Elections Etc

First Combined Forecast for the 2017 general election

Advertisements

By Stephen Fisher, John Kenny and Rosalind Shorrocks.

The betting markets, polls, statistical forecasting models and volunteer forecasters are all pointing towards a landslide majority of around 132 for the Conservatives in the UK general election on 8th June. Surveys of citizen expectations are more limited and less clear, but by our analysis they suggest that 43% of people in society at large are expecting a Tory majority of over 100.

Historically the idea of combining forecasts from different sources has had a good track record, though it has to be admitted that our attempt to do one for the EU referendum did not work out well. Most recently the pollyvote.com combined forecast of the US presidential election last year was 2 points out on the share of the vote.

This is the first of our Combined Forecasts for next month’s general election in which we aggregate seats and vote share forecasts from a variety of sources including betting markets, polls, statistical forecasting models, volunteer forecast aggregators, and citizen forecasts. We will update weekly and the methodology (as detailed below) will evolve.

All the different sources point to a big Conservative win and there is little sign of doubt that the Conservatives will get at least a majority. Various different complex forecasting models are pointing to majorities for Theresa May of between 110 and 174, with an average of 130. Betting and prediction markets are just a little more bullish with an average majority of 146.  Volunteers contributing forecasts to the Times Red Box sweepstake are only slightly more cautious. Details of the average estimate from each source, and the overall average number of seats expected for each party are in the table below.  In addition to the big Conservative majority they collectively suggest that the Liberal Democrats seat tally will only increase by four and that the SNP will be down seven.

Seats Betting Markets Complex models Simple models Volunteered Average
Con 398 390 394 380 391
Lab 158 178 173 172 170
LD 18 11 9 13
UKIP 0 0 0 0
Green 2 1 1 1
SNP 45 49 51 49
PC 3 3 3
Con majority 146 130 138 110 132

Not only are the headline forecasts from these sources remarkably strong for the Conservatives but the evidence from the polls and the models is consistent. There is also little sign of doubt among punters. The table below shows the implied probabilities of there being a Conservative majority and a Conservative landslide (100+ majority) from the betting markets, forecasting models (where available) and the polls (as judged by the proportions of polls showing the required lead). The citizen forecasts in the table show the proportion of respondents (excluding Don’t Knows) expecting each outcome. They are noticeably much more cautious and both ICM and YouGov surveys have shown as few as 30 per cent expecting a Conservative landslide.

Probabilities Betting markets Statistical Models Polls Citizen forecasts Average
Conservative Majority 0.96 0.92 1.00 0.77 0.91
Conservative landslide 0.83 0.86 0.43 0.71

The sources for vote-share forecasts are more limited but again remarkably consistent across polls, markets and forecasting models. The polling average is an average of seven different polling averages. They are at best nowcasts rather than forecasts. The statistical models give actual forecasts. Even though the polls have underestimated the Conservatives and overestimated in the large majority of general elections, and substantially so in 2015, all but three of the four forecasting models with vote shares are expecting Labour to outperform the polls and the Tories to underperform. The betting markets are expecting the opposite; an even bigger Conservative lead than the polls are showing.

% vote Polling average Betting markets Forecast Models Average
Con 46.8 47.4 45.0 46.4
Lab 28.7 26.9 27.6 27.7
Lib Dem 9.6 12.2 10.9
UKIP 6.4 3.9 6.3 5.5
Green 2.8 2.6 2.7
SNP 4.3 4.1 4.2
PC 0.6 0.6 0.6

Methodology

The basic approach is to combine forecasts by averaging them within each category and then averaging across categories. Since the different sources do not all present equally clear figures that can be averaged on a like for like basis there are various judgement calls we have had to make on how to treat the data.

There are numerous statistical forecasting models this year (and more to come). We have divided them between simple (poll average plus uniform swing seats projection) and complex (anything more elaborate but not necessarily particularly complex). Within these categories we simply average the available estimates of seats and shares.  We should note that not all of the models are the modellers’ favourite. Some are counterpoints to their main models for comparison. We have included these on the basis that they are still talked of and expected to be reasonable estimates. E.g. at Electoral Calculus, Martin Baxter has a local election results based model. We have not excluded any models based on quality, but they do have to be statistical models as opposed to personal guesses.

For the seats forecasts we are using just the mid points of the spread betting for the markets. Note that the markets imply fewer seats forecast for the main parties than there are in Britain. This is because the markets are separate for each party and do not need to be consistent collectively.

Volunteered forecasts so far come from The Times Red Box sweepstake.

Betting markets for shares of the vote are done in bands. We take the mid points of the (up to) three bands with the highest implied probabilities weighted by their implied probabilities from the betting odds, averaging across bookies.

There are seven different polling averages. Some admittedly are quite sophisticated, allowing for house effects, but they are nonetheless estimates of current public opinion and not future votes.

Estimates come from around lunchtime on 12th May 2017. The Speaker’s seat is counted as Conservative.

Sources

There sources we used are listed below in no particular order. Please let us know of any that you think we have missed or misclassified. Some polling averages we know of were not included because they were more than a week old.

Prediction markets:

IG

Predictit

Sporting index  

Oddschecker

Complex Forecasting Models:

ElectionForecast.co.uk (Chris Hanretty):

Electoral Calculus (main and local election forecast):

Forecast UK  

UK-Elect

PME Politics (Patrick English):

Nigel Marriot (Uniform Regional Swing + Tactical Voting Model)

Chris Prosser (GE vote shares from Local elections vote shares)

Simple forecasting models (polling average + uniform swing):

Nigel Marriot

Electionpolling

Principalfish

Polling Averages (less than a week old):

Telegraph

FT

Britain Elects

Electionforecast.co.uk

NewStatesman

Electionpolling

Principalfish

Volunteered forecasts:

The Times Red Box Sweepstake

Citizen forecasts:

ComRes (here)

YouGov (here)

ICM (here and here)

The three authors are equal contributors and our names are in alphabetical order.

 

Advertisements