Deutsche Special Report – UK election night preview
Opinion polls have narrowed significantly as the election approaches, from highs
of a 20 point Conservative lead over Labour to current poll average of 7pp
at the time of writing. While current polling still remains consistent with the
Conservatives increasing their current slim majority, the tightening of polls and
their wide range has increased the likelihood of alternative election outcomes.
Looking towards Thursday’s vote, we analyse the trends and variation in the
recent polls and preview the likely timelines for election night itself as well as the
potential formation of a coalition government. We also identify the major swing
seats to keep under watch as the results unfold including both 2015 marginal
seats and those potentially vulnerable to Brexit-related swings and tactical voting.
Poll narrowing has increased the probability of alternative election scenarios
Following the sharp narrowing of the Conservative’s polling lead in recent weeks,
the Conservatives’ lead over Labour is now down to 7 points. Comparing this
to the historic margins of error seen over previous elections, current polling
remains just in excess of the largest polling miss observed since 1992 – a 6.3pp
overestimate of the Labour vote share in 2001 – and well in excess of the average
polling error over these votes. At the same time, in elections since 1992 the
pollsters have consistently underestimated the Conservative’s vote share, by
2.1pp on average. (chart below).
Given this clear trend for polls to underestimate the Conservatives for various
reasons, including issues of poor sampling together with the “shy Tory
phenomenon” – pollsters attempt to correct their survey results to account for this
dynamic. The risk therefore is that the polls could be over-correcting and therefore
giving the Tories an advantage in the polls.
As the chart below shows, the raw voting intention data behind the most recent
polls from each of the main pollsters is relatively similar, with a limited spread
across companies. There is, however, a much larger dispersion of results when
looking at the final headline poll lead that is published. Indeed, from an average
Conservative lead of 1.3% based on the raw data, the lead increases to 6.3% after
the polling agencies’ re-weighting, with a much wider spread.
In other words, the wide range in the headline Conservative lead across polls
is a result of the statistical re-weighting employed by the polling agencies in an
attempt to correct for previous Conservative polling misses rather than a result
of variation in the raw survey data.
With the Tory poll lead ranging from +1 to +12 across different polls, the key
question for election night will re-weighting will be most correct. The chart above
compares the current re-scaling of voting intention by company relative to the
miss of that polling company in the 2015 general election. Those companies reweighting
the Conservative vote share the most in fact tended to better predict
the 2015 final result.
Thus, while there is significant re-weighting of polls in favour of the Conservatives,
this should not in itself imply the Conservative lead is being over-stated. Indeed,
in 2015 the pollsters currently re-weighting the most – ComRes, IpsosMori and
ICM – tended to get closest to the final result.
Timing on and following election night
■ Polls open from 7am and close at 10pm on Thursday.
■ The publishing of polls is not permitted on the day until after polling
closes– at 10pm Ipsos Mori will publish the official exit poll for the BBC
and Sky News.
■ The 10pm exit poll will also give a projection of the ultimate seats won by
each party. Historically these forecasts have been very accurate (within 15
seats for the winning party) but this means that if the exit poll projections
are tight – i.e. within a 20 seat majority for the Tories – the margin of error
means we may not know by 10pm which of the potential outcomes across
small Conservative majority, Conservative minority or Labour coalition are
the most likely.
■ The vote count begins at 10pm with the first results typically declared
between 11pm and midnight.
■ The bulk of the declarations begin to flow in from about 2am. By around
4am we should have ~50% of the vote counts declared and a good idea
of the result and by 6am close to 90% of the vote count (see chart below).
■ In the case of a clear majority for the winner, the tradition is for the leader
of the winning party to wait for the leader of the losing Party to concede
before claiming victory. In 2015, David Cameron accepted victory just
before 6 am.
Process for forming a coalition government
Given the tightening of the polls, there is a risk that no party gains an overall
majority. This would lead to a hung parliament scenario with either a minority
Conservative government or a Labour-led alliance supported by the Lib Dems and
In terms of the process for forming a coalition – the 2010 negotiations lasted five
days, beginning with the leaders’ speeches on the Friday morning and concluding
with Cameron and Clegg’s joint press statement on the following Wednesday.
The next state opening of Parliament (the Queen’s Speech) is scheduled for the
19th June which would be an important test in the event of a Conservative
minority government. If the vote on the Queen’s speech fails to pass then either a
new government would attempt to be formed or we move towards fresh elections.
Seats to watch on election night
In the UK’s first-past-the-post constituency level system, marginal seats are
crucial in defining the outcome of elections. In this section we compile the
potential swing seats to watch based on different criteria, ranked by the expected
timing of their declarations on election night.
Seats to watch 1: 2015 marginal seats (<5% winning margin)
Marginal seats are typically defined based on the size of the majority of the current
MP relative to the next best party in the last election. In Figure 6 below we list the
most marginal constituencies in Great Britain in the 2015 election – listing the 52
seats in which the margin of the winning party ahead of the runner-up was less
than 5%. We list the seats by the time they are expected to declare their result
according to the Press Association. This expected timing should be seen as an
indication only , with close results increasing the chances of later declarations.
Seats to watch 2: seats vulnerable to Brexit preferences and tactical voting
However, with the Brexit referendum last year and leadership changes across the
major national parties since the last election in 2015, it may be that a traditional
margin-based seat list is less appropriate than in the past. Limiting our analysis
to England & Wales, which contain the majority of marginal seats, we consider
which seats currently held by the two largest parties could be vulnerable based
on attitudes to Brexit and potential scope for tactical voting.
To make a list of potentially vulnerable Conservative seats we apply two criteria:
■ the seat voted more ‘Remain’ than the UK average (or is estimated to
have ) and
■ the Conservative vote share in 2015 was less than the combined share
of Labour, Liberal Democrats and Greens leaving the seat potentially
vulnerable to ‘progressive’ tactical voting
For vulnerable Labour seats we apply the criteria of:
■ the seat being more ‘Leave’ than the UK average
■ the combined vote share of Conservatives and UKIP being greater than
that of Labour, Lib Dems and Greens
The gives two lists of 21 and 33 seats respectively presented in Figure 8 and 7.
A number of these seats, such as Conservatives’ Croydon Central and Labour’s
Halifax appear both on these lists and on the list of marginal seats. However,
this combination of criteria also suggests seats such as the Conservatives’ Bristol
North West and the Labour seat of Wrexham, the latter expected to come quite
early in the night, could be bellwethers as to whether either side is able to take
advantage of Brexit sentiment or tactical voting.
Of the two seats that have changed hands since 2015, one makes the list
– Copeland, which was lost by Labour to the Conservatives earlier this year.
However, Richmond Park – regained by Lib Dems from Tories in late 2016, does
not make it, as the 2015 Conservative majority was too large to meet our criteria.
Seats to watch 3: Scottish seats vulnerable to a swing from SNP to Conservatives
The dominance of SNP in Scotland was the key theme in the 2015 election,
with the 50 seats gained by the SNP thanks to its 50% vote share in Scotland
accounting for almost half of the seats that changed hands in 2015. The question
this year is can the SNP hold on to its gains. Polls for Scotland are limited, but
they suggest that Conservatives should significantly improve from the 15% they
achieved in 2015, most of their gain coming at the expense of SNP. To keep track
of the potential for the SNP to Conservative swing story playing out, we compile a list of
SNP seats in which the Conservatives finished second in 2015 (Figure 9).
This includes the marginal seat of Berwickshire, Roxburgh & Selkirk that already
made an appearance in the marginal seat table above as well as six others.
Special promotion for a limited time
30€/monthly for yearly subscription
6 month offer
40€/monthly for 6 month subscriptions
1 week free trial of Quintessence Signals is available with all offerings