For the first time since our model launched, last week’s forecast saw Labour become very slight favourites (51-49) to win the most seats at next May’s General Election.

A slight improvement for the Conservatives in the polls (they’re up 2 since last week in the UK Polling Report average, with UKIP down 2 and Labour and the Lib Dems unchanged) puts them back in front this week – but only just. The model now gives them a 56% chance of emerging as the largest party, to Labour’s 44%.

The central forecast is still for a Hung Parliament, with the Tories on 304 seats and Labour on 290 (with 326 needed for a majority). The model continues to make a Hung Parliament a roughly even-money bet – giving it a 49% chance this week.

(Note that the vast majority of the polling in the UK Polling Report Average took place before this week’s Cabinet reshuffle.)

Date of forecast: 18 July 2014
Days till the election: 293

Inputted current average poll shares (from UK Polling Report)
Con: 33%
Lab: 36%
LD: 9%
Others (inc. UKIP): 22%
– UKIP: 12%

Forecast Election Day Shares (with 95% Prediction Intervals)
Con: 36.2% (±7.6, i.e. 29% – 44%)
Lab: 32.0% (±5.7, i.e. 26% – 38%)
LD: 12.8% (±8.5, i.e. 4% – 21%)
Implied point estimate shares for:
– Others (inc. UKIP): 19.1%
– UKIP: 10.4%

Forecast Election Day Seats (with approximate 95% Prediction Intervals)
Con: 304 (222 – 398)
Lab: 290 (201 – 366)
LD: 29 (23 – 36)
(Prediction intervals assume LD & others shares at central forecast, Con & Lab shares vary as per prediction intervals)

Central forecast: Con largest party, but short of a majority by 22

Approximate probabilities of key outcomes
Con largest: 56%
… with a majority: 30%
Lab largest: 44%
… with a majority: 20%
Hung Parliament: 49%
… with Con largest: 26%
… with Lab largest: 23%
(probabilities may not sum due to rounding)