“Could Labour lose 25 seats to the SNP?” That was the question we asked last week, after finding a surge of support for the SNP since the Scottish referendum in YouGov’s Scottish “sub-polls”.
We suggested the SNP were now polling above 40 per cent, with Labour in the mid-20s.
That would mean Labour had lost four in ten of its voters since 2010, with the SNP more than doubling their support.
Today’s Ipsos MORI/STV poll has now endorsed that finding, made on our new elections website, May2015.com. The poll put the SNP on 52 per cent, with Labour on 23 per cent – even worse than we predicted.
How many seats might this mean? STV and Ipsos have suggested the SNP will win 54 – up from the 6 they hold now – and Labour will lose 36, keeping hold of only 4 seats.
That prediction doesn’t have much of a basis. Using traditional “national swing” models doesn’t work well when the swing is 20-30 per cent and a third party (the Liberal Democrats) is also collapsing.
We need Ashcroft’s seat-by-seat polls for a clear idea of how many seats Labour could actually lose. But one academic estimate, constructed by Chris Hanretty, Ben Lauderdale and Nick Vivyan of UEA, suggested that the SNP would win 23 seats before today’s poll. His measure discounts polls this far out from an election and uses demographics to more accurately plot seats.
You can plug in today’s poll and see how many seats the SNP would win here.
It now seems likely that Labour, like the Tories, will fail to win a majority next year. If neither party wins more than 300 seats, partnering with the Liberal Democrats may not be enough for a majority coalition.
Could Labour be forced to work with the SNP – with the promise of a second referendum that would likely involve?