Sunday, April 24, 2016

Did the rise of the SNP really spook Lib Dem voters in England?

Last July I began a post like this:
A myth is growing up about the Liberal Democrat debacle at the last general election. It holds that we lost almost all of our seats because the Conservatives ruthlessly targeted them and won over former Liberal Democrat voters. 
So they did, but there is little sign that our lost voters went to the Conservatives instead.
My assurance was based on my reading of an article by Seth Thévoz and Lewis Baston on the Social Liberal Forum site.

Here are a couple of the paragraphs I quoted back in July:
The Conservative-facing seats showed a remarkably consistent pattern; the main factor at play was Lib Dem collapse rather than Conservative recovery. In each of the 27 seats lost to the Conservatives, the collapse in Lib Dem votes was sizably larger than any increase in Tory votes, by a factor of anything up to 29.
This means that although the Lib Dem position in many Tory-facing seats is dire following a collapse of the party’s vote, the Conservative position is not necessarily ‘safe’ or stable; the Conservatives have won many of these seats on relatively small popular votes, and there still exists in these constituencies a reasonably large non-Conservative vote which could potentially be mobilised around a clear anti-Conservative candidate with a more appealing pitch than that of the 2015 Lib Dem campaign. 
Nor is the Conservative vote appreciably growing much in such areas. In seats like Lewes, Portsmouth South, St Ives, Sutton and Cheam, and Torbay, the increase in Conservative votes was negligible, and Lib Dem defeat can be laid down entirely to so much of the Lib Dem vote having vanished.
I thought of this article when I read the review of David Laws' new book Coalition that Nick Thornsby has written for Liberal Democrat Voice.

Or, to be more accurate, when I read the comments on that review.

In one of them Nick himself says:
The conclusion he [Laws] comes to is that the coalition was probably worst for the party in terms of 2015 results, but that whatever route we took was always going to result in a fairly significant loss of seats, either in a later election in 2010, or in 2014/5. 
The particularly big factor in that is Scotland, and the SNP’s rise there would almost certainly been as drastic whatever we did, which had the double-edged effect of denying us seats in Scotland and scaring our voters in the south-west into voting Tory.
In reply Glenn says:
The Lib Dem vote was not scared by the SNP or Miliband or The Greens or frankly even UKIP. Many more former Lib Dem voters voted for these parties than for the Conservatives. The vote simply split enough in enough seats to give Cameron an edge. This is a government formed on a small majority, not a landslide victory or masses of popular support.
And, Adrian Sanders - the defeated Liberal Democrat MP in Torbay - agrees:
“our voters in the south-west into voting Tory.” No, no, no, this is not what happened. Firstly there was no great swing to the Tories – 500 votes in my seat while I lost over 7,000. Our voters mostly stayed loyal. It was tactical voters who deserted us for Ukip, Labour and the Greens, not the Tories.
This debate matters, because our analysis of what went wrong at the last election must be central to our attempts at recovery.

Are we trying to soothe people who voted Conservative last time and praying for something to change in Scotland? Or are we trying to reassemble the coalition of anti-Conservatives that returned us in these seats between 1997 and 2015?

My feeling, backed by the original article by Thévoz and Baston, is that we should adopt the latter approach,


Andrew Hickey said...

I think it's very obvious that far bigger than the SNP thing was tactical unwind -- all the Labour voters who'd voted for us in order to "keep the Tories out" voted Labour in seats where it really was a "two-horse race". Labour still came third, but we came second in those seats where we'd previously won.

I think the "SNP scare" thing is a very convenient way for the people in charge to avoid all blame for a catastrophe that could be seen coming from four years before, and which they did little or nothing to avert.

Was said...

I did some analysis in a former Lib Dem seat comparing the General Election and District Council voting in 2015. Owing to a non-uniform set of candidates across the district, I had to apply some corrective variables to the data and credit the other parties votes in wards they didn't stand with reference to the vote share they got in the ones where they did. Not perfect, but I hope indicative.

On the face of it 3,000 more people had voted Tory in the GE than they had for them in the DC. 3,000 didn't vote Lib Dem in the GE but did in the DC. The UKIP vote was 4,000 more than in the locals, a marginal increase for the Greens and Labour flat-lined. They were the same voters and the DC LD total vote cast on the same day in the same polling stations would have narrowly won the seat.

It would seem in this seat that, using the DC votes as guides, at the GE level there was no evidence of a move to Labour, the Eurosceptic Tory vote was leeching to UKIP with the soft-LD vote moving to the Tories. This would seem to back up the SNP scare scenario, but it would be dangerous to extrapolate from a single seat. Other seats would obviously have their own local factors in play.

I've not tried this anywhere else as it takes up a lot of time to extract the data from local authority web sites but if anyone has the time to do a full country wide analysis it might be more indicative of the real situation than using the raw GE data alone.

robbieb said...

Fascinating, Wes. Did you write about this in greater detail elsewhere?