The 2017 General Election, Brexit and the Return to Two-Party Politics: An Aggregate-Level Analysis of the Result1
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
The Political Quarterly
The 2017 General Election, Brexit and the Return
to Two-Party Politics: An Aggregate-Level
Analysis of the Result1
OLIVER HEATH AND MATTHEW GOODWIN
Abstract
The outcome of the 2017 general election—a hung parliament—defied most predictions. In
this article, we draw on aggregate-level data to conduct an initial exploration of the vote.
What was the impact of Brexit on the 2017 general election result? What difference did the
collapse of UKIP make? And what was the relative importance of factors such as turnout,
education, age and ethnic diversity on support for the two main parties? First, we find that
turnout was generally higher in more pro-remain areas, and places with high concentrations
of young people, ethnic minorities and university graduates. Second, we find that the Con-
servatives made gains in the sort of places that had previously backed Brexit and previously
voted for UKIP. But, third, we find that the gains the Conservatives made from the electoral
decline of UKIP were offset by losses in the sort of places that had previously supported the
Conservatives, particularly areas in southern England with larger numbers of graduates. The
implication of these findings is that while a Brexit effect contributed to a ‘realignment on the
right’, with the Conservative strategy appealing to people in places that had previously
voted for UKIP, this strategy was not without an electoral cost, and appears to have hurt the
party in more middle class areas.
Keywords: UK general election, Brexit, vote change, turnout
Theresa May, who had become the party’s
Introduction leader only eleven months earlier, the Con-
IN JUNE 2017 Britain went to the polls for a servative party polled 42.3 per cent of the
nation-wide vote for the third time in two vote, an increase of 5.4 points on its result in
years, and for the third time the result con- 2015 and its highest share of the vote since
founded most expectations. Ahead of the Margaret Thatcher’s landslide in 1983. But in
election, all but one of Britain’s forecasters the eyes of many it was a pyrrhic victory.
had predicted a majority Conservative gov- The Conservatives won only 317 seats, thir-
ernment, with the predicted size of these teen fewer than in 2015 when David
majorities ranging from a low of forty-eight Cameron had won a small but surprising
to a high of 124, with a mean majority of majority.
seventy.2 Prior to the election campaign, 335 For the Labour party, meanwhile, an elec-
‘expert’ academics, pollsters and journalists tion that some had feared would culminate
surveyed by the Political Studies Association in an historic loss and perhaps a devastating
similarly forecast an average Conservative split instead ended with heroic defeat. Prior
majority of ninety-two, though most to the campaign, when in some polls Labour
expected a majority of at least one hundred.3 had languished twenty-one points behind
The eventual result—a hung parliament— the Conservatives, some had talked of the
was thus a shock. party slumping to its lowest number of seats
For the incumbent Conservative party, a since the 1930s. But under the leadership of
general election that had been called to radical left-winger Jeremy Corbyn, Labour
stamp its authority on Parliament ended in a outperformed the final opinion polls by an
humiliating retreat. Under the leadership of average of five points, receiving forty per
© The Authors 2017. The Political Quarterly © The Political Quarterly Publishing Co. Ltd. 2017
Published by John Wiley & Sons Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA 1cent of the vote—the party’s highest share cent two years later. The collapse of the UK
since Tony Blair’s second landslide in 2001, Independence Party (UKIP), which had cam-
and the third highest since 1970. Corbyn and paigned for Brexit, was especially noticeable.
Labour won 262 seats—an increase of thirty Compared to 2015, when UKIP had received
on 2015 and the largest number since Blair’s almost four million votes, only two years
third and final victory in 2005. Compared to later UKIP’s number of candidates fell from
2015, Labour increased its share of the vote 624 to 377, its share of the national vote
by 9.5 percentage points, the party’s largest dropped from almost thirteen per cent to just
single advance since the election of Labour 1.8 per cent, and the number of seats in
and Clement Attlee in 1945. which UKIP polled at least ten per cent of
In the end, however, only seventy seats the vote crashed, from 450 to only two (the
changed hands. The Conservatives gained eastern Conservative-held seat of Thurrock
twenty seats but lost thirty-three. Most of and the northern Labour-held seat of Hartle-
their gains came in Scotland, where they pool).4 Compared to their results in 2015, the
gained an extra twelve seats, and most of Scottish National Party (SNP), Liberal Demo-
their losses came in London and the South crats, Plaid Cymru and the Greens also
East, where they lost six seats each, respec- recorded declines in their vote share.
tively, and the South West where they lost In this article, we put the results of the
four seats. By contrast, Labour gained thirty- 2017 general election under the microscope
six seats and lost six seats. They made net to explore, at the aggregate level, the factors
gains across the country, winning an extra that thwarted Conservative hopes for a com-
six seats in Scotland, three seats in Wales, manding majority, defied predictions of a
and twenty-one seats across England. Corbyn-led catastrophe for Labour and pro-
Confronted with a hung parliament, Con- duced the second hung parliament in the last
servative expectations of a commanding decade. Building on our earlier work, we
majority were forced to make way for a ‘con- devote specific attention to the role of Brexit
fidence and supply’ arrangement with the in the election.5 While the 2017 general elec-
Democratic Unionist Party (DUP). Yet despite tion will forever be known as the ‘Brexit
neither party managing to secure an overall election’, what exactly was the impact of the
majority, one feature of the election was the Brexit issue on the outcome? How did
return to dominance of the two main parties the collapse of UKIP, a party that during the
(Figure 1). With a combined 82.4 per cent 2010–15 parliament had paved the way for
share of the vote, the two main parties the national vote for Brexit, affect support
received their largest combined share of the for the main parties? And what role was
vote since 1970, and with 26.5 million votes played by other factors, such as turnout, age,
they received more votes than at any previ- social class and education? After providing
ous election since 1951. After decades when an overview of the campaign, we will turn
elections had provided voters with echoes to examine each of these questions in turn.
rather than choices, the 2017 contest was also
one where there was a more meaningful pol-
icy difference between the main parties: Britain in flux: an overview of the
Theresa May’s vision of a ‘hard Brexit’ versus
Jeremy Corbyn’s anti-austerity platform and
campaign
populist cry to represent the ‘many not the The 2017 general election was one that did
few’. Faced with a clear and compelling not need to be called. Theresa May and her
choice, the 2017 election engaged voters to an team had been encouraged by the polls,
extent not seen for the last twenty years, with which during the fortnight before May called
turnout at sixty-nine per cent, the highest the election had put the Conservative party
since 1997. on an average of 42.8 per cent, Labour on
Such trends were mirrored in a slump of 25.5 per cent, UKIP on eleven per cent and
public support for ‘the others’, with the the Liberal Democrats on 10.3 per cent. With
share of the vote going to parties other than an average Conservative lead of 17.3 per-
Labour or the Conservatives falling from centage points in the polls, May stood on
thirty-two per cent in 2015 to just 17.5 per Downing Street and asked the country to
2 OLIVER HEATH AND MATTHEW GOODWIN
The Political Quarterly © The Authors 2017. The Political Quarterly © The Political Quarterly Publishing Co. Ltd. 2017100.0%
90.0%
80.0%
70.0%
60.0%
50.0%
40.0%
30.0%
20.0%
10.0%
0.0%
1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 2020
Conservave Labour Combined 2 party vote
Figure 1: Historic two-party share of the vote
provide a strong mandate for the vision of ‘Blue Labour’ project, which had urged
Brexit that she had set out in the ‘Lancaster Labour to reconnect with the same groups of
House speech’ in January 2017. This voters, was invited to Downing Street to dis-
included leaving the single market and much cuss the strategy with May’s team.6 Between
of the customs union; negotiating a new free the autumn of 2016 and the 2017 election
trade agreement with the EU; ending the free campaign, May and her team launched a suc-
movement principle as it applies to Britain; cession of populist interventions and policy
ending the jurisdiction of the European proposals designed to lure back traditional
Court of Justice in Britain; and putting the social conservatives who had defected to
final deal to a vote in both houses of Parlia- UKIP or felt alienated by Cameron’s ‘mod-
ment. From the outset, therefore, Prime Min- ernisation’, and win over pro-Brexit and eco-
ister May had framed the contest as one that nomically left-behind workers in more
was chiefly about Brexit, an opportunity to traditional Labour areas. Aside from ‘Brexit
‘strengthen the hand’ of the Prime Minister means Brexit’, this narrative of a more inter-
ahead of the Brexit negotiations, and ‘stick ventionist conservatism was cultivated
to our plan for a stronger Britain and take through promises to cap energy prices,
the right long-term decisions for a more increase the national living wage, develop a
secure future’. May had also suggested that new industrial strategy, support grammar
opposition parties and an unelected House schools, give workers more input into the
of Lords were seeking to block or overturn governance of companies, and end the free
the public vote for Brexit, a claim that was movement of EU nationals. May also publicly
reflected on the following day’s front page criticised a ‘liberal elite’ and ‘citizens of
of The Daily Mail: ‘Crush the Saboteurs’. nowhere’, prompting the Financial Times to
But May and her team had also sought to quip how Britain’s Prime Minister ‘has even
broaden the Conservative electorate. Ever mastered the rhetoric of anti-elitism, an
since becoming party leader, May, alongside impressive feat for a woman whose life story
her advisors Nick Timothy and Fiona Hill, takes in the Home Counties, the Bank of Eng-
had sought to downplay Cameron’s more lib- land and Downing Street’.7
eral brand of conservatism and rebuild links Theresa May’s electoral gamble, therefore,
with blue-collar Britain and ‘ordinary work- was that the path to a commanding majority
ing families’ (the so-called ‘OWFs’), including ran through not only retaining the 330 seats
those who had previously voted Labour. that the Conservative Party held at the close
Even Maurice Glasman, a Labour peer of the 2015–17 parliament, an estimated 247
appointed by Ed Miliband and pioneer of the of which had voted to leave the EU and
THE 2017 GENERAL ELECTION, BREXIT AND THE RETURN TO TWO-PARTY POLITICS 3
© The Authors 2017. The Political Quarterly © The Political Quarterly Publishing Co. Ltd. 2017 The Political Quarterlyeighty-three of which had voted to remain, Corbyn and his team produced a clear and
but also by capturing a large number of compelling vision of where they wanted to
Labour’s 229 seats, especially those among lead the country. They promised to only
the estimated 149 that had voted for Brexit increase income tax for people earning over
at the 2016 referendum.8 May’s plan thus £80,000 each year; to abolish university tui-
rested on an assumption that the Conserva- tion fees; to extend free childcare; to raise
tive party would retain votes in the typically the minimum wage; to cap rents so they can
more prosperous, more highly educated, only rise in line with inflation; to ban zero
middle-class and urban areas that had hours’ contracts; to guarantee the state pen-
tended to back remain, while making big sion ‘triple lock’ for pensioners and a winter
inroads among more economically disadvan- fuel allowance; to invest an additional £30
taged, less educated and working-class areas, billion in the National Health Service; and to
often away from the big cities, which had nationalise energy supply networks and rail
voted for Brexit. In this sense, the election companies. In the final week of the cam-
result depended upon the extent to which paign, YouGov asked a representative sam-
the Brexit fault line trumped traditional par- ple of the population whether they felt that
tisan sympathies, and whether there was Corbyn and May had had a good or bad
any electoral cost to heavily targeting leavers campaign. While forty-eight per cent felt that
while maintaining support from remainers. Corbyn had a good campaign and eighteen
During the campaign, however, there was per cent felt the Labour leader had had a
a decisive shift in the public mood. During bad campaign, these figures were almost the
the remainder of April, after May had called reverse for May: only twenty per cent felt
the election, the Conservatives enjoyed an that May had a good campaign while forty-
average lead of nineteen points. Yet, by the eight per cent felt that she had a bad cam-
second half of May this had dwindled to an paign.10 In the final days, the Conservatives
average lead of only 9.5 points and then, in sought to shift public attention back to
the final days of the campaign in June, to Brexit, but in the aftermath of a second
just 6.6 points. Theresa May’s leadership rat- major terrorist attack during the campaign,
ings also declined sharply. According to data the discourse turned to matters of defence
compiled by YouGov on the question ‘who and security on the one hand, and public
would make the best Prime Minister?’ on the sector cuts and the loss of police jobs on the
day the election was called, May enjoyed a other, which Labour had sought to link to
commanding lead of thirty-nine points. But austerity.
by the final week of the campaign this had
slumped to a lead of just thirteen points. In
the public debate these shifts were attributed
Turnout
to an ill thought-out and un-costed mani- After such a tumultuous campaign, it is no
festo which, days after it was published, led surprise that turnout increased. At sixty-nine
to a Conservative ‘U-turn’ on a social care percent, the turnout was the highest for
policy that would have made people pay twenty years, since New Labour’s landslide
more of the costs of social care (a policy that in 1997, while the overall number of votes
was subsequently branded a ‘dementia tax’). cast (more than thirty-two million) was the
Others pointed to the campaigning energy largest since 1992. Yet, turnout was not even
and enthusiasm of Corbyn, who had been across the country. Compared to 2015, some
met by large crowds across the country. places experienced a sharp increase in turn-
During the campaign, people’s ‘net satisfac- out, while others experienced a decline. The
tion’ with how Corbyn was performing as biggest increases in turnout were in London
leader of the Labour party improved signifi- (+4.5) and the North East (+4.8). The biggest
cantly. Whereas shortly before the election decrease in turnout was in Scotland (-4.7).
was called Corbyn had held a disastrous rat- Most other regions saw an increase in turn-
ing of -58, by the final week of the campaign out of around two or three percentage
this had risen to -2.9 This marked increase in points.
Labour’s support in the polls was probably In the aftermath of the result, much atten-
also driven by a series of policy pledges. tion focused on a ‘youthquake’—which
4 OLIVER HEATH AND MATTHEW GOODWIN
The Political Quarterly © The Authors 2017. The Political Quarterly © The Political Quarterly Publishing Co. Ltd. 2017alludes to an apparent sharp rise in turnout of a multivariate analysis of turnout. The
among the young. We have to be careful first model examines the level of turnout in
about how we interpret aggregate level data, each constituency. The second model exami-
but there is evidence that turnout varied nes how much turnout changed in each con-
according to the age profile of different seats. stituency since the 2015 general election.
Of the twenty constituencies in England with From Model 1 we can see that, across the
the highest concentration of young people country, turnout tended to be higher in con-
aged eighteen to twenty-nine years old, turn- stituencies where there were large numbers
out increased by an average of 4.6 percent- of graduates and which were heavily white,
age points. By contrast, of the twenty while turnout tended to be lower in con-
constituencies in England with the lowest stituencies where there were large numbers
proportion of young people, turnout of young people. Taking these factors into
increased by just 2.6 percentage points. Thus, account, turnout was also lower in Scotland
even though ‘older’ constituencies still tend than it was elsewhere in the country. How-
to register higher turnout, in 2017 it was the ever, if we look at seats where turnout chan-
younger areas that saw turnout increase the ged relative to 2015 then a slightly different
most. There is also evidence that turnout pattern emerges. From Model 2, we see that
tended to increase more in seats that had turnout tended to increase most in seats that
voted to remain in the 2016 referendum, have lots of young people, graduates and
with pro-remain seats such as Oxford West people from ethnic minorities. This provides
and Abingdon, Winchester, Twickenham us with some evidence to suggest that at
and Hornsey and Wood Green all recording least with regard to age and ethnic diversity
some of the highest rates of turnout at the there was an upsurge in turnout in areas of
election. According to Hanretty’s (2017) con- the country that two years ago had been
stituency level estimates of the referendum less likely to vote. Also, the seats where
vote, of the twenty constituencies in England turnout was most likely to increase were
which registered the highest support for also those which our earlier analysis
leave in the EU referendum, turnout revealed had also been the most likely to
increased by an average of just 1.6 percent- have voted remain at the 2016 referendum.12
age points.11 By contrast, of the twenty con- This evidence suggests there may have been
stituencies in England which registered the a counter mobilisation against Theresa
highest support for remain in the EU referen- May’s vision of a hard Brexit in places
dum, turnout increased by an average of 5.5 where support for leaving the EU was low.
percentage points. By contrast, people living in those places
We are able to get a sharper understand- which had most strongly supported Brexit
ing of the factors that influenced turnout by in the referendum were rather less likely to
considering a number of different factors vote this time out than they had been in
simultaneously. Table 1 presents the results previous elections.
Table 1: Multivariate analysis of turnout, linear regression (England, Scotland and Wales)
Model 1: Turnout Model 2: Turnout Model 3: Turnout
2017 change change
Coefficient Std. Err. Coefficient Std. Err. Coefficient Std. Err.
% Age 18–29 years -0.42*** 0.03 0.05* 0.02 0.03 0.02
% with degree 0.46*** 0.01 0.03*** 0.01 -0.01 0.02
% non-white -0.04*** 0.01 0.04*** 0.01 0.04*** 0.01
Scotland -2.57*** 0.41 -7.46*** 0.30 -8.13*** 0.45
% Voted Leave -0.04* 0.02
Constant 63.84 0.51 0.96 0.38 4.58 1.85
N 630 630 630
Adjusted R-square 0.66 0.56 0.56
Notes: ***denotes pWe can test this more formally by using our earlier research had found were among
Hanretty’s (2017) constituency level esti- the most likely to back Brexit—seats with
mates of the leave vote. From Model 3, we large numbers of people without educational
see that when we add this variable to the qualifications, a large proportion of working-
analysis, we find that turnout tended to class residents, and which were predomi-
decline in places that had backed leave and nantly older and white. By contrast, the
that the effect of age and education is Conservatives were less likely to make gains
washed out, indicating that perhaps one rea- in places where there were a lot of young
son why places with lots of young people voters and university graduates. Thus, the
and graduates were more likely to vote this social and political factors most strongly
time is because they wanted to register their associated with support for Brexit were
dissatisfaction either with the outcome of the also strongly associated with changes in the
2016 referendum or with the subsequent level of support for the Conservatives. The
direction of the Brexit process. This supports Conservative strategy of focusing on a hard
the general view that people in places which Brexit—and appealing to UKIP voters in
had not voted to leave the EU were particu- particular—therefore appears to have made
larly keen to turn out and vote in this elec- a difference to how well they performed.
tion—which potentially did not bode well Indeed, there is a strong correlation between
for May’s chances of securing a large man- the collapse of the UKIP vote and increased
date for her vision of Brexit. support for the Conservatives.
However, somewhat surprisingly the pat-
tern of changes in support for Labour is not
Analysing the result nearly as polarised along social or political
We now turn our attention to analysing the lines. The correlations are generally much
result of the 2017 general election and the weaker, which indicates that there is not
factors that motivated the change in support such a clear structure to the places where
for Labour and the Conservatives. Given Corbyn and Labour tended to perform par-
that the contest was supposed to be a ‘Brexit ticularly well or badly. This implies that
election’, we can probe a number of key Labour managed to attract a broad coalition
questions. Did areas that had backed Brexit of support, and one that was not especially
at the 2016 referendum move towards the socially distinctive. Thus, arguments that
Conservatives? Did areas that had voted to frame Labour’s performance rather narrowly
remain in the EU lend more support to as a ‘remain backlash’ or a ‘youth revolt’
Labour? How were these shifts, if any, medi- over-simplify the true picture. Nonetheless,
ated by the fate of UKIP, whose support col- a few noteworthy patterns emerge. As
lapsed? Did May and her team succeed in expected, Labour tended to make bigger
capturing the UKIP vote and did this strat- gains in places where there were lots of eigh-
egy come with any electoral costs? teen to forty-four year olds, in places where
To better understand the factors that influ- turnout increased and where support for the
enced support for the two main parties, we Greens had declined. They also advanced in
restrict our analysis to England and Wales, areas that were more ethnically diverse and
as a rather different set of factors are rele- which had a larger proportion of university
vant for understanding electoral competition graduates.
in Scotland, where the SNP is dominant and Some of these patterns are worth explor-
divisions over Scottish Independence are ing a little more. As we have seen, the pro-
more salient. We start by considering some file of places where turnout increased the
social and political factors that plausibly most was similar to the profile of places
may help to shed light on how well the which had been most likely to support
Conservatives and Labour performed in dif- remain. This suggests that there may have
ferent areas. Figure 2 shows the correlation been a backlash against the vote for Brexit
between change in support for the two par- and direction of the Brexit process, which
ties at the constituency level and a number may have mobilised people to vote, and per-
of different factors. The Conservatives were haps also to vote against the Conservatives.
more likely to make gains in places which Some further evidence for this is presented
6 OLIVER HEATH AND MATTHEW GOODWIN
The Political Quarterly © The Authors 2017. The Political Quarterly © The Political Quarterly Publishing Co. Ltd. 2017Leave vote 0.78
-0.27
Turnout change -0.25
0.18
UKIP decline 0.78
-0.07
Green decline -0.51
0.15
LibDem decline 0.15
0.03
Incumbent -0.12
0.13
Non white -0.43
0.16
University degree -0.74
0.12
No qualificaons 0.69
-0.16
Working class 0.71
-0.18
Middle class -0.56
0.00
Age 65 plus 0.42
-0.23
Age 45 to 64 0.41
-0.32
Age 30 to 44 -0.53
0.14
Age 18 to29 -0.33
0.35
-1.00 -0.80 -0.60 -0.40 -0.20 0.00 0.20 0.40 0.60 0.80 1.00
Conservave vote change Labour vote change
Figure 2: Correlates of vote change for the Conservatives and Labour in England and Wales
in Figure 3. Labour tended to make greater constituencies with larger numbers of young
gains in places that recorded an increase in people, whereas the Conservatives tended to
turnout, whereas the Conservatives tended fare somewhat worse. In the twenty ‘young-
to fare somewhat worse. Revealingly, the est’ seats Labour made an average gain of
Conservatives were only able to make about fifteen percentage points, whereas the
greater gains than Labour in those places Conservatives made an average gain of just
where turnout had declined. In the twenty one point.
constituencies where turnout increased the Next, Figure 5 shows the relationship
most, Labour made an average gain of 12.7 between Hanretty’s (2017) constituency level
percentage points, whereas the Conserva- estimates of the leave vote and change in
tives made an average gain of only 0.8 support for Labour and the Conservatives.
points. There is a slight tendency for Labour to per-
One reason why turnout may have made form better in places which supported
a difference is that it brought new groups of remain, though there is a lot of variation
voters to the ballot box who were dispropor- around the line and the gradient of the slope
tionately more likely to vote Labour—such is not particularly steep. Moreover, even in
as the young. Figure 4 shows the relation- places which voted leave, Labour’s share of
ship between the age profile of the con- the vote still improved. By contrast, the rela-
stituency and the change in support for tionship between support for leave and
Labour and the Conservatives, respectively. change in the share of the vote for the Tories
Labour made more substantial gains in is much stronger. Whereas the Tories lost
THE 2017 GENERAL ELECTION, BREXIT AND THE RETURN TO TWO-PARTY POLITICS 7
© The Authors 2017. The Political Quarterly © The Political Quarterly Publishing Co. Ltd. 2017 The Political Quarterly30
Change in party vote share
20
10
0
-10
-10 -5 0 5 10 15
Change in turnout
Conservatives Labour
Figure 3: Change in turnout and change in support for Labour and Conservatives
30
Change in party support
20
10
0
-10
10 20 30 40 50
% Age18 to 29 years old
Conservatives Labour
Figure 4: Age and change in support for Labour and Conservatives
votes in places that were very pro-remain, Conservatives), the Conservatives won two
they gained votes in places that were very seats (down from three) and the Greens
pro-leave. However, even in those places retained one. In these most strongly pro-
that were strongly leave, they did not gain remain seats, Labour’s average vote share
many more votes than Labour did, and it increased by over thirteen percentage points,
was only in the most staunchly leave areas while the Conservatives’ declined by over
of the country where their gains outstripped three points. This represents a swing of eight
those made by Labour. In the vast majority points from the Conservatives to Labour. By
of places, including many which preferred contrast, of the twenty seats in England with
leave to remain, Labour were able to outper- the highest estimated leave votes, Labour
form the Conservatives, and capture an won nine and the Conservatives won eleven
increased share of the vote. (the Conservatives retained seven, gained
Out of the twenty seats in England with three from Labour and one from UKIP).
the highest estimated remain vote, Labour However, even in these pro-Brexit seats,
won seventeen (gaining one from the while the Conservative average vote share
8 OLIVER HEATH AND MATTHEW GOODWIN
The Political Quarterly © The Authors 2017. The Political Quarterly © The Political Quarterly Publishing Co. Ltd. 201730
Change in party support
20
10
0
-10
20 40 60 80
Vote Leave (%)
Conservatives Labour
Figure 5: Estimated support for leave and the change in support for Labour and
Conservatives
rocketed by fifteen points and UKIP’s share ground than in places where UKIP’s vote
plummeted by twenty points, Labour’s aver- held up. By contrast, once again there is a
age vote share still went up by close to much clearer pattern with respect to the
seven percentage points, thus making it Conservatives, who did well in places where
harder for the Conservatives to translate an UKIP lost a lot of votes, but badly where
increased vote share into seat gains. This UKIP only lost a few votes. UKIP needed to
represents a swing of just four points from lose close to ten percentage points of the
Labour to Conservative, which means that vote before the Conservatives saw any
the swing to Labour was about twice as increase in their own share of the vote. This
large in the most pro-remain areas as the suggests that whatever the Conservatives
swing to the Conservatives was in the most gained from UKIP voters moving to the
pro-leave areas. Theresa May and her team Conservatives were offset by losses
had targeted pro-leave Labour areas but in elsewhere.
the end they only captured six seats from Two rather distinct patterns emerge, there-
Labour areas of the country that had voted fore. Labour made large gains across the
for Brexit. country, performing relatively well both in
This suggests that going after such a hard areas that had voted to leave the EU and
Brexit may not have paid the electoral divi- areas that had voted to remain. By contrast,
dends that Conservatives had expected. No the Conservatives made progress in areas
doubt one reason why May, who had cam- that had voted to leave but were punished
paigned for remain during the referendum, in areas that had voted to remain in the EU.
opted for this hard line approach was to try Although they appear to have successfully
and appeal to the nearly four million social harnessed votes that in 2015 had gone to
conservatives who had voted for UKIP in UKIP, they do not appear to have gained
2015. Indeed, before the election, the expecta- these votes without a cost.
tion was that UKIP voters would defect en
masse to the Conservatives. From Figure 6,
we can see how the collapse of UKIP
Was it a Brexit election?
affected support for the two main parties. So far, we have presented evidence to sug-
There is not much of a pattern between the gest that the educational, ethnic and age
collapse of UKIP and change in support for composition of different constituencies had
Labour. On average Labour did not do much an impact on how well the two main parties
worse in places where UKIP lost a lot of performed relative to 2015, and that Brexit,
THE 2017 GENERAL ELECTION, BREXIT AND THE RETURN TO TWO-PARTY POLITICS 9
© The Authors 2017. The Political Quarterly © The Political Quarterly Publishing Co. Ltd. 2017 The Political Quarterly30
Change in party support
20
10
0
-10
-40 -30 -20 -10 0
Change in UKIP vote
Conservatives Labour
Figure 6: Change in support for UKIP and change in support for the two main parties
turnout and the collapse of UKIP may have However, when we look at Models 3 and
influenced the final result. Yet, we need to 4, a slightly different picture emerges. The
be a little careful in terms of how we inter- vote share for the Conservatives tended to
pret these bivariate correlations. For this rea- decline in places where there are lots of
son, we will now focus on the joint impact graduates, and in places that are more ethni-
of these key social and political factors on cally diverse. By contrast, the vote share for
support for the two main parties. Table 2 Labour tended to increase in places where
presents results from a series of linear regres- there were a lot of young people, and where
sion models. In the first two models, we turnout increased. From the model fit statis-
examine how well the Conservatives and tics we can see that there is quite a clear
Labour did in the 2017 election. Thus our structure to the change in the vote for the
dependent variables relate to the vote share Conservatives (R2 = 0.61). By contrast the fit
that each party received in each con- to the data for Labour is substantially
stituency. In the next two models, we exam- weaker (R2 = 0.14), implying that there was
ine how much they improved since 2015. not such a clear pattern to where they made
Thus, our dependent variables relate to the gains and losses, and that their vote gains
change in the vote for each party. were more evenly distributed across differ-
From Model 1 we can see that the Conser- ent areas. Further analysis (not shown) also
vative vote tended to be higher in seats with reveals that Labour also profited from a con-
largely white, highly educated populations, solidation of the anti-Tory vote, and picked
and lower in seats with younger popula- up support in seats at the expense of the Lib-
tions. The Conservatives also did slightly eral Democrats and Greens.13
worse in constituencies where turnout These patterns have many similarities with
increased. Controlling for all of these factors, the patterns that we observed in our analysis
they also tended to do slightly worse in Lon- of the 2016 vote for Brexit. Places where the
don than elsewhere in England and Wales. Tories underperformed were places that were
To a fairly striking degree, Labour’s vote likely to have supported remaining in the EU.
tends to be higher in exactly the opposite The places where Labour made their stron-
kind of places, which is what we would gest gains were—in a slightly different way—
expect in a predominantly two-party system. also the sorts of places where remain had
Both models provide a fairly good fit to the polled most strongly. As Table 2 shows, in
data, and even these parsimonious models 2015 the Conservatives had performed well
are able to explain a lot of the variation in places that were highly educated, but only
between constituencies in the vote share that two years later, they performed relatively
each party receives. poorly in these places. Why did places with
10 OLIVER HEATH AND MATTHEW GOODWIN
The Political Quarterly © The Authors 2017. The Political Quarterly © The Political Quarterly Publishing Co. Ltd. 2017Table 2: Multivariate analysis of support for Conservatives and Labour in England and
Wales, OLS regression
Model 1: Conser- Model 2: Labour Model 3: Conser- Model 4: Labour
vative 2017 2017 vative gains gains
Coef. Std. Err. Coef. Std. Err. Coef. Std. Err. Coef. Std. Err.
% Age 18–29 -1.50*** 0.10 1.46*** 0.11 -0.06 0.03 0.27*** 0.04
% degree 0.46*** 0.06 -0.88*** 0.06 -0.46*** 0.02 0.05* 0.02
% non-white -0.15*** 0.04 0.29*** 0.05 -0.09*** 0.01 0.01 0.02
Turnout -1.53*** 0.20 1.89*** 0.22 -0.12 0.07 0.18* 0.08
change
London -4.17*** 1.87 5.13*** 2.10 1.34* 0.64 -1.91** 0.72
Constant 63.16 33.37 2.21 19.94 0.64 3.95 0.76
N 570 570 570 570
Adjusted 0.54 0.58 0.61 0.14
R-square
Notes: *** denotes pTheresa May and the Conservatives were at registered any benefit from UKIP’s collapse
translating UKIP losses into Conservative in places where there were relatively few
gains. If the Conservative strategy of appeal- graduates. But, in places where there were
ing to pro-Brexit UKIP voters drives down many more graduates, the Conservative vote
support among well-educated Tory remain- share suffered.
ers, then the gains they make among the for- The substantive effect of these patterns can
mer may be offset by losses among the be illustrated by calculating the estimated
latter. In this case, the Conservatives would impact of UKIP’s decline on the Conserva-
only expect to see a net increase in their vote tive vote in different types of areas, as
share in places where there are relatively shown in Figure 7. We can distinguish
few graduates—and might actually see their between areas where there are slightly fewer
vote share decline in places where the con- graduates than average, where there were an
centration of graduates is higher—even if average number, and where there were
they still succeed in capturing votes that slightly above average. All other things
have come from former UKIP supporters. To being equal, Conservative gains are much
explore this possibility, we can investigate higher in places where the UKIP vote
whether there is an interaction between declined a lot than where it declined a more
changes in the UKIP vote and the education modest amount. However, in relatively low-
profile of a constituency. If there is, this skilled areas, the Conservatives were very
would imply that the Conservatives made effective at turning UKIP losses into Tory
more effective gains from the UKIP vote in gains. In these sorts of places an eight-point
low skilled areas (such as North East) than drop in UKIP’s share of the vote translated
they did in high skilled areas (such as South to about a five-point gain for the Conserva-
East and South West). tives. However, in relatively high-skilled
Model 2 in Table 3 provides clear evidence areas the Conservatives were not nearly as
that the Conservatives made stronger gains effective in turning UKIP losses into Tory
in places were UKIP’s share of the vote gains. In these sorts of places, an eight-point
declined (b=-0.82). However, from the posi- drop in the UKIP vote translated into just a
tive sign of the interaction between the one-point gain for the Tories.
change in UKIP’s vote share and the educa- How can we interpret this pattern? One
tional profile of the constituency, we can see possible explanation concerns the behaviour
that the Conservative gains from UKIP were of former UKIP voters themselves. Perhaps
offset by the losses they received in places in high skilled areas UKIP voters were more
where there were larger numbers of gradu- likely to revolt against the Conservatives?
ates (who had previously tended to back the This seems unlikely given that, in places
Conservatives). Thus, the Conservatives only where UKIP’s vote declined by five points,
12
11.1
Change in Conservative vote share
10
8
8
6 5.3
4.8
4
3.1
2
1
0
Ukip lose 8 pts Ukip lose 18 pts
Low % degree Average % degree High % degree
Figure 7: Estimated Conservative gains for UKIP losses
12 OLIVER HEATH AND MATTHEW GOODWIN
The Political Quarterly © The Authors 2017. The Political Quarterly © The Political Quarterly Publishing Co. Ltd. 2017the Conservative share of the vote actually considerable support from the key social
declined, which implies that at least some groups they had been directed toward. Our
people who had previously voted for the analysis provides evidence that the Conser-
Conservatives did not do so this time. A vatives gained in more economically left
more plausible interpretation concerns the behind, heavily white and older areas that
behaviour of Conservative voters, who had previously given disproportionately
might have been alienated by Theresa May’s strong support to Nigel Farage and UKIP.
vision of a ‘hard Brexit’. There thus appears But this narrow strategy also came with
to be a trade-off between the appeals that costs and was much less popular in more
the Conservatives made which had particu- high skilled areas that in 2015 had given
lar resonance in more deprived areas, there- lower support to UKIP, but had still turned
fore allowing them to make substantial gains out for the Conservatives.
at the expense of UKIP, and how those same Labour, meanwhile, managed to attract
appeals were received in more high-skilled support from a broader coalition and one
areas, where they lost votes and needed big that is not nearly as socially distinctive as
swings away from UKIP in order to just that of the Conservative party. The gains
hold support and tread electoral water. that the party made came from a variety of
May’s strategy of aggressively courting the different sources. Jeremy Corbyn’s accep-
2015 UKIP vote might, therefore, have back- tance of Brexit while wanting to protect jobs
fired and been at least partially responsible does not appear to have cost Labour votes in
for the Conservative party losing seats, par- areas that had voted for Brexit and it was,
ticularly in London, the South East and the simultaneously, a much more popular option
South West. in the younger and more diverse areas that
had voted to remain in the EU. Labour’s
impressive performance also appears to have
Discussion: implications of the been helped by an increase in turnout, which
changed the composition of voters. Turnout
result was higher in younger and more ethnically
The 2017 general election was called by diverse seats, and where there were larger
Theresa May in the hope of securing a numbers of graduates. Thus, there were
strong mandate both for her premiership more voters in places sympathetic to Labour,
and her vision of Brexit. The election was a and these voters were in turn more likely to
gamble, with May and her team essentially support Labour than previously. Labour also
placing the opposite bet to that which had benefited from a consolidation of the anti-
been placed by Tony Blair and New Labour Tory vote, particularly from the Liberal
more than twenty years earlier. Whereas Democrats and the Greens
Blair and Co. had gambled that they could Whether Jeremy Corbyn and Labour are
retain support from their traditional blue- able to maintain or even expand this coali-
collar and socially conservative workers tion further to secure a majority government
while reaching into the more liberal, urban remains to be seen, as does the extent to
and university-educated middle-classes, which the Conservative party is able to
May and her team gambled that they could rebuild its support among younger, better
retain support from the more middle-class educated and more pro-remain groups, who
and pro-remain wing of the Conservative will be required for any return to a majority
party while reaching into the more pro- government. The move towards two-party
Brexit, left behind and Labour areas of the politics may also turn out to be a blip rather
country. Our analysis of the results, at the than a permanent realignment, and there is
aggregate-level, has revealed the factors that still the potential for minor parties to reassert
combined to ensure that this gamble would themselves. One thing, however, appears
—ultimately—lose. fairly certain. While the 2017 general election
May’s strategy and a retro manifesto, result defied predictions and also easy expla-
which had included calls to restore fox hunt- nations, the next election—whenever it is
ing alongside strong support for a hard called—will also have the capacity to
Brexit and grammar schools, did win surprise.
THE 2017 GENERAL ELECTION, BREXIT AND THE RETURN TO TWO-PARTY POLITICS 13
© The Authors 2017. The Political Quarterly © The Political Quarterly Publishing Co. Ltd. 2017 The Political QuarterlyNotes 7 ‘Theresa May’s balancing act and the neutering
of UKIP’, Financial Times, 24 February 2017.
1 We’re grateful to Chris Hanretty, Georgios 8 These estimates are taken from data compiled
Xezonakis and Elie Pelling for helpful com- by Chris Hanretty. Aside from the paper in
ments and feedback on earlier drafts. point 11 below see also ‘Revised estimates of
2 Forecasts compiled and shared by Simon Hix; leave vote share in Westminster constituencies’,
https://twitter.com/simonjhix/status/872724 Medium, 18 August 2016; https://medium.c
388768092160 (accessed 18 June 2017). om/@chrishanretty/revised-estimates-of-leave-
3 Political Studies Association, Expert Predictions of vote-share-in-westminster-constituencies-c4612f
the 2017 General Election, London, Political Stud- 06319d (accessed 24 June 2017).
ies Association, 2017; https://www.psa.ac.uk/ 9 Data on net satisfaction of party leaders com-
sites/default/files/PSA%20GE2017%20Expert% piled by YouGov. See the YouGov archive;
20Survey.pdf (accessed 18 June 2017). https://yougov.co.uk/publicopinion/archive/
4 On electoral support for the UK Independence (accessed 24 June 2017).
Party see M. J. Goodwin and C. Milazzo, 10 The remainder said neither/did not know.
UKIP: Inside the Campaign to Redraw the Map of YouGov/Sunday Times, 1–2 June 2017.
British Politics, Oxford, Oxford University 11 C. Hanretty, ‘Areal interpolation and the UK’s
Press, 2015. referendum on EU membership’ Journal of Elec-
5 M. J. Goodwin and O. Heath, ‘The 2016 tions and Public Opinion, published online 17
referendum, Brexit and the left behind: an March 2017, pp. 1–18.
aggregate-level analysis of the result’, The 12 Goodwin and Heath, ‘The 2016 referendum’.
Political Quarterly, vol. 87, no. 3, 2016, pp. 323– 13 When we include data on change in the Liberal
332. Democrat and Green share of the vote in Model
6 ‘May’s manifesto chief held talks with former 4, the R-square increases from 0.14 to 0.44.
Miliband aide’, Financial Times, 9 May 2017.
14 OLIVER HEATH AND MATTHEW GOODWIN
The Political Quarterly © The Authors 2017. The Political Quarterly © The Political Quarterly Publishing Co. Ltd. 2017You can also read
