Forecasting English Local Election Seat Gains/Losses 2017

by Stephen Fisher

I am planning on forecasting the general election, but first things first. There are local elections this week.

My seat gains/losses forecasts for the English council elections this year are definitely more for curiosity value than to be taken very seriously. They are based on a simple model which uses change in party support in the polls since the equivalent round of elections four years ago to predict seat changes at the national level. The Conservatives are at 46% in the polls; up a massive 15 points on 2013. Labour are down 11 points in the polls since last time; extraordinary for an opposition party. With such big changes in the polls my model inevitably predicts very big net seat tally changes for these parties. But it does not take the electoral geography into account. Many of Labour’s seats are likely to be very safe and the Conservatives might find it hard to recoup many more than the 337 they lost last time.

With that caveat and with more below, my forecasts, together with those from Colin Rallings and Michael Thrasher, are presented in the table below. They also forecast UKIP -105. My modelling approach still cannot manage a sensible UKIP forecast.

 2017 Forecast Range Rallings & Thrasher
Con +430 +190 to +670 +115
Lab -315 -555 to -75 -75
LD -30 -265 to +210 +85

Whereas last year my forecasts were very similar to those of Rallings and Thrasher, the differences are quite dramatic this year.  The Rallings and Thrasher model is based on the results of local by-elections. They apply their predicted national changes in the share of the vote to county electoral division and ward level results (adjusting for boundary changes).

As I have said before, the Rallings and Thrasher model should perform better and usually it does, as detailed in previous posts on this website and below. Probably it will be a similar story this year. But the potential advantage of my model this time is that it has picked up the changes in party support in the wake of the announcement of the general election in June. These benefit the Conservatives especially. Since there have not been many local by-elections for several weeks, the Rallings and Thrasher predictions reflect slightly older political trends. If the local election results reflect the recent rally of support for Mrs May then the Rallings and Thrasher forecast is likely to be an underestimate.

A further issue of interest here is that local by-election results have not tracked changes in the opinion polls since general election for the Conservatives and Liberal Democrats as they usually do and have done for Labour and UKIP since 2015. In particular the Conservative poll share has risen while their by-election performance has been flatlining. Another reason to be sceptical about my forecast.

The situation the reverse for the Liberal Democrats. They are one point down in the polls compared with four years ago but achieved some remarkable successes in local by-elections. The local elections this week will be a crucial test as to whether they can replicate their successful campaigns on a bigger scale. Even if they do that will not necessarily mean that they are heading for better general election success than the opinion polls suggest.

Given wide margin of error shown in the range, my model is essentially ambivalent as to whether the Liberal Democrats will be net winers or losers of seats.

Although it is very clear that the Tories will gain and Labour will lose, I do not take the upper bound for the Conservatives or the lower bound for Labour at all seriously. The former implies that the Conservatives will win 75% of the seats up, while the latter implies that Labour will lose all but a dozen or so of their seats. The main message I take from this is that the opinion polls and previous rounds of county elections suggest that this week’s will be a very bad for Labour and extremely good for the Conservatives.

Perhaps though, the electorate will focus more on local issues this week and wait till the general election to deliver their verdicts on Mr Corbyn and Mrs May.

Performance of the forecasts last year

The table below shows the forecasts from last year and the actual outcome. Rallings and Thrasher did very well for the Conservatives and Liberal Democrats. The fact that the Labour performance was better than expected gave Jeremy Corbyn’s leadership a boost. But in truth it did not bode well for Labour that they were the first opposition party in over thirty years to lose seats in standalone council elections.

2016 Forecast Range Rallings & Thrasher Actual
Con +19 -73 to +112 +50 +48
Lab -151 -276 to -26 -150 -18
LD +93 +54 to +132 +40 +45

Methodology

Since my previous forecasting models showed that years, like this one, with English shire county council elections do not fit so neatly into a broader model covering all years and all kinds of councils. So I have restricted my model this year just to the results of elections that have happened on four-yearly intervals before this year, going back to 1993. I fitted OLS regression models for the raw numbers of gains and losses in England with just one predictor: the change in the poll share. The intercepts have been forced to zero, so that increases in opinion poll ratings inevitably lead to predictions of seat gains. These models were estimated separately for each party and then the coefficients and forecast standard errors were averaged across parties to produce the predictions and 95% prediction intervals (referred to as the range in the tables above).

Acknowledgements: Thanks to David Cowling, Colin Rallings, and Michael Thrasher for help with the data.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s