Sadiq Khan, one of the many mayors elected under the Supplementary Vote system.

New research by Make Votes Matter has concluded that some of the key justifications for extending First Past the Post to mayoral and Police and Crime Commissioner elections are not reflected in the data of recent elections. 

Researchers at MVM analysed the 217 elections held under the supplementary vote (SV) in England and Wales. This research directly contradicts some of the key justifications by the government for getting rid of SV in these elections.

How does SV work? 

Elections under SV are not that different to those under First Past the Post. Voters are still presented with individual named candidates. However, rather than putting a cross next to the name of their preferred candidate, they put “1” next to their preferred candidate, and “2” next to their second preference candidate. 

After all first preferences are counted, the two candidates with the most votes go through to a runoff round, and the other candidates are eliminated. Anyone who put an eliminated candidate as their first preference, now has their second preference counted in the runoff round (provided that their second preference wasn’t also eliminated), while first preferences for candidates that went through for the runoff round also count in the runoff round. 

Most of the time, the candidate who came first in the first round, also comes first in the runoff round. However, it is occasionally the case (17 times in the 217 elections under SV), that the second-placed candidate receives enough transfer votes from other candidates to make up the difference between them and the first-placed candidate. 

As mentioned above, it is rare for this to happen. When it does, it is usually because  either both candidates got similar numbers of votes in the first round, or voters particularly disliked one candidate, but had a large choice of alternative candidates to choose from, who split the vote in the first round. 

Two preferences are certainly better than one, but it does not prevent votes from being ignored. As the Electoral Reform Society notes: 


“As the Supplementary Vote only lets voters express two choices, it is possible for a high number of voters’ first choices to be excluded in round one and for their second choice to not be in round two.”

This often leads to the transfer rate, the percentage of second preference votes that are actually registered in the run-off round, being low, at 46%. 

A simple system 

The idea that SV is confusing has been used as a justification to extend FPTP several times by the government, as detailed in a press release

“In this May’s London Mayoral elections, the Supplementary Vote system saw hundreds of thousands of void, wasted or blank votes cast, reflecting voter confusion and the complex system.”

However, this ignores the fact that the London mayoral election in 2021 was held with a new ballot design, which confused voters, and led to a substantially higher rate of invalid ballots than usual. Other elections held under SV that year, such as the Greater Manchester mayoral election, had much lower rates of invalid ballots. 

Historical data for London mayoral elections also shows that the rate of invalid votes is typically much lower, at around 1.8%. 

This rate, however, is consistently higher than rates of invalid ballots in FPTP elections. The 2019 General Election, for instance, had an invalid ballot rate of 0.4%. However, it is often the case that “new” aspects to our democracy, such as new voting systems, are a little harder to understand. For example, in the 2010 General Election, 3.8% of postal votes were marked as invalid. Rather than getting rid of the system, the Electoral Commission instead worked to make postal votes easier to fill in, and this rate had reduced to 2.4% in 2017. 

Thus, rather than disenfranchising the millions of people who use their second preference vote in SV elections, we believe that it is better for the Electoral Commission to look into making it easier for voters to fill in their ballots, and to provide more thorough voter education.  

Moreover, the overwhelming majority of voters have no difficulty using the preferential ballot. Although MVM’s research found that the transfer rate is low (46%), this rate increased to 76% when only three candidates were running for election. 

This is likely due to the fact that SV only allows voters two preferences. Therefore, the more candidates in an election, the likelier it is that a voter picks two candidates who don’t make it through to the runoff round. As our research showed, the more candidates in a race, the lower the transfer rate:

  • In elections with 3 candidates, the transfer rate was 76%

  • In elections with 4 candidates, the transfer rate was 58%

  • In elections with 5 candidates, the transfer rate was 44%

  • In elections with 6 candidates, the transfer rate was 41%

  • In elections with 7 candidates or more, the transfer rate was 36%

An interesting addition to this is data released by the Greater London Authority (GLA) about second preference votes. While most other authorities only count second preference votes for the two candidates in the run-off round, the GLA counts all second preference votes, regardless of whether the candidate makes it into the final round. 

These figures show that voters consistently use their second preference vote over 80% of the time, with this number hitting 85% in 2004 and 2016. However, because SV only counts these votes if they go towards the two finalists, many of them are ignored. 

An ideal preference voting system would have as many successful transfers of preferences as possible.

However, these rates are also improving. While only 35% of voters used a valid second preference in the first SV election in 2000, this number increased to 56% in the 2021 local elections. The message is simple: the more voters use SV, the more they understand its nuance

Aiding in accountability

Another justification used by the government to extend FPTP in the Elections Bill was that FPTP makes local officials more accountable to constituents. As the former Minister for Constitution, Chloe Smith said: 

“The second important feature [of first past the post] is accountability, which with first past the post is linked to place. It can be achieved in other ways, but with first past the post we all get the kind of accountability that comes when an MP walks down the street and looks into the eyes of their constituents.”

However, MVM’s research has shown that there is no empirical link between FPTP and more accountability. If anything, elections held under SV have so far produced fewer “safe seats” than elections held under FPTP. Safe seats can allow representatives to ignore their constituents, secure in the knowledge that they cannot be voted out. The fewer safe seats in our democracy, the more responsive our representatives 

MVM’s research has found that, of the 67 positions currently or formerly elected by SV in England and Wales, only 34 of them (51% of the total) have been held by the same political party since their creation, despite the fact that most have only been elected three times or fewer – in other words, seats change hands in a way which does not happen in Westminster elections, illustrating that the public can vote out representatives who they do not think are delivering for them. 

In comparison, 453 of the 650 Parliamentary constituencies (70%) have been represented by the same party since the 2010 general election, despite the significant electoral realignment which went on in this period. 

Ms. Smith also stated that: 

“It is important to seek a system that has that in its design, so that there is not some relatively difficult to understand method of apportioning the number of votes, but instead a clear method of who came first—the clue is in the name. That gives citizens, voters, a clear method of holding somebody to account.”

This ignores, however, the fact that, had these elections been held under SV, the winner would frequently been elected with a majority of residents preferring another candidate. 

In the 217 elections held under SV: 

  • No candidate got a clear majority of first preferences in 159 of them (77% of the total).

  • No candidate got over 40 percent of first preferences in 78 instances (34%)

  • No candidate got over 33 percent of first preferences on 35 occasions, meaning that in 17% of cases, nearly every 5 elections, at least twice as many voters preferred another candidate to the one who would have been elected under FPTP. 

While FPTP may be able to identify which candidate got the highest total of votes, it routinely fails to select someone with the support of the majority of the community

Bipartisan benefit

Moreover, it is not just third-party candidates who benefit from this system, it is in fact Labour and the Conservatives who benefit the most. Of the 159 run-off elections conducted under SV: 

  • The Conservatives have won 2,136,886 second preference votes in 124 elections 

  • Labour has won 2,393,784 second preference votes in 123 elections 

  • Independent or third-party candidates have won 951,641 second preference votes in 71 elections 

It is also a system which is only becoming more advantageous for the two main parties. In 2012, only 22 of the Police and Crime Commissioner races had both the Conservative and Labour candidates in the final round. In 2021, this number had increased to 33.

Similarly, the percentage of transfer votes going to independent or third-party candidates has sharply declined, particularly since the decline of UKIP in 2016. It is now the case that SV mediates the differences between Labour and the Conservatives, rather than allowing for any real opposition to the two largest parties to come through. 

An imperfect system, but one far better than the Government’s preference 

One thing to be clear about is that SV is not proportional. Indeed, no voting system that results in single-member representation can be. Although millions vote for third-party candidates, those votes are not registered in the results, since only four of the 64 officials currently elected under SV are not from Labour or the Conservatives.

Moreover, giving voters only two preferences still leaves millions disenfranchised, particularly in elections such as that for the Mayor of London, where 20 candidates stood in 2021. This is part of why SV does not meet the standards set by Make Votes Matter’s Good Systems Agreement

However, two preferences is still better than one, because it results in more accurate representation of voters’ wishes. Extending FPTP to elections currently held under SV will only disenfranchise many more voters. 

To help us oppose this change, please sign our ‘Less First Past the Post, not more!’ petition, and ask your MP to oppose this amendment to the Elections Bill.