General Election Prediction
Current Prediction: Conservative majority 46
| Party | 2017 Votes | 2017 Seats | Pred Votes | Low Seats | Pred Seats | High Seats |
|---|---|---|---|---|---|---|
| CON | 43.5% | 318 | 43.5% | 262 | 348 | 441 |
| LAB | 41.0% | 262 | 32.7% | 136 | 225 | 300 |
| LIB | 7.6% | 12 | 12.5% | 10 | 13 | 42 |
| Brexit | 0.0% | 0 | 2.9% | 0 | 0 | 5 |
| Green | 1.7% | 1 | 2.3% | 0 | 1 | 1 |
| SNP | 3.1% | 35 | 3.7% | 22 | 41 | 47 |
| PlaidC | 0.5% | 4 | 0.6% | 0 | 4 | 5 |
| UKIP | 1.9% | 0 | 0.0% | 0 | 0 | 0 |
| Other | 0.7% | 0 | 1.9% | 0 | 0 | 1 |
| DUP | 10 | 8 | ||||
| SF | 7 | 7 | ||||
| Alliance | 0 | 3 | ||||
| NI Other | 1 | 0 |
Prediction based on opinion polls from 02 Dec 2019 to 07 Dec 2019, sampling 10,827 people.
Probability of possible outcomes
| Conservative majority |
| ||
| Lab minority |
| ||
| No overall control |
| ||
| Labour majority |
| ||
| Con minority |
|
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. Minority government probabilities assume possible alliances between the Conservative and Brexit parties, and between Labour, Liberal Democrats, SNP and Plaid Cymru.
In the headline prediction table, the columns 'Low Seats' and 'High Seats' give the 90 per cent confidence interval for the number of seats won.
Parliamentary arithmetic
How to calculate a majority
This simple calculation causes more correspondence to Electoral Calculus than any other. Here's how we do it.
The majority is defined as the difference between the number of seats that the Government has and the number of seats that all the other parties have. It is not the difference between the number of seats that the Government has and half the number of seats in parliament. For example, if one party gets 340 seats out of 650, then the other parties have 310 and the majority is thirty seats, not fifteen. Not everyone believes that so Electoral Calculus set up a vote on Twitter, which had a clear majority for the correct answer.
Then the question of the Speaker arises. There is one Speaker, Lindsay Hoyle, who was originally elected as a Labour MP. There will also be three deputy speakers, two Conservative and one Labour, who conventionally do not vote either. So, counting Sir Lindsay as Labour for this purpose, the four non-voting MPs cancel each other out and do not affect the majority. As long as you include the speaker as belonging to his/her former party, the majority calculation is unaffected by the Speaker and deputy speakers.
Finally, there is the Sinn Fein party from Northern Ireland. It will have around six to ten MPs and they have historically not attended the House of Commons and not taken part in parliamentary votes. If they continue to do that, then the effective government majority will increase because the number of opposition MPs has decreased. But there is no guarantee that Sinn Fein will never attend in future. If it was in their interests to do so because, say, it could result in a united Ireland, then there is nothing to prevent them from choosing to attend. As such we do not assume that Sinn Fein will necessarily always be absent, and the majority is not adjusted for that.
With the current prediction of 348 seats for the Conservatives (including two deputy speakers), then the opposition MPs (including the Speaker, one deputy speaker and Sinn Fein) would have 302 seats, giving a Conservative majority of 46 seats. On those days where, say, the seven, Sinn Fein MPs do not attend, then the majority would be 53.
Posted 8 December 2019
Prediction Models
Predicting the election with modern statistics
Electoral Calculus is a long-term supporter of using modern mathematical and statistical techniques to get the best election prediction. While no prediction method can be completely accurate, our testing has shown that the new regression-based techniques are better than or no worse than the classical polling analysis performed by most pollsters. We have tested this with pre-election polls from 2017, 2015 and 2010 (details).
Our method works as a three-stage process:
- Take a medium-sized poll of respondents and train an algorithm to learn (or regress) how people's voting intention typically depends on aspects of their demographic and political profile, such as age, sex, class, occupation status, past vote, etc.
- Apply those relationships to the known characteristics of groups within each constituency using (anonymised) census data and past election results, to get a forecast for how people in the seat are likely to vote on average. We call this the baseline prediction.
- Make a small adjustment to every seat to reflect changes in public opinion since the baseline was last calculated. This adjustment is based on uniform national swing (UNS) modelling. That modelling is not recommended for large changes in public opinion, but should be reasonable for small movements.
This allows us to have a regression-based prediction which can change day-by-day as public opinion evolves. Our testing shows that these methods can be very effective with sample sizes of about 5,000 respondents.
Comparison with recent YouGov MRP poll
Respected pollster YouGov conducted a large-scale MRP poll in late November. The abbreviation MRP stands for "multi-level regression and poststratification", which means it is a regression-based poll which uses similar techniques as Electoral Calculus. The poll surveyed around 100,000 respondents, which is a very large compared with typical polls of between one thousand and two thousand respondents. The sample dates were from 20-26 November 2019.
To compare the YouGov and Electoral Calculus predictions we need to adjust for the fact that they were conducted at different times when the parties had different national levels of support. We can do that by applying the third-stage UNS adjustment mentioned above. This can be done either to switch the Electoral Calculus prediction to be on the same basis as YouGov, or vice versa. The results of the two experiments are shown in the two tables.
| Pred Votes | YouGov Seats | El Calc Seats | |
|---|---|---|---|
| CON | 42.7% | 359 | 345 |
| LAB | 31.7% | 212 | 223 |
| LIB | 13.7% | 13 | 18 |
| Brexit | 3.1% | 0 | 0 |
| Green | 3.2% | 1 | 1 |
| SNP | 3.7% | 43 | 41 |
| Plaid | 0.5% | 4 | 4 |
Table A. Both baselines adjusted to YouGov MRP support figures.
| Pred Votes | YouGov Seats | El Calc Seats | |
|---|---|---|---|
| CON | 41.8% | 346 | 336 |
| LAB | 33.3% | 223 | 231 |
| LIB | 13.1% | 11 | 15 |
| Brexit | 4.1% | 0 | 0 |
| Green | 2.7% | 1 | 1 |
| SNP | 3.8% | 47 | 45 |
| Plaid | 0.6% | 4 | 4 |
Table B. Both baselines adjusted to Electoral Calculus poll of polls at 30-Nov-2019.
The first table (A) shows the predictions on the basis of the YouGov vote share where the Conservatives are exactly 11pc ahead of Labour. YouGov calculated the Conservatives would win 359 seats to have a majority of 68. The Electoral Calculus estimate for the Conservative seats is slightly smaller at 345, implying a majority of 40.
The second table (B) shows both baselines converted into the Electoral Calculus poll-of-polls as at 30 November. This has a reduced Conservative lead of 8.5pc. (That poll of polls was based on a sample of 8,341 respondents from 21-28 November.) The YouGov baseline translates into 346 seats for the Conservatives with a majority of 42. The Electoral Calculus baseline gives 336 seats to the Conservatives, with a majority of 22.
Neither table shows a massive difference between the two baselines. Electoral Calculus is slightly lower for the Conservatives and the SNP, and slightly higher for Labour and the Liberal Democrats. Historically, a prediction is quite good if it is within 20 seats of actual result for the major parties, and these two baselines are well within that tolerance.
One baseline might be a bit better than the other, or the truth could be somewhere in between, or they might both be miles off. We will have to wait and see.
The Electoral Calculus prediction will be regularly updated from its own baseline until election day to give the public the most up-to-date information about likely political opinion.
Posted 30 November 2019
General Election Prediction
Conservative majority cut by updated political geography
Using the latest detailed opinion poll analysis, using regression-based techniques, Electoral Calculus has a new baseline political geography. This is used as the basis for all our predictions by estimating how votes for the parties are distributed over the country.
This latest update has cut the predicted Conservative majority from over a hundred to around fifty. We have also improved the handling of seats where parties do not stand candidates, and improved the tactical voting feature in the user predictor.
The major change is the new baseline. Our modelling works by applying regression techniques (sometimes also called MRP or RPP) to detailed polling information to learn how voters with particular demographic and political characteristics tend to vote on average. This information is then applied to the census and political data of each seat to get an estimate of the current voting intention of each seat. The new baseline is based on recent polling and should give a more accurate "shape" of political opinion through the country.
Following the Brexit party's decision not to stand in half the seats, our predictors now assume that Brexit support is only counted by respondents who live in a seat where there will be a Brexit candidate. Using this convention, measured Brexit party support will appear to have halved, being around 4pc-5pc rather than 8pc-10pc. YouGov have announced they will poll on this basis, but other pollsters might not. When using the user predictor, use the smaller number to avoid over-estimating Brexit support.
This issue is a regular cause of confusion at election time. A simple way to understand it is to imagine that Brexit party support is spread evenly across the country. (It isn't, but let's pretend.) Before the party withdrew from half the seats, a national support figure of 10pc meant that they had 10pc support in each seat. But following their withdrawal, a national support figure of 10pc means they have 20pc support in half of the seats and they are not standing in the other half. If their support in the contested seats is actually constant at 10pc, then their national support figure must drop to 5pc. This is the drop that YouGov have observed.
Since many of the polls in the current poll-of-polls use the old method of quoting the Brexit party, this may result in a temporary overstatement of Brexit support and an understatement of Conservative support. These errors should diminish as more polls are published using the new methodology. For example, using the recent YouGov poll in isolation would predict a Conservative majority of around 96 seats. If other polls confirm this, then the predicted Conservative majority may increase in the coming days.
Independent candidates have now been included in Devon East (Claire Wright), Beaconsfield (Dominic Grieve), Broxtowe (Anna Soubry), Guildford (Anne Milton) and Hertfordshire South West (David Gauke). Due to lack of specific constituency polling data, the independent vote share has been manually estimated and may not be a good guide to the election result.
Posted 15 November 2019
Northern Ireland
New NI Poll shows Alliance threaten DUP in Belfast
A new Northern Ireland poll carried out by LucidTalk on behalf of Electoral Calculus and Remain United shows the DUP maintaining its lead, despite its problems over Brexit. The Alliance party, which had received a bounce after the European elections, has fallen back slightly but is still on course to take Belfast South from the DUP, and potentially in Belfast East as well thanks to tactical voting.
Also the prediction has been updated to reflect parties decisions to stand aside in particular seats, and Sylvia Hermon's decision to retire from Down North. Votes from those parties have been transferred to other parties in the proportions given by respondents' tactical voting preferences in the LucidTalk poll.
See full details on our Northern Ireland home page.
Posted 14 November 2019
Uncontested Speaker's seat
The Speaker's seat has often been contested by major parties
Despite some misty-eyed nostalgia about a tradition of the Speaker's seat being uncontested by the major parties, the so-called unbroken tradition dates only back as far as 1997, when the Conservatives did not contest Betty Boothroyd's seat of West Bromwich West. The previous election in which the Speaker ran as a candidate was 1987 when Bernard Weatherill was opposed by both Labour and the then SDP in Croydon North East. Though its has been observed at elections before that.
The nationalist parties do not seem to believe in this particular tradition. Not only did Plaid Cymru run against George Thomas in Cardiff West in 1979, but the SNP ran against Michael Martin in both 2001 and 2005. And the smaller parties have little regard for the convention. Both UKIP (three times) and the Greens (twice) ran against John Bercow in Buckingham.
The tradition that does seem to have been lost is that Speakers often step down just before a general election and let the new parliament choose their successor. This was done by both George Thomas (in 1983) and Bernard Weatherhill (in 1992). That is probably a good tradition as it lets voters in all seats choose between the major parties and avoids voters in any single seat being effectively disenfranchised. John Bercow chose not to follow that tradition, resulting in a likely uncontested election in the marginal seat of Chorley, which deprives voters there of a say. That's not the best example of democracy in action.
Posted 13 November 2019
General Election 2019
Poll of polls at campaign start has 76 seat majority for Conservatives
With the election now almost certain for 12 December, our monthly poll-of-polls marks the start of the general election campaign. According to the polls, the Conservatives under Boris Johnson have a lead of about 10pc over Jeremy Corbyn's Labour party. A straight translation of that hefty lead into seats gives a predicted Conservative majority of 76 seats, which would be a comfortable victory for the Conservatives.
These figures are an explanation of why the Conservatives think that an election is a good idea for them, and the predicted gains for the Liberal Democrats and the SNP have encouraged those parties to support the election. But do the Conservatives really have a lock on this election, or could it unravel again?
The headline prediction table as at 30 October is:
| Party | 2017 Votes | 2017 Seats | Pred Votes | Pred Seats |
|---|---|---|---|---|
| CON | 43.5% | 318 | 35.3% | 363 |
| LAB | 41.0% | 262 | 25.3% | 186 |
| LIB | 7.6% | 12 | 18.1% | 31 |
| Brexit | 0.0% | 0 | 11.3% | 0 |
| Green | 1.7% | 1 | 4.1% | 1 |
| SNP | 3.1% | 35 | 3.2% | 48 |
| PlaidC | 0.5% | 4 | 0.7% | 3 |
| UKIP | 1.9% | 0 | 0.4% | 0 |
| Other | 0.7% | 0 | 1.7% | 0 |
| DUP | 10 | 9 | ||
| SF | 7 | 7 | ||
| Alliance | 0 | 1 | ||
| NI Other | 1 | 1 |
Prediction based on opinion polls from 01 Oct 2019 to 25 Oct 2019, sampling 11,304 people.
There are three main factors which could change this prediction: pollster error, campaign swing and tactical voting. The pollsters have got it wrong before, and they can't all be right now because there are big disagreements between them. Some pollsters see Labour doing much better, with ComRes measuring the Conservative lead at just 4pc which would translate into another hung parliament. Conversely, Opinium measured the lead at a chunky 16pc which would give a Conservative landslide. The truth is probably somewhere between those extremes, but it could be nearer one end than the other.
The power of campaign swing was also demonstrated in 2017. The initial Conservative lead dwindled as Labour ran a strong campaign and Theresa May led a weak one. Labour are hoping that something similar happens again. That's not guaranteed, since Boris Johnson has a better track record of campaigning, but it can't be ruled out.
And finally there is the potential of tactical voting to turn things on its head. Politics has re-aligned and polarised around the Brexit issue, which has overtaken traditional left-right economic questions in voters' priorities. The Conservatives have been fairly successful in unifying Leave supporters by reducing support for the Brexit party. On the other side, Remain supporters are still split between Labour and the Liberal Democrats. Although the parties are unlikely to co-operate, we may see voters vote tactically to support Remain-leaning candidates against Leavers. This factor could change the result in many marginal seats, and there are far more marginal seats than usual this year.
Electoral Calculus will try to keep abreast of all these issues as the campaign develops and bring you the best estimates of their likely effects.
Posted 30 October 2019
September poll update
Pollsters disagree between hung parliament or big Conservative majority
Our regular month-end poll of polls shows an average Conservative lead of seven per cent over Labour. This is just enough for a small majority in the House of Commons, which is the headline prediction above.
But that average disguises wide variation between pollsters. The chart below plots the most recent poll in September from each of ten pollsters, using the horizontal axis for their estimated Conservative support, and the vertical axis for Labour. The range of Conservative support runs from a low of 27pc (ComRes and Survation) to a high of 38pc (Kantar). Labour supports runs from 22pc (YouGov) to 28pc (Deltapoll and Panelbase).
The crucial number is the lead of the Conservatives over Labour (or vice versa if Labour is ahead). That ranges from 0pc (ComRes) to 14pc (Kantar). If this lead is around 6pc, then the Conservatives should get a bare majority in parliament, which is marked by the light blue diagonal line. But if this lead is at least 9pc, then they will get a more comfortable majority, marked by the dark blue line. On the other side, Labour need a lead of about 6pc for a bare majority themselves, shown by the light red line.
The current average prediction is shown by the red dot, marked "Average" which is within the bare Conservative majority region. Fascinatingly, only one pollster (Hanbury) thinks the parties are actually in this zone. Five of the other pollsters are predicting another hung parliament, while four indicate a large Conservative majority.
This increases the uncertainty about the possible election outcome. It could be another a hung parliament, or it could give a comfortable Conservative majority.
Posted 30 September 2019
General Election
Who would win a General Election?
As an election appears possible, Electoral Calculus looks at the question of who would be likely to win it, according to the opinion polls and our models.
Our analysis looks at out current headline prediction, and at the relatively sophisticated statistical modelling behind it. A special section looks at how seats are likely to change at a regional level, showing the pressures that the two major parties are under. It also highlights the ongoing division of opinion between the pollsters and how that could make a big difference to the outcome.
Read the full story only on Electoral Calculus.
Posted 5 September 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:
- Economic: running from left-wing (more taxes and spending) to right-wing (freer markets and tax cuts)
- National: from globalist co-operation to nationalist sovereignty
- Social: from socially liberal to socially conservative
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
