Thursday, March 26, 2015

Three post-election permutations that might work in the SNP's favour

There's been a unnecessary degree of confusion in recent months over the differences between coalition, confidence-and-supply, and a vote-by-vote arrangement.  Much of the problem is caused by "creative fuzziness" in the language used by the political parties - for example, it helps the Tories in their aim of whipping up haggis-phobia if they pretend that there is no real distinction between a full Labour/SNP coalition and a looser deal.  But I get the impression that a fair chunk of the media has been genuinely lost at sea on this topic all along.

One crucial point that needs to be borne in mind is that it's not necessarily an either/or choice - it's perfectly possible that this election could eventually result in spells of both confidence-and-supply and vote-by-vote.  There's a relatively recent precedent for that.  The parliament that was dissolved in the spring of 1979 after Callaghan lost a confidence vote had been elected way back in October 1974, but that election ultimately produced no fewer than four distinct types of government - firstly a Labour government with a tiny majority, then a relatively secure Labour minority government that was able to operate on a vote-by-vote basis without its existence being seriously threatened, then a formal confidence-and-supply deal with the Liberals, and finally a return to vote-by-vote, but by now with the administration under constant threat of being brought down.  (It's too often forgotten that the Liberals eventually joined the Conservatives, the SNP and the Ulster Unionists in defeating Callaghan by a single vote - which according to Murphy-logic presumably means that Willie Rennie must be held personally responsible for Thatcherism.)

What we might see this time is Labour refusing to do a formal deal to begin with, but then quickly finding themselves worn down by the guerilla tactics that Alex Salmond and others have been hinting at.  You could imagine that such a weak government might end up trailing badly in the opinion polls, so a snap election wouldn't be a realistic escape strategy, even if the problem of the Fixed Term Parliaments Act could be circumvented.  A confidence-and-supply deal with the SNP might start to look like Miliband's least worst option, because at least it would guarantee him the passage of certain key parts of his programme, and give him some kind of track record to take to the electorate in 2020.  In my view it would also suit the SNP, because a comprehensive deal on Home Rule is surely easier to attain via confidence-and-supply than vote-by-vote.  I must admit I've been slightly puzzled by the apparent preference of both Alex Salmond and Nicola Sturgeon for vote-by-vote, although it may be significant that they've always stressed that confidence-and-supply is nevertheless very much "possible".

But what if the arithmetic doesn't fall in the SNP's favour, and the Tories are close enough to a majority that it's the Lib Dems, UKIP and the DUP that effectively hold the balance?  I think there's still a big opportunity there, which doesn't rely on the SNP directly pulling the strings.  If the Lib Dems refuse to do a formal deal this time around, a minority Tory government could end up losing an uncomfortable number of votes.  Cameron might find himself becoming tempted by the idea of instantly transforming his administration into a majority government on English domestic matters, by putting forward a plan for genuine Scottish Home Rule tied to English Votes for English Laws.  It's inconceivable that legislation to that effect wouldn't pass if both the Tories and the SNP voted for it.  And there would be absolutely no bar to the SNP voting for it, because it wouldn't be an issue of confidence in the Tory government.

Another possibility is the one we were constantly talking about a few months ago - the "sweet spot for UKIP" scenario, whereby the arithmetic works out perfectly for Nigel Farage, and even with a relatively small number of seats he has enough leverage to force an EU referendum this year, without any prior renegotiation of Britain's terms of membership.  If the UK voted to leave the EU, a quick independence referendum would be back on the agenda with a bang.  However, that scenario is looking increasingly improbable - partly because UKIP's hopes of even a modest electoral breakthrough are gradually receding, and partly because opposition to EU membership seems to be dropping.  But you just never know what twists and turns are around the corner.

*  *  *

I had a Labour canvasser at the door on Tuesday for the first time ever in a Westminster campaign - which tells you something important in itself.  I must admit I failed to practice what I preach by posing as an undecided voter and engaging him in a pointless ten minute conversation, but in my defence I did manage to hold his stare enigmatically for at least three seconds before telling him I was planning to vote SNP.  He made no effort at all to dissuade me, and he didn't even bother asking for my reasons - presumably every last drop of nervous energy must be kept in reserve for the people who might still be won back.

He was a nice enough guy, although the all-red outfit seemed a bit "angry" to me, somehow.


  1. The selfish part of me [a big part?] wants the SNP to bring down the whole rotten Westminster palace until it is a massive pile of stourie bricks blocking the all the roads to and from Windsor castle and Eton college.

  2. Well, fieldwork for the full Scottish Yougov I completed on the 24th should be done now. I wonder who its for - looked like a newspaper one.

    On top of that we have the mysterious Yougov mega poll with a big Scottish subset. Times had the crossbreaks for Scotland and released some of the results, but tables did not go up on Yougov strangely. Rumours gave 45% SNP / 26% Lab; close to my own calculations based on 2010 Lab-Lib-Con to SNP numbers. As it's a cross-break, it wouldn't have been weighted properly potentially. Using a simple correction between aggregated Yougov UK crossbreaks and Election Forecasts re-weighted poll based on this (from ncpolitics) gives something like this:

    49% SNP
    28% Lab
    18% Con
    5% Lib
    4% Green
    4% UKIP

    Which would be the SNP up three maybe, with Labour still static.

    For the Yougov Scottish - normally it's the Times that does these but they've just done one. We could wildly speculate that their mega poll crossbreak might have tempted them to do another? Any change such as an increasing SNP lead might prompt that. 50% is a real headliner...

    All speculation anyway, but we may find out soon what that full scottish poll got.

    And then we have our mysterious big rise in SNP identifying respondents in UK yougov polls yet massive 50% down-weighting of these. If tomorrow's is like that then something is going on as people have noted it on UKPR so AW would be aware and maybe check in case it was e.g. a calculation error or something.

    If populus looks normal the morn as survation did, the our Yougov UK thing looks even weirder.

  3. I've just completed a Panelbase poll that seemed to be for a newspaper, possibly the Sunday Post.
    A mixed bag of questions about voting intention, best leader, power generation, schooling and immigration.

  4. Excuse me for interrupting with daft questions

    What is a "cross break"?

    Do I understand that people were asked which party they identify with and the answers were then "down weighted" to make them more in line with previous results? If that is correct what is the reasoning, given the large increase in SNP membership since the referendum? Does that not tend to indicate that they have simply changed their minds?

    1. Cross-tabbed, cross-breaks... subsamples, subsets...

      Basically it's looking at poll break downs by various factors such as age, sex... region. All the above mean the same really.

      In UK polls of 1000 we get what maybe 100 Scots said.

      The Scots sample is not weighted fully for the Scottish demographic / is too small. However, lots of them can give some ideas as to what might be happening in the absence of Scotland full polls (1000 scots).

      Political weightings designed for UK polls may not work well for Scotland due to its rapidly changing political climate, two General Elections where people vote different ways and very high anti-Tory tactical voting levels.

      Such methods may only depress SNP by 1% UK-wide so means little for a UK poll. However, it could make for SNP and Labour level in Scotland when they are not.

      Opinium is bad for this. Yougov was ok until something very strange happened with its last two polls.

    2. I see. Thank you Scottish-Skier :)

    3. I see. Thank you Scottish-Skier

  5. You missed your chance there, James, to try to convert your caller to our cause.

    I have played this scenario over and over in my mind wondering what I would say to a Labour-in-Scotland chap at the door. Apart from telling him/her that I no longer believed anything they say I see it as an opportunity to invite them to ditch their toxic allegiance and start canvassing for the SNP who are the only party speaking for the people of Scotland. It's worth a try.

    1. I take that tactic with religious callers on the doorstep. A few years ago I had two JW's, young and old. Young stupidly asks 'why's that then' when I tell them I'm atheist. Well how wide a door do I need? I then spent the next 20min doing my scientific level best to insert mind-worms of doubt into this young man.

      Since then on a number of occasions I have seen and heard them do the neighbours either side and over the road but not us. You can just step over the wall from the other semi's doorstep to ours but they go out along their drive and walk past us to do the neighbour on the other side. So I think I scored a hit.

      So go for it.

      BTW if anyone has any moral qualms about what I did, it was arms equal. They chapped my door seeking to proselytise me but got proselytised right back. Them's the risks.

  6. Maybe it was for the Scottish Sun, that full YG poll you did, Skier. I'm sure they have used them in the past.

    Seen Lord Ashcroft Tweet ;

    Lab 34% Con 34% UKIP 15% Green 6% Libs5%

    Now my found Standard Grade Maths adds that up to 94%. So, the usua 5-6% SNP/Plaid?

    P.S Does anyone have any links to Wales? I've only just recently how poor Plaid are polling. They are 4th in the polls. Shame really, as Leanne Wood seems a very good speaker and politician.

    1. Not quite, because (very unusually) it's a full UK poll and therefore includes Northern Ireland. No regional / national cross-breaks are given. 31 SNP responders are down-weighted to 24.

    2. "and voters were also weighted to match the 2010 UK election result"

      Nuff said.

    3. Although weighted it was 3.5% SNP (3.48) which would be 42% roughly in Scotland assuming reasonably similar turnouts, a tad lower assuming a higher turnout in Scotland. Upper limit ~54% without 2010 down-weighting, so seems in line with other polls.

  7. Roger Scully (Welsh equivalent of John Curtice) is worth following on twitter if you are interested in Welsh politics.

    His research unit (with the BBC) and ITV Wales both conduct monthly Welsh polls, conducted by YouGov and ICM. Scully said the other day that his next monthly poll is due on Monday.

    He gave a talk the other day about the situation in Wales, which is available on youtube.

    One curious thing about Plaid is that they performed well in the Ashcroft polls of Welsh seats, even though he has only polled seats where Plaid are not in contention. This improvement in Plaid fortunes isn't supported by the Welsh national polls, which basically show them flat-lining since 2010.

    The general theory about Plaid is that they were doing pretty well up until the early 2000s, but then they had a very lacklustre leader and fell back. Wood has made some impression but they are having to build again from a low base.

  8. Thanks, James.

    I'll go have a look at Scully and that video too.

  9. SNP gain a seat on Fife Council.