STV Is A Poor Choice For Reform


Following the failure of the AV referendum, the electoral reform movement is now undergoing a major appraisal of where to aim for next. PR in the Commons is the ultimate goal of course, but reformers need to ensure that we don’t blindly stumble into the next campaign trying to sell a complicated, unpopular electoral system to a disinterested public.

STV, or the Single Transferable Vote, is the favoured system of the Liberal Democrats and the Electoral Reform Society. It is essentially AV for multi-member wards; people rank the candidates in order of preference and three to six MPs (usually) are elected for an area.

Here is the description from the Electoral Reform Society (ERS):

STV uses preferential voting in multi-member constituencies. Each voter gets one vote, which can transfer from their first-preference to their second-preference and so on, as necessary. Candidates don’t need a majority of votes to be elected, just a known ‘quota’, or share of the votes, determined by the size of the electorate and the number of positions to be filled.

If your preferred candidate has no chance of being elected or has enough votes already, your vote is transferred to another candidate in accordance with your instructions. STV thus ensures that very few votes are wasted, unlike other systems, especially First-Past-the-Post, where only a small number of votes actually contribute to the result.

However, there are inherent problems with this system:

STV is NOT properly proportional

The defining characteristic of a proportional system is the link between the percentage of the vote and the number of MPs elected for each party in Westminster. STV does not have that. The only ‘proportionality’ that exists is within each super-constituency, and with only three to seven MPs being elected there is very little room for making sure that each party gets their fair share.

With just 3 to 7 candidates the bar for getting elected is very high (between 25% and 13%) , and the level of proportionality is small. If we increase the number of seats per constituency to improve the proportionality, we also further reduce the link between voters and their MPs, and increase the complexity of both the voting and counting. For it to approach acceptable levels of proportionality we would need seats of 10-15 MPs I think, which is unworkable.

This lack of ‘granularity’ means that, while it is a fairer system, it does not create a proportionally representative House of Commons. At best it can be described as ‘semi-proportional’.

STV favours the Lib Dems

Due to the nuances of the system, STV disproportionately favours the third placed party. The Electoral Reform Society did a study of the 2010 General Election results and combined it with  ComRes survey data that looked at how people might have voted under STV and AV. This is the resulting distribution of seats that it achieved:

2010 results modelled under STV

Note that the Liberal Democrats would have received something in the region of 162 seats under STV in the 2010 election. However, a look at the actual votes cast showed that they received 23% of the vote, which should have earned them around 149 seats in the Commons.

This shows that not only does STV benefit the third placed party, it can actually over compensate them and penalise the more popular first and second placed parties.

It is hardly surprising that the Lib Dems support this system and it smacks of personal interest rather than a desire to promote a genuinely proportional system.

Obviously voting patterns would change with the electoral system, but this remains a difficult sell to the public.

STV does little for smaller parties

A major appeal of Proportional Representation is that it is supposed to give all parties (and thus voters) their fair share of MPs in Westminster, especially the smaller parties. There are around 2 million voters in this country without a single MP representing their political views.

For example, in 2010 UKIP received 919,546 votes (3.1%) but no MPs. Under a purely proportional system, they should have had around 20 MPs elected. STV looks unlikely to change that.

Let’s take a look at a specific region, the South East using the same ERS data as above:

Conservative Labour Lib Dem other
FPTP 75 4 4 1
STV 50 11 23

As you can see, under STV it is likely that the Greens would have lost their only seat in Brighton. Similarly UKIP, which received 4.1% of the vote in the South East, should have had at least three MPs elected under an ideal PR system. Under STV they still would not have any.

STV does not assist the smaller parties and requires quite a hefty minimum percentage of the vote in any given region before they can win an MP.

STV removes the constituency link

A major problem of STV is that it removes the constituency link. Whereas small multi-member wards might work for council elections where the areas involved are still relatively small, STV for General Elections would require constituencies the size of counties or small cities. For example, we might have a constituency the size of Bristol or Somerset, with half a dozen MPs for the area.

The result would be the same problem we have with the European Elections – no one can name their MEPs. Constituency MPs are important because they provide a single person who will focus their attention on a specific area and residents know who to contact about their problems.

STV will be seen as complicated

One of the main causes of rejection of the AV system in the 2011 referendum was its alleged complexity. Now imagine trying to promote the same system being used to elect multiple MPs for a region.

But STV is actually MUCH more complicated than AV. Not only do the votes for unelected candidates get potentially redistributed, the excess votes from those who are elected also get passed down. Here’s how it works (taken from Wikipedia):

Setting the quota

In an STV election, a candidate requires a certain minimum number of votes – the quota (or threshold) – to be elected. A number of different quotas can be used; the most common is the Droop quota, given by the formula:

\mbox{votes needed to win} = \left({{\rm \mbox{valid votes cast}} \over {\rm \mbox{seats to fill}}+1}\right) + 1

The Droop quota is an extension of requiring a 50% + 1 majority in single winner elections. For example, at most 3 people can have 25% + 1 in 3 winner elections, 9 can have 10% + 1 in 9 winner elections, and so on.

Finding the winners

An STV election proceeds according to the following steps:

  1. Any candidate who has reached or exceeded the quota is declared elected.
  2. If a candidate has more votes than the quota, that candidate’s surplus votes are transferred to other candidates. Votes that would have gone to the winner instead go to the next preference listed on their ballot.
  3. If no one new meets the quota, the candidate with the fewest votes is eliminated and that candidate’s votes are transferred.
  4. This process repeats until either a winner is found for every seat or there are as many seats as remaining candidates.

STV damages local campaigns

Because of the much larger constituencies required for STV to function, this means that local campaigns such as the Save Kidderminster Hospital group would never be able to make an impact. They managed to get one of Britain’s very few independent MPs elected based on an issue that just affected one area. Had the Wyre Forest constituency been part of a much larger constituency it is quite likely that their local campaign would have been airbrushed out of our political history.

Similarly, the small Green party enclave in Brighton would also be averaged out over a much larger area.

STV is bad for independents

Just as with the problems it causes for local groups, the larger constituencies make STV almost impossible for independent candidates to campaign. How can one person possibly hope to make an impact in an area the size of Somerset for example? STV is good for political parties with their established brands but bad if we would like some independent opinions in Westminster.

STV encourages ‘Donkey Voting’

In a six seat constituency, each party is expected to put up six candidates. If we have the six largest parties standing in each area that will mean that there should be 36 candidates plus however many independents there are. All of those names have to appear on a ballot paper and then voters will be expected to rank as many as they like.

Most people will still vote for a party, and will be encouraged to rank those party candidates in order of preference. In reality, the vast majority of voters will have no idea who the candidates are, nor have any rational basis on which to rank them. The result is known as ‘donkey voting’ – a voter just ranks them by alphabetical order or a random distribution of preferences. Is that really progress?

To prevent this happening, parties commonly only put in the number of candidates that they expect to be elected. The result is that they are essentially offering a ‘party list’ of candidates and the rankings become irrelevant if all of the candidates are elected. This completely undermines the whole point of STV and we might as well use a list system as we do for the EU elections.

STV encourages ‘core vote’ campaigning

One of the key selling points of AV was that it encourages candidates to appeal to a much broader spectrum of voters in order to ensure enough second preference votes to push them over the 50% mark. The opposite effect comes into play with STV.

Because candidates compete with other members of their own party for the first placed spot (and thus the most chance of being elected), and they now only need between 13% and 25% of the vote to be elected, they are therefore encouraged to focus their campaign internally. The result is that they will spend more time campaigning to their core vote and ignoring the wider electorate.

  1. #1 by Will on May 21st, 2011

    Of course, you are proposing an unproportional system yourself. The Lib Dems get under-represented under the system advocated on this site. Indeed, it hurts third parties more than STV.

    The problem with the ERS’s calculations is that they assume everyone voted the way they did in 2010, and did second preferences from there. Of course, in the case of Brighton, Caroline Lucas probably would’ve won on a hypothetical “Brighton and Hove”; her win in Pavillion was spurred by a belief that they could actually win. Given that Brighton is a relative Green stronghold, and that STV first preferences cannot be tactical, she would most likely be returned under STV.

  2. #2 by Anthony Butcher on May 21st, 2011

    RTU goes a long way towards improving the proportionality of the existing electoral system that the public are comfortable with. It’s not perfect, but would improve significantly as voting patterns change to reflect the system, as you mention would happen under STV too.

  3. #3 by Will on May 22nd, 2011

    One of my main problems with most top-up systems as proposed is that it allows parties to “parachute” candidates in by putting them on the list. Peter Mandelson would never win these days in most constituencies. But put him first on the list where Labour don’t win constituencies, and he’ll surely get elected.

    I think you’re overselling how bad STV is for a local campaign. There’s an entry-level Vernon Bogdanor book, “What is Proportional Representation?”, where part of the chapter on STV dissects the 1982 election results for Carlow-Kilkenny in Ireland. He found that voters in Carlow wanted to elect candidates of the same party from Carlow over candidates from Kilkenny, and vice-versa.

    And with a lower quota, independents find it easier to be elected; 10% of the TDs in the Dail are independents. Sure, Richard Taylor may not have made that many waves in Worcester, but he could build up a cross-party support base in the north of the county and easily hit the 14% or 16% quota for election.

    Finally, I like STV as it allows for more intelligence in creating boundaries. Current constituencies are based on ward boundaries, which often cut through street-level communities to keep balance; I know that my student flat, in Leeds, is in a different ward to the flats ten yards away, across the road, but there’s a lot of socialising between the sets of flats. STV would only need to do district and county boundaries, which are more tied to history and as such, wouldn’t split communities.

  4. #4 by Anthony Butcher on May 22nd, 2011

    I should point out that RTU doesn’t use party lists – the lists are created from the most popular runners-up, so everyone is elected by the public.

    I take your point about the lower bar for independents making it easier, and Dr Taylor might still have got in due to the strength of the vote in one local area. However, that would still mean that large numbers of people in his new supersize constituency would never have heard of him.

    Let’s not forget that they are recommending constituencies of 5-7 seats for the Lords, so presumably the same would apply to the Commons. It is hard enough for an independent candidate to cover an electorate of 75,000, let alone 525,000. Do we really believe that a single independent candidate from a field of 30+ candidates can have any kind of relationship with half a million people?

  5. #5 by Nate Simpson on March 27th, 2014

    While it is interesting to look at – comparing the results as “what you could have won” is overly simplistic – there is no way of knowing what the results would have been they are an extrapolation with significant assumptions (you have not detailed these)
    It also ignores tactical votes for the big parties and the potential to increased turnout if it is always worth voting for.

  6. #6 by Anthony Butcher on March 27th, 2014

    Of course, but it is the only information available to us that we can use to get a feel for what things might be like under other systems. The numbers used are from the Electoral Reform Society who specialise in this kind of thing.

(will not be published)


  1. No trackbacks yet.
eXTReMe Tracker