Voters Left Voiceless: The 2019 General Election

The 2019 Election

The 2019 general election was pitched as a ‘People versus Parliament’ vote, with voters urged to break the Brexit ‘deadlock’ of the past few years. But the so-called deadlock was a crisis of Westminster’s own making. A system built on unearned majority rule and confrontation had ground to a halt when faced with the realities of a 21st century electorate.

December’s election saw the system fight back. The electoral system did what it is supposed to do in manufacturing a majority for one party at the expense of voters’ choices. The Conservatives gained an extra 48 seats (7.4% increase in seats from 2017) on a 1.3 percent increase in vote share, delivering a majority of 80 seats, the largest for the Conservatives since 1987. This is an extraordinary shift given the previous election had seen the Prime Minister lose her majority on a similar vote share.

A quarter of votes went to parties other than the largest two, but they returned less than 13 percent of seats. Bolstered by the discrepancy of vote and seat share for the SNP in Scotland, this figure conceals an even greater inequality for smaller parties. Nationally, over 865,000 votes were cast for the Green Party, but they elected just one representative.

Our analysis shows that across the UK, over 22 million votes (70.8%) were ignored because they went to non-elected candidates or were surplus to what the elected candidate needed. In total, 14.5 million people (45.3% of all voters) cast their vote for a non-elected candidate. Of course, not every candidate or party can or should secure representation – but First Past the Post is brutal in denying millions of voters any representation at all. Other electoral systems redistribute preferences to ensure the outcomes reflect more accurately voters’ choices and diversity of opinion.

Tactical voting and electoral pacts loomed large over the election, with YouGov polling for this report revealing that one in every three voters (32%) opted for a tactical vote, instead of choosing their preferred party or candidate. This is a significant increase on the last general election, as reflected in the campaign itself which often focused more on electoral tactics than parties’ policies. Both parties and voters are increasingly being forced to game the system – electoral pacts and tactical voting are symptoms of a system that isn’t working for either.

There is much on which to reflect in this report from all perspectives. For the Labour Party, the concentration of the Labour vote in certain areas meant that it took on average 50,835 votes to elect a Labour MP, whilst only 38,264 votes were needed to return a Conservative MP. This vote concentration is also seen in the top 10 largest majorities, nine of which are in Labour seats, where parties pile up votes without securing real representation.

While the Conservative Party is benefiting from Westminster’s voting system across the UK as a whole, in Scotland, a substantial Conservative vote share (25.1%) yielded just six seats (10.2%). Smaller parties contesting seats spread across the whole of Great Britain – the Liberal Democrats, Green Party and Brexit Party – continue to lose out. For the Liberal Democrats, an 11.5 percent vote share across Britain resulted in just 1.7 percent of Commons seats.

Drawing on exclusive YouGov polling, we have modelled how the results would look under different electoral systems. These results can, of course, only approximate what a different electoral system would produce – voters behave differently under different systems as our tactical voting polls show – but they give a clear indication of how much the system affects the outcome and how much First Past the Post is distorting the electoral map.

Acknowledgements:

We would like to thank Jess Blair, Doug Cowan, Josiah Mortimer and Jon Narcross for their help in this report.

2019 General Election under First Past the Post

In an election campaign characterised by uncertainty and volatility, it came as a surprise to many that the result would deliver such a decisive majority for one party. The Conservative Party made a net gain of 48 seats – an increase of 7.4 percentage points in their seat share compared to the 2017 general election and the largest majority for the Conservatives since 1987.

The final polls had predicted Conservative seats ranging between 311 and 367¹Abraham et al. (2019). ‘The key findings from our final MRP poll’. YouGov, 10 December. https://yougov.co.uk/topics/politics/articles-reports/2019/12/10/key-findings-our-final-mrp-poll. That the difference between a hung parliament and a large majority for one party rested within a polling margin of error shows just how erratic the electoral system can be, particularly when there are more than two parties in contention.

Due to the oddities of First Past the Post (FPTP) – or one-party-takes-all results – the Conservative Party was rewarded with a majority of seats (56.2%) on a plurality of the vote (43.6%) – with a 1.3 percentage point increase on its 2017 vote share giving the party a 7.4 percentage point increase in seats. The Scottish National Party (SNP), who support a move to a proportional system at Westminster, also benefited from FPTP, gaining 7.4 percent of seats in Westminster on only 3.9 percent of the vote.

While the Labour Party’s results were much more proportional, the Liberal Democrats were again disadvantaged by FPTP – the party saw an increase of 4.2 percentage points in its overall share of the vote compared with 2017, but actually suffered a net loss of seats at this election.

Once again, smaller parties were penalised by Westminster’s broken electoral system, with the Green Party only securing one seat, despite winning almost three percent of the vote. Brexit Party voters were denied any representation despite getting two percent of the vote.

This election saw an increase in vote share for smaller parties – from 18 percent in 2017 to around 25 percent in 2019 – that is more in line with previous elections, and reflects the long-term trend towards multi-party politics. This is despite a significant two-party squeeze which took place during the campaign, with the two ‘big’ UK parties struggling for 50 percent of the vote earlier in the year, but ending up with 75 percent of the final vote share in this election.

Table 1: 2019 UK General Election Results

Figure 1: Percentage Point Change since 2017

England

The two-party squeeze is most evident in England, with the Conservatives and Labour together taking over 98 percent of the seats with 81 percent of the votes.

This election, however, over five million votes went to parties other than Labour and the Conservatives in England (nearly 18.9% of the vote), yet resulted in just 1.7 percent of the seats.

The Conservatives gained an additional 49 seats in England: a 9.2 percentage point increase in seats⁵This percentage increase includes the former Speaker’s seat (Buckingham) as a gain for the Conservatives. for a 1.8 percentage point increase in votes. The Liberal Democrats lost a seat despite a 4.6 percentage point increase in votes.

Despite the Conservatives picking up seats in traditionally Labour-voting areas, large regional divides continue, with the Conservative Party over-represented in the South East (88.1% of seats on 54% of the votes in this region) and Labour over-represented in the North East (65.5% of the seats for 42.6% of the votes).

Table 2: Election Results in England

Scotland

In Scotland, the SNP’s performance delivered one of the most disproportionate results with a 22 percentage point increase in seats for an eight point increase in votes. The SNP now hold 81 percent of the seats on 45 percent of the votes.

The results show how FPTP struggles to translate votes into seats in multi-party contests but also how precarious victory can be under this system. Scottish constituencies feature prominently among the top 10 smallest winning margins and smallest winning majorities UK-wide. These small majorities are typical when FPTP is used in seats where more than two parties have a significant amount of support.

Table 3: Election Results in Scotland

Wales

Wales returned one of the more proportional results for most parties except Labour, which received a majority of seats for only 41 percent of the votes. Though – as is often the case with FPTP – a decline in support gave the party a disproportionate drop in seats, with an eight percentage point decrease in votes leading to a 15 percentage point decrease in seats.

A large number of votes went unrepresented – with no seats for the Liberal Democrats, Brexit Party or Green Party despite all these parties increasing their vote share.

Table 4: Election Results in Wales

Northern Ireland

Northern Ireland’s multi-party politics was again, as in 2017, squeezed into a warped two-party shape with 83 percent of the seats going to just two parties despite 47 percent of votes going to others.

Northern Irish seats feature quite prominently in seats with the lowest winning vote shares, with two of the three lowest, including the lowest (32.4% in South Down) coming from Northern Ireland. As in Scotland, tiny winning margins – where a candidate can slip in on a fraction of the vote, and all others are discarded – are the result of trying to force a voting system designed for two-party politics onto a diverse, multi-party contest.

Table 5: Election Results in Northern Ireland

UK: Votes Needed per MP

The number of votes needed to elect an MP differed quite significantly for each party. On average, it took 38,264 votes to elect a Conservative MP, while it took 50,835 votes for a Labour MP. Strikingly, it took 865,697 votes nationally to elect just one Green Party MP and 336,038 votes for a Liberal Democrat – demonstrating how punitive Westminster’s warped system is on parties whose votes are not concentrated in specific constituencies, but spread out across the nation. The Brexit Party did not win any seats, despite having received 644,255 votes nationwide, while it only took 25,882 votes to elect an SNP MP.

Turnout

The 2019 general election was the first election to be held in December since 1923. Prior to polling day, there was speculation that bad weather and reduced daylight hours might affect not only campaigning but also turnout at the election itself. However, this does not appear to have been the case.

Turnout at the 2019 general election was 67 percent across the UK – 1.4 percentage points lower than in 2017, but higher than at any other election since 1997. There was some variation in turnout among the nations, with Scotland having the highest turnout (68.1%) and Northern Ireland the lowest (61.8%). English regions were similarly diverse, with turnout being highest in the South West (72%) and lowest in the North East (64.2%).

Figure 2: Turnout since 2005

Table 6: Turnout at the 2019 General Election

Parliamentary Representation

This election saw 220 women elected, 34 percent of the total number of MPs. In the last election 208 female MPs were elected (32% of the total), up from 191 in 2015. Once again, we see only minor increases in the proportion of women elected. At this rate, it will take another nine general elections (45 years) for women to reach parity in the Commons.

There have been some milestones achieved, however, with the Labour Party and the Liberal Democrats having more than 50 percent female MPs this election.

Table 7: Number of Women Elected by Party

Even before a single vote had been cast in the general election, it was clear that it would be difficult to improve women’s representation in parliament because women were not equally represented amongst candidates. Of the 3,322 candidates standing, there were 1,124 women – just 34 percent of the total.^{6BBC News (2019). ‘General election 2019: Record number of women set to stand’, 18 November. https://www.bbc.co.uk/news/election-2019-50441274} Only the Labour Party selected more than 50 percent of female candidates (53%), with the Green Party closely following with 41 percent female candidates. The SNP, Liberal Democrats and Conservatives selected around one third female candidates (34%, 31% and 30% respectively), while Plaid Cymru selected 25 percent and the Brexit Party 20 percent.^{7Candidate data taken from Democracy Club’s crowdsourced candidate list https://democracyclub.org.uk, downloaded December 2019}

Female candidates do much better under PR systems than under First Past the Post.^{8ERS (2018). ‘How Westminster’s voting system is holding back gender equality’, 13 February. https://www.electoral-reform.org.uk/how-westminsters-voting-system-is-holding-back-gender-equality} ERS research has shown how ‘seat-blocking’ by male MPs is a key obstacle to increasing female representation in the House of Commons.^{9Mortimer and Terry (2015). Women in Westminster: Predicting the Number of Female MPs. https://www.electoral-reform.org.uk/latest-news-and-research/publications/women-in-westminster} The longer a seat has been held by an MP, the more likely it is that the incumbent is male. With an electoral system which encourages a high proportion of safe seats, long-term male incumbents block new women MPs from coming through.

There is a clear relationship between the number of women selected in (winnable) seats and the resulting gender balance in the Commons. We are calling for the government to enact section 106 of the Equality Act 2010 to ensure that parties publish diversity data of their for candidates who are both successful and unsuccessful in their selection processes. Currently, there is no official, consistent and good quality information on the diversity of those who put themselves up for candidate selection at any level of government in the UK. This lack of data on candidate diversity means that it is hard to identify if, where, and why selection processes might be hindering greater diversity of candidates, including female representation.

The lack of candidate data means that information on diversity itself is also difficult to obtain. We do know, however, that the election returned the most diverse parliament ever not just in terms of gender, but also race and sexuality.^{10BBC News (2019). ‘Election 2019: Britain’s most diverse Parliament’, 17 December. https://www.bbc.co.uk/news/election-2019-50808536} One in 10 MPs (64) are non-white, though none of them represent a constituency outside of England. One in five Labour MPs come from a black, Asian or minority ethnic background, compared with six percent of Conservatives. Forty-five MPs are openly gay, lesbian or bisexual, with almost one in five SNP MPs being openly gay.

Proportionality

A well-established measure of disproportionality is the Deviation from Proportionality (DV) score. The DV score shows the extent to which an election result deviates from proportionality, i.e. from what it would look like if seats were proportional to votes gained by each party. It gives a percentage of seats in parliament which are ‘unearned’ in proportional terms, and which the party would not have obtained under a more proportional voting system.

There are various ways of measuring DV scores. We have used the Loosemore-Hanby index, which is calculated by adding up the difference between each party’s vote share and their seat share, and dividing by two. This gives a ‘total deviation’ score – the higher the score, the more disproportionate the result. A typical proportional system will give a DV score of 5–8.

Table 8: Deviation from Proportionality Score in the UK and in the Nations

The DV score for the UK overall is 16, which is higher than the 2017 election’s score of nine,^{11The low DV score in 2017 can in part be explained by the ‘two-party’ squeeze at that election and consequent more proportional results (especially for Labour).} but much more in line with the DV scores for previous elections. The DV score for the 2015 general election was 24, while for the 2010 election it was 22. The prior post-war record was 23 in 1983.^{12Some historic DV scores can be found here: Dunleavy and Gilson (2010). ‘How unfair or disproportionate is the UK’s voting system for general elections?’. LSE British Politics and Policy Blog, 16 March https://blogs.lse.ac.uk/politicsandpolicy/how-unfair-or-disproportionate-is-the-uk%E2%80%99s-voting-system-for-general-elections}

The DV score for the UK is already quite high, but this overall score actually masks some significantly high DV scores across the nations.

The DV score for England is 17.5, while Scotland’s is much higher at 36. The high Scottish DV score can be explained by the SNP’s performance in terms of seat share, compared to its share of the vote. The DV for Wales is relatively low (14.2), while in Northern Ireland the figure is much higher (30). These results indicate that over a third of seats in Scotland were ‘unearned’ in proportional terms, while slightly less than a third were unearned in Northern Ireland.

Table 9: Deviation from Proportionality Score in English Regions

Even though the DV score for England as a whole is lower than Scotland and Northern Ireland’s, there are significant differences across English regions. Around a third of seats in the South West, South East and East of England were ‘unearned’ in proportional terms, while over a quarter of seats were ‘unearned’ in the East Midlands. Around a fifth of seats were unearned in London, the North East, West Midlands and Yorkshire and the Humber.

As we found in our report on the 2019 local elections, it would seem that these differences even each other out, with different parties overperforming in specific regions.^{13Simpson (2019). Democracy Denied: The 2019 Election Audit. https://www.electoral-reform.org.uk/latest-news-and-research/publications/democracy-denied-the-2019-election-audit} This helps explain the lower DV score for England as a whole.

Smallest Share of the Vote Needed to Win

In seats where more than two parties were in contention, winners were frequently elected on a small percentage of the vote. The smallest of these being in South Down where the winning MP gained just 32 percent of the vote share – this means that over two-thirds of voters in South Down voted for another candidate.

Overall, 229 of the 650 MPs were elected on less than 50 percent of the constituency vote – in other words, 35 percent of all MPs lack majority support.

Table 10: Top 10 Smallest Winning Vote Shares

Table 11: Top 10 Smallest Margins of Victory

Largest Winning Majorities

Another symptom of Westminster’s electoral system are candidates winning with huge majorities – piling up votes far beyond the amount needed to claim victory. Though indicative of a party’s support in specific areas, such large winning majorities mean that thousands of votes have no effect on the overall outcome.

Table 12: Top 10 Largest Winning Margins

The huge majorities won in these constituencies reflect a broader trend of certain votes consolidating in certain areas. The problem of this geographical concentration of votes for parties is that this increased support does not result in greater representation, only larger majorities for those MPs who have already crossed the line. FPTP rewards the most geographically efficient vote spread – which means it wastes a lot of votes which are geographically concentrated.

Ignored Votes

Under Westminster’s one-party-takes-all voting system, only a small subset of votes secure representation – those that are decisive in securing a candidate’s election. Votes cast for non-elected candidates, or for winning candidates which are over and above what they need to be elected, are thrown on the electoral scrapheap and do not influence the outcome of the election.

At the 2019 general election, a staggering number of voters saw their vote count for nothing. Of the 32 million votes cast, only 9.4 million votes (29.2% of the total) were ‘decisive’ in securing a candidate’s election (i.e. were needed to elect the winning candidate).

Across the UK, 14.5 million people (45.3% of all voters) cast their vote for a non-elected candidate, while 8.1 million votes (25.5%) were ‘surplus’, i.e. they were cast for the elected candidate but did not contribute to their election. All in all, over 22.6 million votes (70.8%) did not contribute to electing an MP. In seven constituencies, over 90 percent of the votes went to waste in this way.¹⁴ These constituencies are: Liverpool Riverside (92.2% of votes were ignored), Knowsley (91.9%), Bootle (90.8%), Liverpool West Derby (90.6%), Manchester Gorton (90.5%), Liverpool Wavertree (90.3%), and Liverpool Walton (90.1%). In contrast, with STV, surplus votes are redistributed to voters’ other preferences.

Figure 3: Votes cast in 2019

Table 13: Top 10 Constituencies with the Largest Number of Ignored Votes

Table 14: Top 10 Constituencies with the Largest Number of Votes for Non-Elected Candidates

Looking at the votes cast for non-elected candidates in more granularity shows that voters in Scotland and Northern Ireland fare particularly badly, with the choices of 54 percent (Scotland) and 55 percent (NI) of voters going to non-elected candidates. This means that over half of voters in these areas do not have an MP they voted for.

Looking further at the proportion of votes going to non-elected candidates by party, reveals how the voting system has treated voters of different parties unfairly. Overall, across the UK, over half (50.6%) of Labour voters saw their votes go unrepresented, compared to just under a quarter (24%) of Conservative voters, with even fewer (19.7%) votes for the Conservatives going unrepresented in England.

Supporters of parties with strength spread more thinly throughout the UK fared even worse than Labour supporters. The Liberal Democrats achieved nearly 3.7 million votes, yet 92 percent of their votes are unrepresented. Over 96 percent of 865,697 Green Party votes are unrepresented,¹⁵This includes voters for the Green Party of England and Wales, the Green Party of Scotland and the Green Party in Northern Ireland while all of the Brexit Party’s 644,255 votes are unrepresented.

Given the increasing geographical concentration of votes, voters of different parties suffered from this unfairness in different parts of the UK. Labour voters were particularly disadvantaged in the English midlands and southern England, outside of London. For example, in the East of England, where Labour received just under 750,000 votes, 84 percent of these voters saw their vote go unrepresented, compared to just five percent of Conservative voters in the region.

In the South East of England region, the Liberal Democrats recorded just under 850,000 votes (18.2% of the total, up 7.7 points on the 2017 general election), yet saw only one MP elected (down one on the last election). That means 96 percent of their voters in this region went unrepresented.

While on this occasion Conservative voters fared best overall in achieving representation for their votes, they still suffered in some places. In London, over half (55.2%) of Conservative voters went unrepresented, while only 17 percent of Labour voters did.

In Scotland, voters of both the Conservatives and Labour suffered, with 80 percent of Conservative voters and 95 percent of Labour voters going unrepresented, compared to just 15 percent of SNP voters.

Green Party voters were shortchanged in the South West region of England, where they received 115,011 votes (3.8%, up 1.5 points on the last general election). All of these voters went unrepresented. The over-100,000 Brexit Party voters (8.1%) in North East England suffered the same fate, with every single one of them going unrepresented.

Electoral Pacts and Tactical Voting

One of the most striking features of the 2019 election campaign was the focus on electoral pacts between parties and tactical voting. That these issues were key talking points reflects the dysfunctional nature of Westminster’s electoral system. Under proportional systems, tactical voting is far less of an issue: what you vote for is what you get.

During the run up to the election and throughout the campaign, we asked voters whether they intended to vote tactically.^{16BMG polling for ERS (fieldwork 7–12 August 2019; 8–11 October 2019; 5–8 November 2019; 27–29 November 2019), sample 1,502–1,630 GB adults 18+. Data weighted.} The percentage of voters thinking of opting for a tactical vote, instead of voting for their first choice of candidate or party, increased as we got closer to the election. Between August and November 2019 (the start of the official campaign) between 22 and 24 percent of voters said they would choose ‘the best-positioned party/candidate to keep out another party/candidate that I dislike’. Our final poll before the election found nearly a third of voters saying they would vote tactically in this way (30%).^{17BMG polling for ERS (fieldwork 27–29 November 2019), 1,630 GB adults 18+. Data weighted.}

This hardening of the tactical vote perhaps reflects voters coming around to the realities of the contest under FPTP and the increased media discussion around tactical votes and various tactical voting websites. Certainly there was a greater focus on tactical votes during this election campaign than there had been in 2017, when one in five people were planning a tactical vote.^{18Garland and Terry (2017). The 2017 General Election: Volatile Voting, Random Results. https://www.electoral-reform.org.uk/latest-news-and-research/publications/the-2017-general-election-report}

After polling day, we asked voters whether they had in fact cast a tactical vote. In a large post-election poll, conducted by YouGov for the ERS, 32 percent of voters said they voted tactically.^{19YouGov polling for ERS (fieldwork 13–19 December 2019, online), 8,237 GB adults. The figures have been weighted and are representative of all GB adults (aged 18+).} Tactical voting was slightly higher amongst those who had voted Labour or Liberal Democrat (36% and 39% respectively) compared to Conservative voters (30%). Perhaps unsurprisingly, Remain voters were more likely to vote tactically (35%) than Leave voters (29%).

Results from certain seats suggest tactical voting took place. For example, in Canterbury, where the previously selected Liberal Democrat candidate withdrew and recommended a vote for Labour’s Rosie Duffield,^{20BBC News (2019). ‘General election 2019: Lib Dem candidate stands down to avoid Remain vote split’, 12 November. https://www.bbc.co.uk/news/election-2019-50398820} the Labour vote share went up by three points and the Liberal Democrat vote share went down by two points. This bucked the overall trend in the South East, where the Labour vote share went down by 6.5 percentage points and the Liberal Democrat vote share went up by eight percentage points.

Conversely, in Esher and Walton (also in South East England), where it had been widely reported that the Liberal Democrats could defeat Dominic Raab, the Foreign Secretary,^{21Woodcock (2019). ‘Dominic Raab among Brexiteers at risk of losing seats due to tactical voting, latest polling analysis reveals’. The Independent, 11 December.  https://www.independent.co.uk/news/uk/politics/general-election-dominic-raab-brexit-tactical-vote-boris-johnson-a9242236.html} the Liberal Democrat vote share increased by 28 points and the Labour vote share decreased by 15 points.

Electoral pacts also played their part in this election. In November, the Brexit Party announced that they would not stand in any seat that the Conservatives won at the 2017 general election – half of all the seats in Britain –  in a move to consolidate the pro-Brexit vote.^{22BBC News (2019). ‘General election 2019: Brexit Party will not stand in Tory seats’, 11 November. https://www.bbc.co.uk/news/election-2019-50377396}

On the other side of the Brexit debate, the Liberal Democrats, Green Party and Plaid Cymru formed a limited agreement that saw only one of their number stand in 60 seats in England and Wales.^{23Walker and Stewart (2019). ‘Lib Dems, Plaid Cymru and Greens to launch pro-remain pact’. The Guardian, 6 November. https://www.theguardian.com/politics/2019/nov/06/lib-dems-plaid-cymru-and-greens-launch-pro-remain-electoral-pact} Although covering less than one fifth of the number of seats and thus having a smaller impact on the overall result than the Brexit Party withdrawal, this meant that voters in a number of seats were denied the chance to vote for their first-choice party. In Northern Ireland too, the SDLP and Sinn Féin stood down in three seats,²⁴Mairs (2019). ‘Sinn Féin stands aside in key seats in bid to boost pro-Remain candidates against DUP’. PoliticsHome, 4 November. https://www.politicshome.com/news/uk/political-parties/news/107752/sinn-f%C3%A9in-stands-aside-key-seats-bid-boost-pro-remain in order to try to boost the prospects of each other’s candidates (and other Remain supporting candidates) in these seats.

Given the withdrawal of the Brexit Party from so many seats, tactical voting among Remain-inclined voters was the focus of much attention. At least five tactical voting websites were set up by different organisations offering advice on which ‘Remain’ or ‘anti-Conservative’ candidate was best placed to win in particular constituencies – sometimes containing contradictory advice for the same seat.

The variation in some of the tactical voting advice is understandable given different data can be used when producing a tactical voting recommendation. Usually these sites employ either the result in the constituency at the last election or more recent polling data for the constituency, or a combination of the two. But conflicting advice not only has the potential to confuse voters, it also means that, rather than arguing about who has the best policies, parties become embroiled in arguments about which of them is best placed to beat another party in a particular seat.

It was notable that a great deal of media attention was focused on party tactics rather than policies at the election. In-depth research conducted by academics at Loughborough University found that discussion of electoral process issues – including tactical voting, electoral pacts, party divisions, and electoral integrity – dominated media coverage throughout the election campaign.^{25Deacon et al. (2019). General Election 2019. Centre for Research in Communication and Culture. https://www.lboro.ac.uk/news-events/general-election} Electoral pacts, tactical voting sites and parties’ focus on who is best placed to ‘win here’ all detract from the basic democratic premise that voters should be able to vote for who they want to win. More than that, they have a stagnating effect on politics: if only a select few big parties ‘can win here’, voters’ desires for change or new representation are crushed.

Tactical voting and electoral pacts are nothing new. In 2017, the Greens, Liberal Democrats and UKIP all entered electoral pacts going into the election.^{26Garland and Terry (2017). The 2017 General Election: Volatile Voting, Random Results. https://www.electoral-reform.org.uk/latest-news-and-research/publications/the-2017-general-election-report/} Although a perhaps understandable reaction to the iniquities of FPTP, tactical voting is not something that voters should have to consider. They should be free to vote for their first-choice party without fear that their vote would not count. Both the need for parties to step aside and widespread tactical voting would be eliminated by a move to a proportional voting system.

Safe Seats

Another recurring feature of FPTP that was again in evidence at the election was the number of safe seats, where parties are almost certain to win. Before the election, the ERS predicted the outcome in 316 seats, half of all seats in Great Britain. These predictions had a 100 percent success rate.^{27ERS (2019). ‘Electoral Reform Society correctly predict 316/316 safe seat General Election results’, 16 December. https://www.electoral-reform.org.uk/latest-news-and-research/media-centre/press-releases/electoral-reform-society-correctly-predict-316-316-safe-seat-general-election-results} The certainty of safe seats can breed complacency among parties and lead to voters being taken for granted, with safe seats ignored during election campaigns while seats that may change hands are lavished with attention.

BMG polling for the ERS revealed that those living in seats classed as marginal received far more election literature than those seats classed as safe for one party or another.^{28ERS (2019). ‘Research: Voters in swing seats bombarded with leaflets – while safe seats voters are ignored’, 12 December. https://www.electoral-reform.org.uk/latest-news-and-research/media-centre/press-releases/research-voters-in-swing-seats-bombarded-with-leaflets-while-safe-seats-voters-are-ignored/} Just one in four people (25%) in safe seats reported receiving four or more election leaflets or other pieces of communication through their door compared to almost half (46%) of those in potential swing seats. Nearly three times as many people in swing seats (14%) reported receiving 10 or more leaflets or other pieces of communication, compared to just five percent of those in safe seats.

Before the election, the average UK constituency had not changed hands for 42 years, with 192 seats (30% of the total) last changing party in 1945 or earlier, and 65 seats (10% of the total) being held by the same party for over a century. These ‘one-party’ constituencies mean that other parties can build up substantial vote shares in particular areas, yet never achieve the representation they merit.

As with recent general elections, not many seats actually switched party at the 2019 general election, with just 79 doing so.^{29This excludes the seats of the previous Speaker (Buckingham) and the new Speaker (Chorley).} Although this represents a small increase on the 70 seats that changed at the 2017 general election, it still represents just 12 percent of seats across the UK and it is a smaller number than changed hands at either the 2015 general election (111 seats) or the 2010 general election (117 seats).

Changes in votes are not being represented by changes in the House of Commons, with two artificially inflated blocs propped up by safe seats and many voters locked out of having a meaningful influence on our politics.

Electoral Systems

The UK remains the only democracy in Europe to use First Past the Post (FPTP) to elect its MPs. As we have shown in the preceding chapters, this system is totally unfit for purpose, leading to warped outcomes, ignored votes, electoral pacts, and tactical voting. There are many electoral systems which fare much better than FPTP in terms of proportionality, voter choice, and representation. In other words, working for voters.    

Using exclusive polling from YouGov for this report, the ERS has projected the results of the 2019 election in Great Britain under three other electoral systems: Party List Proportional Representation (List PR), the Additional Member System (AMS), and the Single Transferable Vote (STV).^{30As the data covered Great Britain only, our projections under alternative voting systems only look at potential results in GB. YouGov polling for ERS (fieldwork 13–19 December 2019, online), 8,237 GB adults. The figures have been weighted and are representative of all GB adults (aged 18+).}

These systems were chosen as they not only deal with the deficiencies of FPTP but are also already in use in the UK. STV is used in Scottish local elections and all elections in Northern Ireland, apart from UK general elections. It is also currently being considered for elections to the Senedd and for Welsh local elections. AMS is used to elect the Scottish Parliament, the Welsh Senedd and the London Assembly, while List PR was used in Great Britain for elections to the European Parliament.

A full methodological note for these projections can be found in Appendix A. It is important to note from the outset that it is impossible to predict with certainty what electoral results under different voting systems would be. The projections that follow are merely an indication of what the results of the 2019 general election – conducted under FPTP – would have looked like using a different electoral system. It is of course impossible to account for the other changes that would accompany a switch to an alternative electoral system, such as changes in voter behaviour, party campaigning, or the number of parties standing candidates.

Party List Proportional Representation

Party List Proportional Representation (List PR) is one of the most commonly used electoral systems around the world. List PR systems vary depending on whether voters cast their vote for a party (closed list) or can vote for their preferred candidate within a list (open list). Between 1999 and 2019, closed List PR was used in Great Britain to elect members of the European Parliament (MEPs).

Table 15: 2019 List PR Projection

We include votes for the Speaker in the Labour figure in the projection to ensure results reflect the fact that the Speaker’s constituency was a Labour seat prior to him taking on this role

We modelled the results of the 2019 general election using the D’Hondt system of List PR, as previously used in Great Britain for European Parliament elections. Our projection shows that the Conservatives would still have been the largest party in the House of Commons with 288 seats, which is in line with their having received a plurality of votes at the election. But, based on our model, they would not have gained an overall majority. Conversely, parties with a less geographically concentrated support base would have obtained seats more in proportion to their share of the vote.

List PR systems score highly in terms of proportionality, but – especially in the closed list variant – they limit voter choice, because electors are forced to vote for a list pre-determined by a party and cannot nuance their choice by ranking candidates, as in preferential systems. Though the open list variant can increase voter choice, there is often a weaker constituency link in List PR systems as voters elect a slate of candidates from a larger area than under other electoral systems. Reducing constituency sizes might improve on local representation, but this would then affect proportionality.

Additional Member System

The Additional Member System (AMS) – also known as Mixed Member Proportional Representation (MMP) outside of the UK – is a hybrid voting system. It combines elements of First Past the Post (FPTP), where voters choose one candidate to represent their constituency, and Party List Proportional Representation.

In AMS elections, voters choose a constituency candidate (elected under FPTP) and have a second vote for their preferred party. Each party will publish a list of candidates in advance. Voters can cast both votes for the same party or vote for different parties in their constituency and regional ballots. List seats are then allocated to parties on a proportional basis, usually applying some form of electoral threshold (generally 5%).

List seats ‘top up’ and partially compensate for the disproportionality associated with the FPTP element of the system, by taking into account how many constituency seats have already been won by a party. For example, if a region has 10 seats (five constituency seats and five list seats), and a party wins half the list votes and three constituency seats, then it should win an additional two list seats to reach ~50 percent vote share it received.

The design of AMS systems can differ quite considerably. One significant difference is the ratio of constituency to list seats, which has consequences on the proportionality of this voting system. While the Scottish and Welsh versions of AMS are quite similar to each other, especially when compared to those used in Germany or New Zealand, they differ with regards to the proportion of constituency and list MPs. In Scotland, the proportion of ‘top-up’ list MPs is much higher than in Wales (43.4% compared with 33.3%), which means that the Scottish version of AMS returns a more proportional parliament. In our modelling, we have opted for a 50:50 ratio of constituency to list seats, which leads to a much more proportional outcome than the versions of AMS used in Scotland and Wales. We also applied a five percent electoral threshold.

Table 16: 2019 AMS Projection

Table 17: Percentage Seats with AMS and under FPTP

Given the relatively high proportion of list seats, our AMS projection delivers a broadly proportional result, with seats more closely matching how people voted at the 2019 election. The Conservatives, Labour and the SNP all lose seats in our projection, while smaller parties benefit considerably from the list seats.

AMS is often considered to be a compromise system, as it combines the constituency link of FPTP with the proportionality of list PR. As shown above, with a higher percentage of list seats, it can typically produce relatively proportional results, but has the same problems of safe seats and wasted votes as FPTP. Further, as AMS uses closed party lists, parties still have a lot of control over who gets elected. Voters who particularly dislike a candidate at the top of their preferred party’s list, or like a candidate from a party they otherwise do not support, are unable to express this at the polling station. Voter choice thus remains constrained – though compensated by fairer representation than under FPTP.

Single Transferable Vote

The Single Transferable Vote (STV) is a form of proportional representation which uses preferential voting in small, multi-member constituencies of around three to six MPs. It is used in Northern Ireland for all non-Westminster elections, Scottish local elections, the Republic of Ireland, Malta and the Australian Senate.

STV maintains a constituency link and strong representation, while enhancing voter choice and leading to much more proportional outcomes than FPTP. Under STV, each voter has one vote, but they can rank candidates in order of preference. Voters vote by putting a ‘1’ next to the name of their favoured candidate, a ‘2’ next to the name of their next favoured candidate, and so on. Voters can rank as many or as few candidates as they like. If a voter’s preferred candidate has no chance of being elected or has enough votes already, their vote is transferred to another candidate according to their preferences.

STV ensures that very few votes are ignored when compared with FPTP. It also ensures maximum voter choice, as electors can rank their choices both within and between parties and independents. As a slate of MPs is elected from a slightly larger area than under FPTP, STV also keeps the constituency link while ensuring that the diversity of opinion in the country is fairly represented in parliament.

Table 18: 2019 STV Projection

We include votes for the Speaker in the Labour figure in the projection to ensure results reflect the fact that the Speaker’s constituency was a Labour seat prior to him taking on this role.

Table 19: Percentage Seats using STV and under FPTP

We include votes for the Speaker in the Labour figure in the projection to ensure results reflect the fact that the Speaker’s constituency was a Labour seat prior to him taking on this role.

Our STV projection shows a result that is more proportional and more in line with how people voted at the 2019 general election, with no party gaining an overall majority of seats. Based on our STV projection, the Conservative Party secures 312 MPs (49.4% of all GB MPs), just shy of a majority of seats in the House of Commons and more in line with their percentage of the vote in Great Britain (44.7%).

Though Labour and the Conservatives slightly outperform their vote share in terms of seats under STV, smaller parties’ seat share in our model is much more similar to how people actually voted at the election, with the Liberal Democrats making significant gains in our projection (an additional 48 MPs, leading them to have 9.3% of seats on 11.8% of the vote). As previously stated, smaller parties would have campaigned very differently in an actual STV election.

Table 20: 2019 Alternative Projections

We include votes for the Speaker in the Labour figure in the projection to ensure results reflect the fact that the Speaker’s constituency was a Labour seat prior to him taking on this role.

Conclusion

When the vast majority of votes go ignored, something is seriously wrong with our political system. The 2019 general election showed, once again, that Westminster’s voting system is short-changing both voters and parties. Huge numbers of votes thrown on the electoral scrapheap and significant disproportionality in the results – for all parties – should sound alarm bells. This one-party-takes-all approach is skewing the electoral map and leaving voters locked out of their one chance to have a say in the running of their country.

Our analysis of the results and modelling of those results under different electoral systems shows both the extent of First Past the Post’s failings and that it does not have to be this way. Elections do not have to be held under a system that forces voters into tactical choices – voting against their preferred party – or that forces parties to game the system with electoral pacts, denying their voters the choice to vote for them. FPTP is not fit for purpose for either parties or voters.

The contrast with Ireland’s recent election is stark. As one voter there summed it up to us: “In Ireland, your vote matters, no matter where you live. But First Past the Post makes you feel your vote’s impact depends on your postcode. It is an issue of vibrancy compared to stagnation.”^{31ERS (2020). ‘If people want change in Ireland, they can get it at the ballot box’, 11 February.} Let’s choose the former.

Note on Methodology

Projecting how results of First Past the Post elections would translate into seats under different electoral systems is an imperfect task. Using such results as a baseline means that any projection still incorporates FPTP’s deficiencies – such as tactical voting considerations and the lack of genuine multi-party competition – which would not be the case under more proportional systems.

There are some ways to mitigate against these restrictions to ensure that projections are more similar to what would be the case under PR systems. As in previous years, we commissioned a post-election survey from YouGov to ask how people would have voted if they had been allowed to express a first and second preference. The survey ran online from 13th to 19th December 2019 and returned a sample size of 8,237 adults (18+) in Great Britain. The question asked was:

Please indicate how you would have voted in the General Election on Thursday 12th December, if you had been asked to rank the parties in your order of preference. Put 1 for your most preferred party, then 2 for your second party. You may rank two choices as you wish. If you would not vote, or do not know how you would vote, tick the boxes below.

Given the different party system in Northern Ireland and the fact that our YouGov poll covered Great Britain only, projections under alternative voting systems only look at potential results in the 632 constituencies in Great Britain. Votes for the current Speaker of the House of Commons were reallocated from ‘other’ to the Labour Party to account for the fact that, prior to taking on the speakership, he was a Labour MP and to avoid skewing the results with regards to ‘other’ parties/independent candidates.

List PR Methodology

For projections under this system, we followed the variant of List PR used in Great Britain for elections to the European Parliament (Northern Ireland used STV for these elections). We divided all 632 GB constituencies up into the 11 regions used for European Parliament elections, keeping the number of seats the same as those for Westminster elections.

Votes for each party were added up across constituencies for each region and seats were allocated on the basis of the D’Hondt formula, without applying an electoral threshold. The D’Hondt formula allocates the party with the most votes a seat in rounds, with a party’s votes divided by the number of seats it has won plus one. These rounds continue until all seats have been assigned.

Additional Member System Methodology

AMS combines First Past the Post and List PR seats. The calculations used for our AMS projections thus involved a two-step process.

First, we allocated constituency FPTP seats, using the same AMS constituencies as in our 2017 report. As we opted for a 50:50 ratio of constituency to list seats, these new AMS constituencies were usually created by combining two existing FPTP seats into a single AMS constituency. In some cases, because of an odd number of seats in a region, tricky geography or special exemptions (e.g. the Isle of Wight), single FPTP constituencies were kept. We added up a party’s total votes and calculated their new vote share in each AMS constituency. As these seats are allocated under FPTP, the party with the most votes in each constituency was the winner.

Second, we allocated list seats on the basis of the 11 government office regions. To control for the effects of tactical voting on the actual general election results, we recalculated parties’ vote shares on the basis of voters’ first preference results in the YouGov poll by region. This meant that, for example, if 90 percent of those who voted for the Conservative Party in London ranked the party as their first preference in the poll, then 90 percent of Conservative votes in that region were assumed to be a first preference.

These ‘rejigged’ vote shares allowed us to recalculate the total votes each party received in that region. Similarly to List PR, we then used the D’Hondt formula to allocate seats to each party, based on their recalculated total votes per region. Unlike List PR, we applied a five percent electoral threshold to each region and, as list seats were compensatory, we took into account how many seats each party obtained under the FPTP element to calculate the number of list seats to allocate.

Single Transferable Vote Methodology

For our STV projections, it was necessary to work on the basis of new constituencies. These were the same as in our 2017 report and had been created by aggregating existing FPTP constituencies into new three- to six-member seats.^{32There are two exceptions to this: Na h-Eileanan an Iar (the Western Isles) and Orkney and Shetland remained single-seat constituencies.}  STV constituencies were drawn up to reflect local communities as much as possible.

Parties’ votes were added up and their vote shares were calculated for each STV constituency. We then recalculated each party’s vote share in an STV constituency on the basis of first preference results in the YouGov poll. This was done by region, meaning that the same ‘rejigged’ formula was applied to each STV constituency in an existing government office region.

We then proceeded to allocate seats using the droop quota, which means that, to win a seat, a candidate must receive a vote equivalent to the total number of votes cast divided by the number of seats to be allocated plus one. For example, in a three-seat constituency, the droop quota is equivalent to 25%. Any party which reached the quota was allocated a seat. Seats were awarded on the basis of how many quotas of support (e.g. combinations of 25%) a party won. So, a party winning 50% of the vote in a three-member constituency was allocated two seats.

If no party achieved the quota, the party with the lowest vote share was eliminated and its vote share was redistributed to other parties using a formula based on the second preference results in the YouGov poll. This process continued until all seats were allocated. In very limited cases when awarding the final seat, no party reached a full quota so the party with the highest vote share was awarded the seat.

This modelling is of course only an approximation of the allocation of seats and transfers under STV and relies on a limited number of preferences (in a real-world STV election it is likely that a voter would rank more than two candidates/parties). But it does give an indication of how votes would transfer under STV and offers an insight into how voters’ choices would be translated into seats.

Ignored Votes by Party, and Nation and Region

Ignored votes are the sum of surplus votes and unrepresented votes (votes to non-elected candidates).

Conservative Party

Labour Party

Liberal Democrats

Brexit Party

Green Party

Scottish National Party

Plaid Cymru

Democratic Unionist Party

Sinn Féin

Social Democratic and Labour Party

Alliance

Ulster Unionist Party

Others/Independents (GB)

Others/Independents (NI)

National/Regional Breakdown

UK

England

Scotland

Wales

Northern Ireland

East of England

East Midlands

London

North East

North West

South East

South West

West Midlands

Yorkshire and the Humber