How did votes in the 2017 General Election reflect those in the EU Referendum? This post explores the relationship between the vote share for particular parties at the 2017 General Election and the percentage of estimated Leave votes in UK constituencies. While we don’t know how individuals voted in these elections – due to the secret ballot – we can rely on constituency level data and estimates, as well as survey data on people’s voting intentions.
Party vote share and the Leave vote
It appears that the way constituencies voted in the EU Referendum had a moderate effect on the 2017 General Election results. This chart shows the relationship between how constituencies voted in the EU Referendum and the 2017 General Election results:
Source: Chris Hanretty, Areal Interpolation and the UK’s referendum on EU membership, Journal of Elections, Public Opinion and Parties, 17 March 2017; BBC ward data
Around 19% of the variation in the 2017 Conservative vote share is due to variation in the Leave vote, meaning that other more traditional factors, such as age and class, might have had a stronger impact. Labour vote share appears to have no direct relationship with the referendum vote; constituencies which voted Leave or Remain were equally likely to vote Labour in 2017.
The Brexit effect
One of YouGov’s post-election surveys notes that Brexit did not have the anticipated effect on the main parties’ voter base. Excluding people who did not vote in the referendum, 71% of the 2017 Conservative supporters were Leave voters and 29% Remain. This is compared with 61% and 39%, respectively in 2015. Among the 2017 Labour voters, 71% voted Remain and 29% Leave. There’s little change from Labour’s Remain/Leave divide in 2015: 67% and 33%, respectively.
The link between the Leave vote and UKIP support remains: the higher the Leave vote, the higher the 2017 UKIP vote share. Almost half the variation in the UKIP vote share is explained by the Leave vote. However, this is weaker than the relationship between its 2015 vote share and the referendum vote.
Conservative/UKIP swing and the Leave vote
UKIP’s national vote share declined from 12.6% at the 2015 General Election to 1.8% at the 2017 General Election. Is Brexit the underlying factor and if so, who picked up UKIP’s supporters?
The chart below shows the relationship between UKIP/Conservative swing in constituencies where both parties stood in 2015 and 2017 and the % Leave vote in those constituencies. Electoral swing is used to compare the results of two elections in the same area. It measures the change in vote share for one party in respect to another. A positive value means that there was a swing from UKIP to Conservative, while a negative value denotes a swing from Conservative to UKIP.
Source: Chris Hanretty, Areal Interpolation and the UK’s referendum on EU membership, Journal of Elections, Public Opinion and Parties, 17 March 2017; BBC ward estimates
There is a strong relationship between the % Leave vote in a given constituency and its swing from UKIP to Conservative. This means that the higher the Leave vote, the higher the swing in 2017 from UKIP to the Conservatives. In fact, around 70% of the UKIP/Conservative swing can be explained by the Leave vote. This supports the idea that individuals who previously voted UKIP and Leave in the referendum moved disproportionally to the Conservatives in 2017. This is also in line with data from Lord Ashcroft’s post-election survey, where around 57% of 2015 UKIP voters said they voted Conservative in the 2017 General Election.
*Note that there are no comprehensive official statistics on how individual UK constituencies voted in the 2016 EU Referendum. This analysis uses a combination of estimates of the referendum results by constituency, published by UEA academic Dr. Chris Hanretty and actual results, provided by BBC ward data. More information on his statistical method and a list of referendum results by constituency can found in our previous post: Brexit: votes by constituency
Picture credit: Verification underway, by Epping Forest District Council. Licensed by CC 2.0 /