A. Referendum analysis

Introduction

Appendix A. Referendum analysis begins with a description of the 2016 electorate and those who were included or excluded from it. It then moves on to analysing the result at a high level. As the winning majority was just under 1.3 million and there were 12.9 million eligible voters who did not vote, we consider what might have happened if fewer had abstained.

We also compare the 2026 result with the result of the 1975 referendum which had a much larger majority for staying in the EEC (the predecessor of the EU). Appendix A closes by asking whether the 2016 referendum was the biggest ever democratic exercise in the UK (TLDR, it wasn’t).

2016 EU referendum electorate

The electorate was defined as the 46.5 million people who were entitled to vote in a UK parliamentary election. It comprised:

1. British citizens over 18, resident in UK
2. Irish citizens over 18, resident in UK
3. Maltese & Cypriots over 18, resident in UK
4. Commonwealth-born citizens over 18, resident in UK
5. Commonwealth citizens in Gibraltar over 18
6. British expats, lived overseas for less than 15 years
7. Irish citizens born in Northern Ireland, registered to vote there in the last 15 years

This definition meant that the referendum electorate excluded: 4.6 million potential voters, most of whom would have favoured remaining in the EU.

1. All other EU citizens resident in the UK who are allowed to vote in local elections – estimated at 1.6 million EU registered voters (which includes some Irish, Maltese and Cypriot citizens listed above).
2. Young voters aged 16-18 (unlike the Scottish referendum of 2015) – estimated at 1.5 million voters based on ONS statistics.
3. UK expatriates out of the country for 15 years or more of about 1.5 million, estimated as follows:
   • Government estimates that there are 5 million UK expatriates (but does not estimate how many have been overseas for 15 years or more).
     • for the 1.2 million UK expatriates in the EU, an estimated 300,000 (25%) to 700,000 (58%) were 15-year expatriates.
   • applying a figure at the low end of this range (30%) to the 5 million indicates another 1.5 million potential voters.

As in every election, there were adults who were eligible to vote, but who did not register to vote. However, ONS does not collate data on this group.

Note that the Conservatives had promised in their 2015 election manifesto to restore full voting rights to all expatriates, but had not done so at the time of the referendum. The government reiterated its intention to do so in February 2018.

Eventually, the Elections Act 2022 removed the 15-year limit. Since January 2024, any British citizen living overseas can register as an overseas elector as long as they were previously:

• registered to vote in UK elections, including as an overseas voter, or
• resident in the UK before moving to an address abroad.

When the government introduced the Elections Bill in 2021, the accompanying impact assessment estimated that abolishing the 15-year rule could make 3.2 to 3.4 million British citizens eligible to register as overseas voters.

Sources:
Cabinet Office, A democracy that works for everyone: British citizens overseas Policy Statement, October 2016
Office for National Statistics, Electoral statistics, UK: 2016
House of Commons Library, Overseas voters, April 2026

 

2016 EU referendum result

33.6 million voted – 72.2% of the registered electorate (see Table A.1).

Excluding those that did not vote (Figure A.1):

• 17.4 million (51.9%) voted Leave
• 16.1 million (48.1%) voted Remain
• Leave majority was 1.3 million votes (3.8% of those that voted)

In terms of the eligible electorate of 46.5 million (Figure A.2):

• 17.4 million (37.4%) voted Leave
• 16.1 million (34.7%) voted Remain
• 12.9 million (27.8%) did not vote

Table A.1: 2016 EU referendum result

Figure A.1: 2016 EU referendum result – those who voted

Figure A.2: 2016 EU referendum result – electorate

 

Sources:
Electoral Commission, EU referendum result
House of Commons Library, European Union Referendum 2016, June 2016

 

What if fewer voters had abstained?

Against Leave’s winning margin of 1,269,501 polls showed that most of the 12.9 million voters who did not vote in June 2016 would have backed Remain. See the results of five separate reputable polls based on a report by Business Insider of analysis by Professor John Curtice in Figure A.3.

In round figures, if 10% more abstainers had voted to Remain rather than Leave, the result would have been a tie.

The five polls showed the excess of Remain over Leave ranged from 3% to 31%:

• Leave: 17% to 24% of non-voters
• Remain: 27% to 55% of non-voters
• undecideds: 21% to 50% of non-voters

Let us assume 40% were unshakeably undecided but 60% of abstainers had voted. If their Remain votes had exceeded Leave votes by 16%, the extra votes would have cancelled out the Leave majority. (If only 20% were still undecided, the 16% would have dropped to 12%.)

Although this analysis is purely hypothetical, it demonstrates the importance of campaigning effectively to reduce the ‘don’t knows’ in any election.

Figure A.3: Polls of those that did not vote

Source: Business Insider, Brits who didn’t vote in the EU referendum now wish they voted against Brexit, 23 September 2016

 

Compare 2016 with 1975

The comparison of the 2016 referendum result with the 1975 referendum result is interesting (Table A.2):

• remarkably, in 2016 and 1975, the winning number of votes was virtually the same at 17.4 million;
• the majority in 1975 of 8.9 million (34.4% of the votes) was much greater than the 1.3 million (3.8% of the votes) in 2016:
  • the 1975 democratic mandate was stronger than in 2016;
  • however, turnout in 2016 (72.2%) was higher than 1975 (64.0%).

Table A.2: 2016 EU referendum result compared with 1975 result

Sources:
Electoral Commission, EU referendum result
House of Commons Library, The 1974-75 UK Renegotiation of EEC Membership and Referendum, July 2015

Was 2016 the UK’s largest democratic exercise?

On its own, the size of the electorate does not tell us all we need to know about a democratic exercise. Every General Election is a larger democratic exercise than the last one, because the UK population is growing and the electorate is growing with it.

The other key factor in a democratic exercise is turnout. The turnout for the 2016 referendum was 72.2%. This was much higher than the turnout during the 2024 election of 59.7%, which was the lowest at a General Election since 2001. However, a higher turnout combined with a smaller electorate may lead to more votes being cast. For example, more votes were cast in the 1992 General Election than in the 2016 referendum, which made it the largest ever UK democratic exercise (see Table A.3).

In the history of UK elections, a turnout of 72.2% is not exceptional. Turnout between 1922 and 1997 was 70% or more in 21 General Elections and, in 16 of those, was higher than 72.2%. However, in the eight General Elections since 1997, turnout has been lower, ranging between 59.4% and 67.3%.

Turnout in 2024 was 7.6 percentage points lower than in 2019 of 67.3%. This was the largest drop in turnout between elections since between 1997 and 2001.

Table A.3: UK General Elections — electorate, turnout, votes cast (1974-2024)

Figure A.4 Turnout at UK General Elections , 1918-2019

Sources:
House of Commons Library, Turnout at elections, January 2023
House of Commons Library, 2024 general election: Turnout, September 2024