General Election Prediction

Current Prediction: Conservative short 15 of majority

Party2017 Votes2017 SeatsPred VotesPred Seats
CON 43.5%318 29.7%311
LAB 41.0%262 25.4%224
LIB 7.6%12 18.2%43
Brexit 0.0%0 14.6%0
Green 1.7%1 5.1%1
SNP 3.1%35 3.2%50
PlaidC 0.5%4 0.6%3
UKIP 1.9%0 0.8%0
Other 0.7%0 2.5%0
DUP 10 10
SF 7 7
NI Other 1 1

Prediction based on opinion polls from 24 Jul 2019 to 27 Jul 2019, sampling 7,733 people.

Probability of possible outcomes

Conservative majority
31%
Labour majority
16%
Con/Brexit coalition
9%
No overall control
7%
Lab choice of Lib/Nat
7%
Lab/Lib coalition
6%
Lib choice of Con/Lab
6%
Con choice of Lib/Nat
4%
Con/Lib coalition
4%
Brexit/Con coalition
3%
LibDem majority
2%
Lab/Nat coalition
2%
Brexit majority
2%
Con/Nat coalition
1%

The future is never certain. But using our advanced modelling techniques, we can estimate the probability of the various possible outcomes at the next general election. ('Nat' means SNP+PlaidC)

Election predictions

Conservatives close to, but still short of, a majority

Brexit Outcome implied probabilities 7-Aug-2019

See the facts and figures behind Martin Baxter's election prediction at this week's Chopper's Brexit Podcast.

Topics covered include the overall election prediction, and three useful rules of thumb for understanding polls.

The big division between pollsters continues with some seeing a Labour coalition and others a large majority for Boris Johnson.

And the regular check on the chance of the various Brexit outcomes.

See all the facts and figures here.


Posted 8 August 2019

Brecon and Radnorshire by-election result

Welsh by-election in line with GE forecast

The Electoral Calculus forecasts are designed to predict general elections not by-elections. By-elections are particularly subject to special local factors as smaller parties get squeezed. It would not be sensible to rely on GE forecasts to predict by-elections, or to use by-election results to forecast general elections.

But the by-election result can be compared with the GE forecast at least to check if there are any glaring mismatches between the two. This might help detect if there is serious polling or model error. Let's have a look at the Electoral Calculus general election prediction posted on 29 July with the by-election result from 1 August.

Party2017 VotesGE Predicted
Votes
By-election
Result
Difference
LIB29%31%43%+13%
CON49%27%39%+12%
Brexit0%22%10%−11%
LAB18%10%5%−5%
UKIP1.4%0.4%0.8%+0.4%
PC3%5%0%−5%
Green0%4%0%−4%

The predicted GE forecast and the by-election result agree on the ordering of the four main parties: Lib Dems to win, Conservatives second, Brexit party third and Labour in fourth place. They also happen to agree that the majority is around 4pc. Obviously the absolute vote shares of the two largest parties are higher than the GE forecast. This could be because smaller parties were squeezed and many voters voted tactically. The Liberal Democrats "gained" about 13pc, which was broadly equal to the sum of the votes lost by Labour plus the estimated support for Plaid Cymru and the Greens who stood aside in favour of the Lib Dems. On the Leave side, the Brexit party was successfully squeezed by the Conservative party, with about half of Brexit party supporters deciding to switch to the Conservatives.

The good news for the Liberal Democrats is that a pro-Remain alliance can be effective, albeit in a seat they were predicted to win anyway without one. The good news for the Conservatives is that they squeezed the Brexit party and were convincingly ahead of them. This was better for them than the Peterborough by-election in June, where they finished in third place behind Labour and the Brexit party.

It's also fairly good news for the pollsters and our model. The by-election results are not identical to the GE forecast, but the differences make intuitive sense in terms of squeezes and the absence of Plaid and Green candidates. This doesn't look like strong evidence of failure by the polls or model. That doesn't mean that it validates them, but it is a useful "sanity check".

Posted 2 August 2019

July update

'Boris Bounce' puts Conservatives back in the lead

July's update is based on polling conducted after Boris Johnson was elected Conservative leader and appointed Prime Minister. It shows an immediate swing of more than 5pc from the Brexit party to the Conservatives, with the other parties holding fairly steady. This puts the Conservatives back into the lead as the most popular party with the public, according to the pollsters, with an average lead of about 4pc. This is the first time the Conservatives have been ahead of Labour since Britain failed to leave the EU in March, and the (joint) largest lead they have had since the last general election.

The polls also show the Conservatives now have around twice the support of the Brexit party, which reverses the position from two months ago when they were behind the Brexit party in polling support. On these figures, the Brexit party are unlikely to win many seats directly, but they can still stop the Conservatives from winning marginal seats from other parties.

Some press comment has suggested this 'bounce' is smaller than Theresa May gained in July 2016, when she took over from David Cameron. But using a like-for-like comparison (change in party support using an average of four polls just before and just after the new PM takes over), the 'May bounce' was a bit lower, being less than 4pc compared with more than 5pc for Johnson. Other comment has suggested that the immediate gain in support is transient, and will rapidly fade away. This was not the experience of Theresa May whose lead over Labour was maintained and enhanced for nearly a year, and only evaporated during the general election campaign of 2017.

But while this change is good news for the Conservatives, they would not enjoy the results of a putative general election tomorrow. On these figures they would win about as many seats as Theresa May did in 2017. They would be on the cusp between a minority government and one with a small majority, winning around 310 seats. That would not be enough for the government, even with continuing DUP support, to push through a controversial legislative agenda.

PartyPolls (%)
Old PM
Polls (%)
New PM
Change
(%)
Predicted
Seats
Seat Delta
(Seats/%)
CON24.1%29.7%+5.6%31115
LAB25.5%25.4%0.0%22413
LIB17.5%18.2%+0.7%434
Brexit20.0%14.6%−5.5%01
Green6.4%5.1%−1.2%10
SNP3.2%3.2%0.0%50
PlaidC0.6%0.6%0.0%3
UKIP0.6%0.8%+0.2%0
DUP10
SF7

For the Conservatives to get a working majority, they have to increase their support in the polls. A useful rule-of-thumb is given by the 'seat delta' which indicates approximately how many more seats each party would win if its support increased by 1pc. Both the big two parties have a seat delta of around 15 seats for each 1pc. So an increase in Conservative support of 1pc would gain them about 15 seats, which would take them (just) to an overall parliamentary majority. Each additional increase in support of 1pc would add another 15 seats, which would increase their parliamentary majority by 30 seats (since every new seat adds two to the majority).

This means that, for example, a Conservative increase of 3pc more in the polls could give them a comfortable majority of around 60 seats. And in the other direction, a decrease of 3pc would leave them 60 seats short of a majority and out of government.

For the Labour party, the challenge is to unite the left-of-centre voters who are still split across itself, the Liberal Democrats and the Greens. Around 49pc of the public support one of those parties, but Labour has barely more than half of that total. At the moment, the Lib Dems and the Greens are strong enough to wound Labour, by splitting the left-ish vote, but are not strong enough to replace it. Labour needs to win back many of those voters. A recent poll by DeltaPoll suggested Labour could gain 9pc if Jeremy Corbyn was no longer leader of the Labour party, but it didn't ask about particular successors.

In summary, the Conservatives are back, the Brexit party is down, and the two-party system is more relevant again. But a few percentage points of support can change the result considerably, making a wide range of potential outcomes possible. Electoral Calculus will give you the central forecast, as well as the probabilities of various outcomes to quantify the uncertainties. It's going to be an interesting autumn.

Posted 29 July 2019

Polling error

Pollsters divided on whether Labour will win

There is a disagreement between pollsters about how well the Labour party is doing. Among the established and respected major pollsters, there are two camps. One group thinks that Labour is doing relatively well and will get more votes than any other party. But the other thinks Labour is doing less well and is in third or fourth place.

Labour/Lib Dem Coalition
majority of 48

PartyNon-YouGov
polls
Predicted
Seats
CON 24.3%220
LAB 26.5%281
LIB 18.3%57
Brexit 18.3%20
Green 6.0%1
SNP 3.2%50

Prediction based on 4 non-YouGov opinion
polls from 21 Jun 2019 to 16 Jul 2019,
sampling 6,095 people.

Conservative/Brexit coalition
majority of 8

PartyYouGov pollsPredicted
Seats
CON 23.8%213
LAB 19.8%182
LIB 19.5%68
Brexit 21.3%116
Green 9.0%1
SNP 3.2%50

Prediction based on 4 YouGov opinion polls
from 24 Jun 2019 to 17 Jul 2019,
sampling 7,084 people.

The graphic shows two alternative universes separated by the great pollster divide. On the left-hand side, Labour is predicted to be the largest party, and Jeremy Corbyn is likely to be Prime Minister in a minority government. On the right-hand side, the Conservatives are the largest party and Boris Johnson is the likely to stay in office in a coalition with the Brexit party.

Read the full details of the pollster divide, and how polling error translates into seats at "Pollsters divided".

Posted 23 July 2019

Election Predictions

New prediction baseline for new political environment

The European elections heralded large changes in British politics. The Brexit party came from a standing start to win the elections overall. On the Remain side, the Liberal Democrats and the Greens also both did well, though Change UK had a poor performance.

Although the European elections are not a direct guide to future Westminster elections, recent opinion polls have confirmed the main features. The Brexit party, Liberal Democrats and the Greens are all polling better after the elections than before them.

To reflect the current political environment, the Electoral Calculus baseline for predictions has been updated today. This update is based on recent polling and uses modern regression-based techniques to estimate the support in every seat across the country.

This new baseline will be used for all our own predictions and also for user-defined predictions at a national, regional or Scottish level. Live polling and user-defined polling will be taken as a perturbation to the new baseline using the Electoral Calculus Strong Transition model.

This should result in more accurate predictions, particularly for the Brexit and Liberal Democrat parties whose baseline vote share in each seat is now more accurate. One natural, but paradoxical, consequence is that the support figures for the 2017 election will produce a different prediction from the actual result. This happens because the underlying political landscape has changed since then, even for the same national support figures.

The Change UK party has been polling at around 1pc in popular support since the elections and half of its parliamentary strength has left it. Because of this we will no longer show Change UK as a separate party and it will be classed as one of the minor parties which are shown as "OTH" or "MIN". Support for Change UK is no longer an input to the user-defined predictions. UKIP will continue to be shown as a separate party because it stood in about half of all seats at the 2017 general election, and it is psephologically relevant to see which seats those were and how well it did.

Although our best efforts have gone into preparing and checking the new baseline, there may be the occasional anomaly buried in the data. Please let us know if you see anything which appears strange or incorrect. Click 'Email us' at the bottom of the left-hand menu bar.

Posted 20 June 2019

Brexit process

Brexit Process needs more Democracy not less

UK Polling Station sign

Big decisions are more legitimate and better accepted when taken openly and democratically. The Brexit process has been marred by decisions taken in private by small groups, which has hurt their legitimacy and produced poorly received outcomes.

This article reviews the unfortunate history and suggests a democratic way forward.

Posted 10 June 2019

Election Prediction

Brexit party ahead of Conservatives in latest three polls

This month's poll-of-polls shows the Brexit party ahead of the Conservatives. This is because the three most recent polls (YouGov, Opinium and Delta Poll) all have the Brexit party ahead by between 3pc and 9pc. As our graph posted on 28 May shows (see "Euro elections analysis below"), there is a switch-over effect in seats won if the Brexit party overtakes the Conservatives. According to these polls, this has now happened so the Brexit party is predicted to win a substantial number of seats, with the Conservatives reduced to third equal place with the Liberal Democrats. On average, the Brexit party is slightly ahead of the Labour party and is predicted to be the largest party in parliament.

On these figures, a three-way Brexit/Conservative/DUP coalition would only have 313 seats, which is short of a majority. An alternative three-way coalition of Labour/SNP/Lib Dem would have 323 seats, so would need additional support from the five Welsh nationalists to make a wafer-thin majority of six seats.

Some points of caution need to be made. The polls are quite volatile and do not show a consistent picture. They disagree on which party is in the lead: YouGov says the Lib Dems, Opinium says Brexit, and Delta Poll says Labour. And the Labour and Lib Dem vote shares vary quite widely between the pollsters. Small differences in voter support can translate into a large difference in seats. For example, a 1pc swing from the Brexit party to the Conservatives would move around 50 seats. So the uncertainty in the prediction is higher than usual. And these polls were taken just after the European elections. This may be a temporary effect and the old status quo may return before too long. Or this could be the new landscape.

It is a truth that if either Leave voters or Remain voters could unite behind a single party, and the other group did not, then they could easily get a majority at Westminster. The Leavers are still split, but more of them are favouring the Brexit party than the Conservatives. The Remainers are even more split between Labour, the Liberal Democrats and the Greens. If voters move to unite their forces, or the major parties change their policy stance, then there could be dramatic changes.

Posted 3 June 2019

Euro elections aftermath

Euro elections prediction accuracy

Electoral Calculus used a regression-based methodology to predict the European elections. Many of the predictions were broadly correct, though we underestimated the scale of Labour's collapse. Our traditional track record analysis shows what went well and what didn't, and also lays out the evidence to show the regression methods are better at regional prediction than looking at sub-samples of classic polling.

Using a simple measure of variance, regression methods explain 66pc of the "shape" of regional trends, but classic polling only explains 20pc of the shape.

Posted 31 May 2019

Euro elections analysis

Euro elections point to divided parliament

Our analysis of the Euro election results points to a divided parliament if there were a new general election. Both Conservative and Labour parties would be on around 250 seats, well short of a majority. The Liberal Democrats and Brexit party would each get around 15pc-20pc of the vote, without winning many seats.

Seats won by Conservative and Brexit parties against their vote difference

Also included are insights about the electoral battleground between the Conservative and Brexit parties, the likelihood of the three Brexit outcomes, and evidence-based assessment of the leading contenders for the Conservative leadership.

Read the analysis here on Electoral Calculus or listen to Martin Baxter explain what it means on Chopper's Brexit Podcast (28 May).

See also Martin Baxter quoted on the likely Westminster impact of the European elections in today's Daily Telegraph.

Posted 28 May 2019

Tactical voting in the Euro elections

Practical advice for tactical voting

Can tactical voting work under the Euro elections' PR system? Some commentators have cast theoretical doubt on it, but the evidence shows that it could win lots of seats.

On both sides of the Brexit debate, voters are split between several parties: UKIP/Brexit on the Leave side; and Lib Dem/Green/Change UK/SNP/Plaid Cymru on Remain.

Electoral Calculus recently provided independent forecasting of the Euro election result for pro-Remain group Remain United who want to encourage pro-Remain tactical voting. That analysis suggested the simple tactical voting strategy of voting for the leading pro-Remain party in each electoral region.

Expected England seats won against Tactical Fraction

Some academics and commentators have questioned this advice and suggested that other approaches might be better. We had a look at the evidence, and it seems to support the original advice. Voters, both Leavers and Remainers, are well-advised to concentrate their votes with a single party in the Euro elections and not to split them.

Read our analysis here.

You can also see technical details of the methodology in this PDF document.

Posted 16 May 2019

Political Analysis

Three-D Politics and the Seven Tribes

A major new piece of analysis by Electoral Calculus and pollster ComRes creates a three-dimensional landscape of British politics and identifies the seven political tribes which occupy it.

Electoral Calculus has moved beyond a simplistic one-dimensional view of politics, and even gone further than our own pioneering two-dimensional view, to create a three key dimensions of political attitudes:

Using major polling work by the British Election Study, we can give a political three-D position for each poll respondent identifying their political position on each of the three dimensions.

Groups of like-minded individuals can be spotted to identify seven political tribes of people with similar political attitudes, and gain insight about their demographics and voting behaviour.

The Seven Tribes have also been reported in today's Daily Telegraph. You can see the full details of the Electoral Calculus analysis here on our website.

You can also find out your own 3D political position, and that of your neighbourhood, by taking our short 3D survey.

Posted 20 April 2019

Regression-based prediction

Regression-based ComRes/ElCalc poll shows increased Labour lead

Working with leading pollster ComRes, Electoral Calculus has applied modern regression-based techniques to show an increased Labour lead over the Conservatives.

Regression techniques, also known as MRP, are growing in popularity in the market research industry as a way to make polling more accurate again. Classic polling techniques have had problems with accuracy in recent years as representative sampling becomes harder.

Working jointly with ComRes, we have applied the regression techniques to recent polls. Those polls showed Labour and the Conservatives about equally popular, when they were analysed with conventional polling techniques. But under the new regression techniques, they show a Labour lead of 4pc over the Conservatives.

General Election 2017Current Voting Intention
PartyVote ShareSeatsVote ShareSeatsChange
CON43.5%31832.6%277−41
LAB41.0%26234.7%29028
LIB7.6%128.6%153
UKIP1.9%05.9%00
Green1.7%13.2%0-1
TIG0.0%08.6%00
SNP3.1%353.7%4712
PC0.5%40.7%3-1

An additional feature of regression techniques is that they give a seat-by-seat breakdown which each seat is individually predicted, rather than relying on uniform national swing assumptions.

This regression prediction by ComRes and Electoral Calculus was featured in the Daily Telegraph, "Tory Leavers most in danger of losing seats" on 20 April 2019.

Details of the regression method, and its successful testing with earlier pre-election polls can be found here.

Posted 20 April 2019