The 2024 General Election saw a landslide victory for the Labour party under Keir Starmer who won 412 seats with a majority of 174, just less than Tony Blair achieved in 1997. The Conservatives under Rishi Sunak crashed to a heavy defeat, winning only 121 seats. The Liberal Democrats did well by winning (at least) 71 seats, and the new Reform UK party won 5 seats. The Green party won 4 seats.
How well did these actual results compare to our predictions? The big picture was broadly correct – we predicted a very large Labour landslide, a big defeat for the Conservatives, and gains for the Lib Dems, Reform and the Greens.
Pred Votes | Pred Seats | Act Votes | Act Seats | |
---|---|---|---|---|
CON | 21.8% | 78 | 24.4% | 121 |
LAB | 38.8% | 453 | 34.7% | 412 |
LIB | 11.0% | 67 | 12.5% | 72 |
Reform | 16.4% | 7 | 14.7% | 5 |
Green | 6.3% | 3 | 6.9% | 4 |
SNP | 3.1% | 19 | 2.6% | 9 |
Plaid | 0.6% | 3 | 0.7% | 4 |
Other | 2.0% | 2 | 3.5% | 5 |
All seats declared.
But in terms of seats for the big two parties, we had Labour about 40 seats too high and the Conservatives about 40 seats too low. The other parties were predicted fairly well.
The forty-seat error was mostly caused by the error in the national vote share, as measured by the opinion polls. Our final poll-of-polls showed a Labour lead of 17pc, but the actual Labour lead was around 10pc. If the actual vote shares are fed into our seat calculator it would give:
CON | LAB | LIB | Reform | Green | SNP | Plaid | Other |
---|---|---|---|---|---|---|---|
123 | 411 | 67 | 8 | 3 | 15 | 3 | 2 |
Which is very close to the actual result. This suggests that the overall 'shape' of seats across the country was well-predicted by our MRP polling, but our accuracy was let down by the error in the national polls.
Our final prediction was also inside the confidence bounds which we gave in advance, reflecting the known uncertainty in national polling.
In Scotland, we correctly predicted that Labour would win over half the seats, and that the SNP would lose a large number of their seats. We predicted the SNP would lose 29 of their 48 seats, and they actually lost 39 of them.
We correctly predicted the Lib Dems would win about 70 seats, and our predictions for Reform, the Greens and Plaid Cymru were fairly good. Of the independents, we didn't spot the pro-Palestine winners in some safe Labour seats, but we correctly predicted Jeremy Corbyn's win is Islington North.
In terms of the 'decapitation' seats where a Cabinet minister (or shadow cabinet minister) lost their seat, we correctly predicted 10 out the 13 such seats (75% success), as well as correctly predicting other high-profile departures such as Liz Truss and Jacob Rees-Mogg. Of the seventeen cabinet ministers who kept their seats, we correctly predicted ten of them.
There will be a more detailed analysis of what we got right and what we got wrong in due course.