every constituency where the Conservatives were competitive. This all-but-forgotten move meant that people who supported Brexit or feared a Labour government (and many did, thanks to its deeply unpopular leader) only had one choice — the Conservatives. Those who opposed the Conservatives, or Brexit (or had merely lost interest in it) were forced to make their own choice between multiple parties and the vote was split.
Unite to Remain
The Tories won a clear victory not by being more popular than the main opposition, but by cutting a deal with the party that would have appealed to the same sub-group of voters. That was a smart idea, maybe the other side should have tried it too! Well… they did, but it didn’t work out very well because the largest two parties, Labour and the Liberal Democrats, refused to cooperate with each other. In the end, the only pact made was Unite to Remain (UTR), an agreement between the Lib Dems and the much smaller Greens and Plaid Cymru.
If Labour and the Lib Dems had put their differences aside and agreed to cooperate, how would this have changed the result? We’ll never know for certain because such a change may have influenced voters to choose differently, but let’s keep things simple and assume that if the Lib Dems stood down in a Labour seat then all their votes would have gone to Labour and vice-versa. Now what would the results look like? We can model this simply by combining the results for all of the UTR parties (+ Labour) into a single entity:
We still end up with a Conservative victory, but a very different looking one. Instead of a stonking 80 seat majority, we now have a 40 seat minority! With only two serious contenders in each seat, we also see FPTP working better as the popular vote is very close to the actual result. However despite losing the popular vote, the Conservatives still win 24 more seats than the entire Remain Alliance combined. This is again due to FPTP, but at the local constituency level.
We can also see where the Conservatives are most vulnerable from this simulation — the affluent commuter towns near the big cities where more and more young and educated voters are moving to escape the house prices of the cities. We also see some of the fabled “Red Wall” of Northern seats that flipped to the Conservatives no longer doing so. This perhaps shows that the Conservative victory here was not so much about “Get Brexit Done”, but more like “Never Corbyn”.
Just to illustrate how much FPTP warps things when you have more than two parties, let’s remove one of the small parties — the Green Party — from our Remain Alliance:
Despite gaining only a single seat, by leaving the Alliance the Greens would hand an extra 12 seats to the Conservatives! The greens may have a small number of votes but they are significant. The biggest political divisions in the UK are not defined by things like class or race, but age and education. Younger people and more highly-educated people are increasingly concentrating in and around the cities, meaning their voting power is being compressed into fewer and fewer constituencies. Locally, this means that the splitting of the Left vote across multiple parties has a larger effect. Nationally, this same phenomenon has exacerbated the dominance of retired voters who are more likely to vote Conservative and are spread across more wider rural areas.
A triangular peg in a round hole
Another way of looking at the election results is with ternary plots. A ternary plot is a two-dimensional graph with three axes. Here is the plot for the 2019 election (England only):
We just look at England because we can only show three parties. The three most dominant parties are different in each nation of the UK (and for Wales and NI, you need to consider more than three parties to gain a complete picture).
Each corner of the triangle represents 100% of the votes going to a particular party.. The markers show the result for each constituency, and the lines can be traced back to the previous election result, showing what kind of a swing there was.
We can see here that despite the election being reported as a huge swing to the Conservatives, the popular vote actually shows much larger swings to the Lib Dems! Unfortunately, most of the Tory seats were previously overwhelmingly Tory, meaning that this large swing wasn’t enough to change the final outcome.
Most of the changes in outcome were Labour -> Tory, but as we can see in the plot, these were due to relatively small changes in the local popular vote. The Conservatives focused their campaigning efforts on seats that were already close contests with Labour and where the Lib Dems never had a chance. The Lib Dems, despite trying to appeal to the youth and the Left, actually appealed more to Conservative voters in safe seats who did not approve of the new management of their party. The Lib Dems biggest gains were in places that they paid little attention to and were very difficult to win anyway.
If you look carefully at the plot, you’ll see a couple of seats going to the Conservatives despite being closer to the Labour corner. This is because the main assumption of ternary plots — that the values on the three axes sum to 100% — has been broken. These are the seats where the Brexit party actually stood. The plot shows clearly how BXP ate into the vote of the dominant party (Labour) and by doing so handed a victory to the next most popular party (Conservatives). If BXP stood in seats where they were competing against the Tories, we could have expected to see this happening (probably quite a lot) to the Conservatives too, drastically changing the result.
Here’s the plot for Scotland:
The Scotland plot shows a strong rejection of the two traditional parties for the alternative, SNP. Most of the results are close to the centre of the plot though, showing that despite a nearly clean sweep for the SNP, the Scottish electorate are heavily divided just like everyone else. And just like the Tories, the SNP have benefited hugely from FPTP voting (although in fairness to them, they actually support replacing FPTP with a proportional representation system).
The Grey Wall
It’s been widely reported that these day the biggest divides in political views aren’t attributable to class, wealth or even location. They are age and education. Electoral ballots are secret so we don’t have any data on how different demographics voted. We do however have some of this information from polls, which act as a good proxy. Let’s take a look at our ternary plots again, but this time for YouGov polling data taken slightly before the election.
Here we compare the YouGov polling data prior to the 2017 and 2019 elections in the same way we compared election results earlier. Because we are now looking at polls, we have demographic data along with voting intent.
Looking at age groups, we see swings across the board away from both Labour and the Conservatives. Under a more proportional system, the Lib Dems would have had a very successful election. We also see that the Conservatives don’t become competitive until you start considering the over-40s. It will be interesting to see if these new voting patterns become habits; if they do, the Conservatives are in trouble. At the moment of course, the over-40s are all they really need since the younger demographics are a smaller and less reliable group of voters. The Labour party are currently focused on regaining their Red Wall seats, but rather than obsessing over these geographic entities, perhaps they should turn their attention to demographic ones instead and figure out how to appeal to retired voters.
Another interesting point is that since these polls were taken before the election, UKIP/Brexit Party are presented as a separate option. At that time, no one knew they would sign a pact with the Conservatives. In the older groups, the biggest swings were actually away from the Conservatives and towards the Brexit party. Pretty impressive for a party with no achievements and no manifesto.
Conclusion
The 2019 General Election in the UK is usually reported as a solid Conservative win that shows the British public approves of their performance and direction. The reality is very different. The population is divided and all we can say for certain is that none of the candidates were popular enough in a voting system that fails to handles the realities of modern Britain, and is utterly unprepared for polarising populist politics.
Again, despite the common narrative it’s largely the same story in Scotland. There was no overwhelming mandate for the SNP or Scottish Independence (although opinions appear to be slowly drifting that way over time). The population is horribly divided, and all we can say for certain is not that the SNP is popular, but that the traditional Westminster parties are all incredibly unpopular. The only difference between Scotland and England in this respect, is that the Scots have an extra option.
