Amid the glorious blizzard of graphics with which Sky News and the BBC explained last week’s elections, one stands out. It undermines Rishi Sunak’s claim that the Conservatives are on course to deprive Labour of a majority at the coming general election.
Sunak’s claim is very narrowly based. It’s as if, when Crystal Palace recently took the lead against Manchester City inside 10 minutes, he had said the champions were on course to lose 9-0 after the full 90. Dream on, Palace fans; dream on, prime minister.
Here are the BBC figures that give just one of the reasons why Sunak is wrong. I discuss others below.
On these figures, the overall swing to Labour is 7.5 per cent (half of the total of Con down 12 plus Lab up 3). But where Labour came second last time, the swing is 12 per cent. If that difference were to apply in the general election, Labour would win an extra 40 Conservative seats beyond those they would anyway on the national swing.
In the case of the Liberal Democrats, the swing is 7 per cent in all seats, 9.5 per cent where they came second last time. That extra 2.5-point swing would give the Lib Dems around six extra gains.
There is more direct evidence of tactical voting in last week’s results. In the elections for London’s assembly, the capital is divided into 14 groups of boroughs. South West London is dominated by three Lib Dem parliamentary seats. Whereas the battle for London’s mayor was between Tory and Labour, the battle for South West London’s assembly member was Tory versus Lib Dem. This gives us a compelling social experiment, with the same people voting the same day in two different contests. Without tactical voting, we would expect the two results to be similar; with tactical voting they would be different.
Here is what happened. Labour won 38 per cent in South West London in the mayoral election, but just 25 per cent in the assembly election. In contrast, the Lib Dem vote shot up from 13 per cent (mayor) to a winning 33 per cent (assembly). The difference between the two election results reflects tactical voting on an epic scale. Which means that voters really, really hate the Tories and will do whatever it takes to remove them. As if we didn’t know already.
However, Sunak’s hung parliament claim ignores this factor. He was recycling remarks made by Michael Thrasher, the noted local elections expert, on Sky News on Saturday morning. Thrasher made it clear he was not predicting the outcome of the general election. He was simply doing a mathematical calculation based on Labour’s seven-point* lead in the national equivalent vote in the council elections.
This is how that non-prediction works. Labour needs to gain 127 seats to win an overall majority at the general election. A seven-point Labour lead translates into a 9.5 per cent Con-Lab swing since 2019. If that swing were to happen in every constituency, Labour would gain fewer than 100 seats. That would mean a hung parliament, with Labour stuck below 300 seats.
The key thing is that this is a statistical quirk, not a political truth. Thrasher made clear that this proposition ignores other facts that determine the state of the parties. Tactical voting is just one. Here are other things that Sunak is also ignoring.
First, there were results of other elections last Thursday. One was Labour’s triumph in the Blackpool South byelection. Of course, byelections tend to exaggerate the anti-government mood. But they tell us something—in this case that Labour looks set for a clear majority in the general election. Likewise, Labour’s big victories in the East Midlands and York & North Yorkshire mayoral contests. As these were inaugural elections, there was no incumbency effect to upset wider conclusions about support for each party. Between them, they cover 30 parliamentary constituencies; and both Labour victories point to a much bigger national lead than 7 per cent.
Secondly, Labour did best where it most needed to do well. According to Thrasher’s own estimates for Sky News, Labour achieved a 13 per cent swing in areas that voted Leave most strongly in the Brexit referendum, but only 6 per cent in the most Remain-voting areas. Many of the seats that Labour needs to win are in Leave-voting England—most famously in the red wall areas where the Conservatives did best in 2019. Opinion polls tell the same story: Leave voters are swinging back to Labour more strongly than Remain voters, few of whom voted for Boris Johnson anyway.
Third, Reform stood in relatively few seats in the council elections. Where it did stand, Conservative support fell much more than where Reform stayed out. Again, this confirms what opinion polls tell us: Reform candidates at the general election will hurt the Tories far more than anyone else and so help Labour win more seats. Had Reform stood everywhere last week, I estimate that Thrasher would have been projecting a 10-12-point Labour lead and possibly more, not seven.
Fourth, there were no local elections last week in Scotland. Thrasher himself points out that his calculation takes no account of how Labour is gaining ground north of the border. It seems likely to gain at least 20 seats. Does Sunak really think that the SNP remains as popular as was in 2019, which is what his hung parliament view implies? Of course he doesn’t. Perhaps the biggest sign of how badly the Conservatives did last week is that they had to resort to such a weak argument to defend their performance.
To clear the air, let’s put that seven-point lead into context. Yes, if every seat swung the same way, Labour would need a 14-point lead over the Tories to win an outright majority in the House of Commons. But that ignores tactical voting, Scotland and the bigger swing to Labour in Leave-voting areas. As a rule of thumb, Labour probably needs a five-point lead nationally to secure an overall majority. A seven-point lead would translate into a majority of around 20; and a Reform-adjusted lead of 10-12 points would mean a Labour majority of 50-70. Not a landslide; but Labour falling 30 seats short? Forget it.
* On Saturday morning, Thrasher’s original estimate of Labour’s lead was nine points. Later council results reduced this to seven, which has been the basis of subsequent analysis.
