Predicting the General Election in Scotland: a fragile landslide?
‘Too many moving parts.’ This was the understandable complaint of a political scientist (oh, all right, it was Philip Cowley) about the complexities of writing about the 2015 election. Back in the day, David Butler and Dennis Kavanagh could get a good deal of the story by trotting from one side of Smith Square to the other, but it has all got complicated now. The most unpredictable of these moving parts is Scotland.
In my recent seat by seat analysis for Westbourne Communications, I found it pretty much impossible to speculate sensibly about most Scottish seats from my London vantage point, and just assumed that the SNP would take most of the Lib Dem seats and a fair chunk (10? 15?) from Labour, reflecting a swing back to Labour between now and the election. I have recently spent a few days in Scotland, talking to knowledgeable people from various points on the political spectrum and return, in the words of Reggie Maudling, none the wiser but considerably better informed.
The truth about Scotland in the 2015 election is that nobody knows. We can be reasonably confident that most of the 11 Liberal Democrats will be swept away – only two (Alistair Carmichael and Charles Kennedy) are likely winners. West Aberdeenshire & Kincardine, and Michael Moore’s seat in the Borders, are Con/ Lib Dem/ SNP wild cards. We can also regard the sole Tory MP in Scotland, David Mundell, as a likely winner on a split vote in a strongly No voting area. But these are details – what about the other 54 seats where the winner is likely to be SNP or Labour?
While the SNP won beyond its dreams in 2011, its more usual experience has been to swagger confidently into an election only to underperform when the votes are counted. The last rites were being read for Labour Glasgow in the run-up to the 2012 local elections, but Labour held on surprisingly well.
Estimates of the number of SNP seats I heard ranged from 20-ish to over 40, and I still do not feel able to rule any of those numbers out. The error margins on Scottish electoral predictions are huge because:
- Scottish public opinion on the centre-left can be very volatile, even during campaigns – there was little sign of the impending SNP landslide in 2011 until late in the campaign, for instance. If the election can be seen as a choice of UK government, Labour has a chance of persuading at least some of its lost voters (and alienated Lib Dems) back into the fold, but if it is seen as standing up for Scotland against an unsympathetic bipartisan austerity regime in London, the SNP will walk it.
- The SNP has become a mass party. Its average membership per constituency is approaching 2,000 people, vastly in excess of any other party in Scotland. The SNP is bound to benefit in terms of doorstep activity and donations. But we do not know how much of a difference it will make yet.
- The referendum created a more active, politicised Scotland, and not just on the Yes side of the argument. Turnout seems likely to be higher, and – maybe – the voters more argumentative and political. A campaign developing in the moronic vein that we have seen in the last week will drive voters towards the SNP, possibly in greater numbers than we expect even now.
- We can expect very little tactical voting in Scotland, except from Greens giving tactical support to the SNP. The ‘Unionist’ voters show no sign of lining up against the SNP. But campaigns can also change tactical outlooks.
But there is another, more below-the-bonnet issue in predicting Scotland. The difference between the SNP winning big in seats, and Labour holding on surprisingly well, may be pretty small in terms of national shares of the vote across Scotland. It might be within the opinion polling margin of error at the end of the campaign, and it is certainly within the wider margins of error that one is dealing with in the art of prediction.
Huge swings may make seats that look safely Labour on the 2010 results highly marginal between Labour and the SNP. Lord Ashcroft’s poll found that the SNP had a 9-point lead in West Dunbartonshire, on a monster swing of 25 per cent since 2010. This seat was Labour even in the 2011 Scottish Parliament election. These figures suggest that there has been a realignment in the Yes-voting urban west of Scotland.
But in areas where there are more parties in the mix, or a different history of Labour-SNP competition, or less working class demographics, or a large No majority in the referendum, we can expect a smaller swing to the SNP (if the national swing is 15-20 per cent). However, this smaller swing would deliver those Westminster constituencies to the SNP as well.
This would give the SNP an overwhelming landslide in seats on the basis of a modest lead in the national popular vote and fairly small majorities in most of its constituencies, be they high-swing previous Labour strongholds like Coatbridge or lower-swing long-standing areas of interest for the SNP such as Ochil or Livingston and mixed areas like Edinburgh where the minor parties (Conservative, Lib Dem, Green) are a force.
What makes Scotland so difficult to call is that – assuming this pattern of swing – it would take a relatively small reversion towards Labour for the position to switch round, and for Labour to be winning a swathe of seats with tiny majorities instead of the SNP.
What, therefore, is my prediction? It should not be forgotten that 2010 was Scottish Labour’s best set of elections since 2001 and its 22-point lead over the SNP was nearly as big as in Labour’s 1997 high tide (23 points). In every other election since 2005 the SNP has won by a whisker (2007, 2012 locals, 2014 Euro) or a wider margin (2009 Euro, 2011). For those of us who find it difficult to imagine an SNP tide given the 2010 results, the remedy is to look at the detail of what happened in 2011. It is also worth casting an eye back to the SNP high tide in the 1977 local elections, and the vulnerability of otherwise very Labour councils like Clydebank and Cumbernauld that year, and the history of big SNP swings in by-elections (over 19 per cent in Monklands East in 1994, when Labour was nationally extremely popular, for instance).
I think the SNP will certainly pick off some seats Labour won handily in 2010 where the Yes campaign did well and Labour has been historically complacent and has unimpressive (or retiring) incumbents and local parties which have fallen down on basic electoral keep-fit exercises like voter ID. I further think the lack of any substantial Unionist tactical vote will enable the SNP to do well in a fair few multi-party constituencies. But I can see Labour holding on in some more vulnerable-looking seats like Stirling and Aberdeen South and perhaps even Ochil, whose failure to go SNP has been a source of bafflement since the seat was created. The incidence of SNP gains (beyond the obvious takeaways like Dundee West, Gordon and Aberdeen North) will probably spatter up the entire spectrum, missing some that should go on the national swing but hitting some improbable targets.
For what it’s worth, I would now guess the SNP somewhere around 33 seats, rather than the 21 I thought in early January; I’ll return to Scotland soon to do more thinking and listening. But do remember the main point of this article – that a lot of seats in Scotland will probably be decided with small majorities because of the distribution of the swing. This in turn makes the likely largest single party in the House of Commons sensitive to small variations in voting behaviour in Scotland. It would be an irony not lost on either man if Jim Murphy holds the Labour line in Scotland and thereby puts Ed Miliband into Downing Street.