Why the Single Transferable Vote is the best form of PR.

Before I start, I would like to state that I freely admit that there is no such thing as a perfect electoral system.
Arrow's Impossibility Theorem tells us that. Electoral systems are in some respects like map projections, it is already known that a perfect solution is impossible, however that doesn’t mean we can’t improve upon it.

All electoral systems have some level of unfairness, scope for throwing out an arbitrary or incorrect outcome or opportunity for tactical voting. Improved systems just cause them to become much less likely to occur and make tactical voting much harder.
As for national elections, it’s also important to find the balance between local representation and nationwide proportionality.

If electoral systems are compared to map projections then I think that if the Dutch system of rigid PR is Mollwiede and FPtP is Mercator then STV is the Winkel Tripel.

 

The AV referendum and preferential voting.

In 2011 in the UK, we had the opportunity to change the electoral system.

Contrary to what many Tory MPs now erroneously claim, It was not a form of proportional representation that was on offer, it was a system designed to be more representative at a constituency level.
The Conservatives, in their infinite wisdom, were against this reform, although that is no surprise, they are so far against any kind of electoral reform they let Labour take office in 1974 partially to protect the current system even though it had led to Labour taking the plurality of seats despite them having gained more votes. 
In 2010 however, the Tories really should have been in favour of AV, it could have given any future Tory government elected with under 50% of the vote a bit more legitimacy, without using PR.

 

Most people who favour FPtP say they do so because it leads to ‘strong single party governments’ and they’re right it often does or it’s more likely to. Unfortunately, this is often a strong single party government elected by as little as 35.2% of the population, as Labour got in 2005.
I would argue that if strong single party governments are important to you then the best system for that is not FPtP but instant run-off voting, also known in the UK under the vague title of The Alternative Vote (AV).
The reason for this is that it’s still likely to lead to a single party winning a majority of the seats in many or most elections, it would do this while allowing people to vote for who they really want without fear of letting someone else win, but most importantly of all the national two-party-preferred result would somewhat deal with the minority rules problem of FPtP without going to PR

For example, if after the 2015 election AV had been used, The Conservatives had won and the result showed that after all preferences had been allocated people preferred the Conservatives to Labour at a rate of 51-45% (the other 4% making no preference for either) then there might have been fewer people protesting in the street at how the election result was fundamentally wrong and unfair.

While AV is not to be considered as any form of proportional representation it was thought at the time the lack of a need to vote tactically may have led to a slightly more proportional result nationwide, although this is just incidental and it would also have led to a small yet significant drop in the overall percentage of wasted votes. If Australia is anything to go by it may have led to a handful of independent MPs being elected at each election.
When it came to the 2015 UK General Election, the projection was that the Conservatives majority would have been double what it was had AV been the system used.
David Cameron promised the EU referendum on the run-up to the 2015 election because he saw the rise of UKIP as a threat, he sealed his fate by not backing AV.

 

Proportional Representation

Most countries in Europe and many countries throughout the world use a form of electoral system designed to give some level of proportional representation, however, the systems used and the level of which they’re proportional varies greatly.

General elections normally decide two things; who will be the local representative in parliament and who people want to be running the country.
Some proportional representation systems do away with local representation altogether and everybody is just voting for the government, however, most systems of PR are trying to maintain a certain degree of balance between local representation and proportional representation, the method used to maintain this and the effect it has on the level of proportionality and how local the representation is, varies from system to system.

There are quite a number of proportional representation systems used throughout the world, countries in Europe that have it use a plethora or different systems many of which are either based on large electoral regions or are a variation of the mixed-member proportional system.

MMP

In the mixed member system, a number of the seats are decided by first past the post and the rest are divided up to try and give a more overall proportional result.

There are a number of problems with this system. The first is that it still relies on the overly simplistic and outdated first past the post method for constituencies, (you could use AV for these elections, but for some reason no one does this) you have two kinds of MPs and often, to try and reduce the likelihood of extremist or fringe candidates getting into parliament, there tends to be some kind of arbitrary cap called an election threshold that a party needs to get in order to gain any seats. This is often between 3-5%.
In Germany for example if a party gets 4.99% of the vote it gets no seats and all of the votes for that party are simply discarded. However, gain 5% of the vote and it’s at least 30 seats.
That’s why in Germany the FDP had no representation in the national parliament between 2013 and 2017, only 4 parties did, because 15% of the party-list votes were wasted the CDU were almost able to gain a majority with only 41.5% of the vote, whilst in 1983 they were unable to gain a majority with 48.8% of the vote.

What has happened in Germany to allow the CDU to almost gain a majority, is that because so many votes are wasted on candidates who do not make the threshold it’s inadvertently made easier for a single party to gain an overall majority with significantly below 50% of the vote.
Having said that, I don’t think gaining 50% of the vote should be an absolute requirement of a single party holding a majority of seats in a parliament, I think it’s more troubling under this and other non-ranking systems than it is under a preferential voting system, the reasons for which I will explain further on.

One Nation Party-List

One solution for PR would be to have no local constituencies and no threshold, this would be a system equivocal to the strictly proportional system used in The Netherlands.

The Netherlands has possibly the most proportional system used anywhere in the world in terms of seat distribution. This might be the best form of running elections if elections were not at all about having local representation and were purely about proportionality between parties. Despite the proportionality, The Netherlands doesn’t rank as highly as Ireland on the democracy index.
The entire country is treated as one constituency and elections take place using an open party list, there is no capping so with 150 MPs, just getting 1/150^th of the vote is enough to gain an MP.

While this is extremely proportional and very few votes are wasted, there is no form of local representation in parliament, and it’s a closed party list, so people do not have a say on candidates, only parties. This form of strict proportionality seems to lead to a parliament that is deeply fragmented.
Often even after the result is in, it can still remain highly unclear who will form the next government, even a combination of the two largest parties often won’t have enough seats for a majority and negotiations must be held between several parties.

In 2010 the largest party got just 20.5% of the vote, it took almost 5 months to form a government, which lasted just 2 years also there were no independent members.
In the next election, the largest party did better with 26.6% of the vote and it only took almost 50 days to form a government. No government has lasted a full term since 2002.
Another example is from 1977 the PvdA won the election with 33.8% of the vote which is considered a “landslide” by Dutch election standards; however, it took over 200 days to eventually form a government that did not include the PvdA.

Personally, I think this kind of pure PR seems to lead to the sort of situations that opponents of PR in general tend to allude to.

Regional Party-list

Many countries have PR that works by splitting the country in to regions some with a national top-up to ensure nationwide proportionality such as Denmark, often these will have caps.
However, although these caps are normally only in place to get nationwide top-up seats, so it is less of a barrier for independent or regional party-based candidates at getting some level of representation.
Still, under this system, the regions are very large so “local” MPs may cover entire regions and all votes for candidates who don’t make the threshold are still ignored.

 

STV

Under the single transferable vote, the country would be split in many multiple-member constituencies consisting of between 3 to 6 members. Some people would suggest more than that and/or no variations in the number of members per seat, but I’m going to use this for my example.

Firstly, you might be wondering why the variation in the number of seats and therefore the level of proportionality?
It’s important to remember that under STV, purely being proportional at a nationwide level is not the aim, the single transferable vote system is designed to; give a broad level of representation to a local area and to minimise the number of wasted votes whilst still managing to be quite proportional at a national level, primarily because of the lack of any need of any nation-wide electoral threshold.

Under STV the ability to vary the number of seats is actually quite important, it is so there is less of a need to have arbitrary electoral boundaries for constituencies.
Under STV constituency boundaries would have less need to have odd shapes with panhandles and isthmuses or to cross major cultural, man-made or geological boundaries that set areas apart.
Boundaries can still change, yes, but wherever possible the number of MPs would be set so that distinct areas are kept as one constituency.
For example; any conurbation large enough to warrant 3 MPs or more could form one constituency, if the population rises in that city then the number of MPs could rise to 4 or 5 then if it gets really big it could increase to 2 constituencies of 3 MPs each and so on.

Under STV in a system known as the Droop quota, the number of votes needed to be guaranteed a seat is calculated thusly.
The number of votes divided by the number of seats plus one, plus one vote.
For example; if a 5 seat constituency had 180 voters the number of votes needed to be guaranteed a seat would be 31 that is 180/(5+1)+1 That’s 16.77% of the vote plus 1 vote. This would only rise to 37 (20%+1) votes in a 4 seat constituency and 46 (25%+1) votes in a 3 seater.

The biggest problem with FPtP is that in areas where one party does not do well they can be pretty much ignored by that party, as they have little to no chance of winning seats there regardless of how well they do nationwide.
For example; as it is now, the people of Liverpool or the South Wales Valleys can pretty much be ignored by the Conservatives, they know they stand no chance of winning any seats there.
In those areas, fewer Tory votes are fewer votes for the Tories where they won’t win anyway so that cannot hurt them, and more Labour votes are more surplus votes for Labour that cannot help them.
Under STV when a party like the Conservatives does well nationally, they will stand a good chance of winning a seat in those kinds of areas, not many but they would still have something to gain or lose none the less.
The same can be said for Labour in large rural seats in the South East. Also in areas where a single party single party does consistently well, they will have something to lose, even if they still win most of the seats in bad years. 

This will also mean that those parties will spend more time and effort trying to win support and listening to voters needs come election time, this can already be seen in local elections in Scotland, in the 2012 local election in Scotland Labour councillor Paul Godzik was canvassing in a well-off area of Edinburgh he admitted that it wasn’t a “traditional Labour area” and how it was “not an area that got a great deal of attention from the Labour party in past years, STV has made us change our approach to campaigning… there’s no ‘no go areas’ for Labour right across this constituency”.
Under FPtP not only are you more likely to be ignored by certain parties depending on where you live but the power of your vote varies greatly depending on how marginal a seat you live in, this greatly affects the amount of power your individual vote has. People who live in very marginal seats can have an order of magnitude more power in their vote than people who live in safe seats.

Wasted votes

Wasted votes are defined as all votes for losing candidates as well as surplus votes for winning candidates.
Under FPtP at the 2015 UK general election, the proportion of wasted votes was 74.5%, under STV the proportion of wasted votes tends to be in the mid to high teens.

I think the biggest plus to STV is that not only are people much more equal in terms of voting power but also coupled with the enormous reduction in wasted votes would greatly increase the incentive for people to vote, and most people will have at least one MP that they voted for.
The beauty of the single transferable vote is not only are votes for candidates who get eliminated transferred but also excess votes are transferable as well, just to minimise the number of wasted votes.

If a winning candidate gets double the votes than they need to get elected according to that seat’s election threshold then everybody who voted for that candidate will get half of their vote back.
The proportion of the vote you get back is dependant on the percentage of surplus votes your preferred candidate gets, fractional votes are transferred before lower placed candidates are eliminated to minimise the spoiler effect and votes can be transferred more than once.

This can be complicated but the people at the count know what they are doing they have been doing it in Ireland for decades, and in Scottish Local elections since 2007.
It may take a longer time, but that should be of little consequence considering that this is an election to decide the representatives for the next few years, I’d say the most important thing is that people get the fairest result possible.
This also means, if you like a certain candidate, however, they are not from the party you usually support, you can rank that candidate first, then rank 2, 3,4 etc for candidates from your preferred parties, STV gives you options like this.

STV is better than is looks on paper.

Some may have noticed that 3 to 6 members per electoral district doesn’t seem very big, however as previously stated there would be no nationwide electoral threshold. Coupling this with the preferential nature of STV seems to create a system that is more proportional than if there were small electoral districts that simply dished out seats proportionally to that district using the D'Hondt method.

As previously stated, most countries with PR have national thresholds to try and lock-out fringe candidates and extremists. However, these caps not only make it harder for extremists but also much more sanguine smaller parties, arbitrary caps can't discern between the two.

A similar effect can sometimes be seen under first past the post.
Looking at results in Ireland it would seem that the preferential nature of the system does tend to be kinder to smaller parties than you otherwise might expect.
Denmark which has a very proportional system has a threshold of just 2%, that is, still significantly more than the 1.5% of the national vote the Irish Greens got when they won their first seat in Dáil Éireann in 1989.


In the UK in 2010 The Green party were able to win a seat in parliament for the first time. Their vote share nationwide was just 0.9% however they only ran in 310 seats so they got roughly 1.8% of the votes on average in seats they ran in.

Caroline Lucas managed to win the seat of Brighton Pavilion with just 31.3% of the vote, that is 17 times the vote the Green Party got on average in seats they ran in.

It took years of hard work and campaigning for them to win that seat and now they’ve done what was considered unlikely, and have a seat in parliament.
Chances are while support for the Greens is much higher than average there, it isn’t likely to be 17x more support than average in a single area, many people may have voted for the Greens because they’re sympathetic enough to their cause to feel that they deserve a voice in parliament.
On the other hand, the opposite effect can be seen in a different constituency. In the same year, the BNP averaged roughly double the vote share of the Greens in the seats they ran in, and still had just about no chance of winning anywhere.

In their most popular seat of Barking, there was a big increase in turnout and there was a swing from the Conservatives to Labour, well against the national trend as people actively voted against the BNP.
Case in point, the only reason the BNP were able to win seats in the 2009 European elections is that there were so many wasted votes cast for parties that did not gain the number of votes needed to win seats. They did not get even close to would have been the Droop quota in either constituency. Worse, the Green Party, just by running, inadvertently helped get the BNP elected.
Yes, you read that right.

This kind of voting behaviour goes some way to explain why, when STV is used smaller but more reasonable parties are able to obtain seats quite easily while more extreme parties will find it harder to gain any seats despite getting more first preference votes or simply under-perform in seat count relative to their vote share.
You may think this is unfair, but it’s important to remember that the system isn’t rigged against any specific parties, it’s just that people tend to vote that way.
It’s a lot fairer than having a cap that not only blocks out extremists but also more reasonable parties.

Extremist parties do not tend to gain preferential votes from people who do not rank them first, also people will tend to actively not rank parties they find abhorrent, actively voting against them by filling the list in, while other smaller parties tend to get transfers from people who ranked other candidates from mainstream parties first.
Also coupled with smaller parties concentrating their efforts on winnable seats means the number of smaller parties gaining entrance to parliament in Ireland is quite high.

Some people might still have concerns that for example, the Electoral Reform Society has only projected that the Greens would have won 3 seats in 2015 when they had 3.8% of the vote, this might sound pretty poor. It's no surprise then that the Green Party favour MMP over STV.

This is mainly down to their figures being based on how people voted under the current system and projections on where preferences would go based on polling data. Looking at Ireland, in 2002 the Irish Greens, also got 3.8% of the vote and they were able to win 6 seats out of a parliament of just 150, that would be the equivalent of winning 26 seats out of a parliament of 650, so I find it hard to believe that with a completely different system the Greens would still have won only 3 seats.

It has not always been that friendly to the greens in Ireland, however, they’ve never done as badly as UK Greens were projected to have done with the same vote share.
I think the notion that the greens in England and Wales would only have gained 3 seats in 2015 with the sea change in the way campaigning and elections are ran is quite ridiculous.
Of course under a list system with a 5% threshold, not only would they have no seats but all votes for them would be discarded and this can disincentives honest voting.

The National Result.

I previously stated, getting over 50% of the vote is not an absolute necessity for a single party to gain an overall majority under STV. Under rigid PR systems with no regions or thresholds, it would be very close to 50% needed to get a majority but still possible.

Thanks to election thresholds and regions, under most proportional systems, it’s possible with significantly lower than 50% of the vote.
If a single party gets to the level of about the mid-40s up, you see the potential for them to get an overall majority.
Under STV, once a single party gains over 43% of the vote, you start to get into the territory where it might be possible for a single party to win an absolute majority.
I would argue that this is less troubling under STV than it is under other systems, the only way a party with slightly over 40% of the vote gains a majority of the seats on a minority of the (first preference) vote is if they are being ranked highly by a large number of the electorate who do not rank them first.
However, if a single party got 43% of the vote under STV but you’re despised by almost everyone who voted for anyone else then chances are you will not win an overall majority in fact the party might slightly under-perform in seats relative to the vote share.

Overall, I think STV is a compromise, it’s not the most proportional, it’s not as local as having single member constituencies, however PR is trying to square a circle, and overall, I think this would be the best way to do it.