A quiet revolution underway in the Borders?


Russell Bruce examines what’s going on in the Borderlands

I wrote some weeks ago about the difficulty the SNP would have in changing the political landscape in the three constituencies that lie on the border with England. This was an area that voted No in the referendum, but the story has moved on and the narrative is different.

This is an election in which the issue is about creating a fairer society. The SNP proposition  is about influence to promote Scottish interests and co-operate with centre left parties in other parts of the UK to deliver social justice with policies designed to benefit the many not the few. The polls show this is clearly gaining traction in the Borderlands, with the SNP as the main contender in all three constituencies.

Are the Tories also gaining traction with their attempt to pull the No vote behind Conservative candidates? There is no real evidence this strategy is helping them gain ground. Conservatives, aided and abetted by the Lib Dems and Labour, are the ones intent on re-running the referendum campaign instead of engaging with the issues and challenges facing both Scotland and the UK.

The UK Quarter 1 economic growth figures have just been published. The Coalition was expecting growth to be 0.6% for the first quarter. A result that would have boosted the Tories in their claim the economy was growing. The market thought 0.5% more likely, but the figures just released show growth at only 0.3% – half of what the UK government had been expecting.

Another round of austerity, as promised by all the UK parties, will make things worse and potentially see growth wiped out later this year. The Financial Times is reporting that this is the worst quarterly growth rate since the last quarter of 2012. Far from the economy continuing to grow – recovery in the first half of 2014 was 0.9% and it has been slipping back ever since. Growth dropped to 0.6% in the last two quarters of 2014 and is now only a third of the best quarters the Coalition Government managed to achieve.

Cripes, will I be able to pose outside Number 11 ever again?
Cripes, will I be able to pose outside Number 11 ever again?

George Osborne and Danny Alexander have just had the feet cawd from under them. Economic competence was central to their claim to be steering the UK forward with strong economic growth.

This is important in the Border constituencies because the Conservatives are in a stronger position than in any other part of Scotland and only the SNP can beat them, as the analysis below from a variety of polling and analyst sites shows.

In the last parliament – Dumfries and Galloway was held by Labour, Dumfriesshire Clydesdale and Tweeddale by the Conservatives and Berwickshire Roxburgh and Selkirk by the Liberal Democrats. Today all three are beginning to resemble three-way marginals with the SNP in challenger position across the Borderlands.

The SNP is coming from well behind in these constituencies compared with the 2010 result when the SNP were third in Dumfries and Galloway and fourth in the other two Borderland constituencies. But, as in other parts of Scotland, the 2010 result does not seem to be influencing voting intention or predictions of the winners in 2015.

A Conservative government with the Lib Dems in tow changed everything. We saw the effect of this, combined with a widespread belief that the SNP had been an effective and popular government in their first term, in the SNP landslide of 2011.

The Labour vote collapsed and spilt voting for Westminster and Holyrood is no longer looking a long-term feature of voting intentions in Scotland. Labour campaigning with the Tories in the Referendum was just too much for many. The disintegration of Scottish Labour as an electable force has continued to gather pace.

Ashcroft: polling all over the place
Ashcroft: polling all over the place

All three Borderlands constituencies have been polled by Ashcroft polls. In the case of Dumfriesshire, Clydesdale and Tweeddale, Ashcroft has polled it twice and found a contest that was a tie in the first poll is now showing a 2% lead for the SNP. – a trend reflected in recent national polls that show the SNP consolidating their lead all over Scotland.

“Adjustment” in the tables is my calculation and prediction for which I take full responsibility for the methodology adopted. Figures in bold are the ‘winner’ in each column and show the trend for the best placed candidates.

I have used the least favourable figures for the SNP, Election Forecasting 2, and adjusted them by calculating the average of the Electoral Calculus and Election Forecasting 2 figures for Labour and the Lib Dems and added the difference to the SNP vote. All other Election Forecasting 2 figures are unchanged. The reason for this change is that the data suggests Electoral Calculus is over estimating the fall in the Labour and Lib Dem vote share and Election Forecasting are overestimating their vote share. Further notes on the tables are at the end to this article.

Dumfries and Galloway

Borderlands vote predictions.numbers

Ashcroft polled this seat in between 19th and 26th February and found the SNP ahead of Labour who had held the seat since the boundary changes of 2005. The former seat of Galloway and Upper Nithsdale was long a conservative stronghold although won by the SNP in 1997. In 2001 it became the only Tory seat in Scotland. The addition of Dumfriesshire, held by Russell Brown for Labour, changed the make up of the new constituency and Brown in 2005 ousted the then last Tory in Scotland.

Ashcroft’s full constituency poll found a 4% SNP lead over the Tories with Labour in third place. Poll trends since February show the SNP vote consolidating.

Today the seat is a three way marginal with the SNP reviving strongly in the area on all predictions. Electoral Calculus predicts an SNP win with 59% probability, giving them a clear 5% lead over Labour and Conservatives. By a 0.3% margin, Electoral Calculus place Labour in 3rd place. As polling information is fed in constantly Labour have fallen back further to 28.7% since I compiled these tables on Monday.

Election Forecasting, on the other hand, who were predicting a clear 6% lead for the SNP last Friday had changed that to a narrow Conservative (1%) win by late Saturday afternoon. Electoral Calculus may be more accurate than Election Forecasting in this constituency. The ‘Adjustment’ calculation is more cautious in predicting the SNP narrowly ahead.

Late Update Election Forecasting now have the SNP in the lead on 32% ahead of the Tories on 31%

Dumfriesshire, Clydesdale and Tweeddale

Borderlands vote predictions.numbers

Conservative, David Mundell, won the seat in 2010 to become the only Conservation MP in Scotland. In the first Ashcroft poll (10th -19th February) he found the Tories were tied with the SNP on 34%. In the recent, second poll (9th to 15th April) Ashcroft found the SNP had moved ahead by 2% to 36% with Mundell stuck on 34%.

Electoral Calculus gives the SNP a more significant lead in their current analysis, suggesting the SNP lead by 6.3%. Mundell has also slightly improved his share, as all the following columns confirm, but is still behind his share of the vote in 2010. Electoral Calculus give the SNP’s Emma Harper a 68% probability of winning and Mundell 34%.

Election Forecasting predict the Labour and Lib Dem vote is holding up more with less movement to the SNP. My ‘Adjustment’ calculation is again cautious, but gives the SNP a 2% lead, agreeing with the differential Ashcroft found this month.

Late Update Election Forecasting now have Tories unchanged, Labour down 1 and SNP up 1

Berwickshire, Roxburgh and Selkirk

Borderlands vote predictions.numbersThis is where it gets a bit personal as I live in this constituency. I am sticking rigidly to the methodology for the ‘Adjustment’ though. This constituency is held by the Lib Dems and includes part of the old Roxburgh, Selkirk and Peebles constituency, won 50 years ago by David Steel in the 1965 by-election.

The table shows this is not helping Michael Moore much as he is persistently in third position. Still a three way marginal, but like the other two Borderland constituencies it is the Tories and the SNP that are the main contenders.

Ashcroft in his recent constituency poll found the Tories in the lead with SNP candidate, Calum Kerr, just 1% behind. Electoral Calculus predicted a larger drop in the Lib Dem and Labour vote shares giving Calum Kerr a 3.3% lead.

A few weeks ago Electoral Calculus was predicting an SNP win with 57% probability. As I write Electoral Calculus is giving the SNP 64% probability of winning. The Conservative candidate is on just 28% probability and Michael Moore trailing significantly on 7% probability of holding on to this seat.

At the end of last week Election Forecast also predicted the SNP ahead in a closer contest. By late Saturday afternoon, Electoral Forecasting had suddenly changed with the SNP dropping 5% and the Tory candidate up 5%. Although data is constantly fed into the model there does not seem an obvious reason for this sudden change when the weekend polls were showing the SNP consolidating their lead.

Yes, it's that Jesse Rae
Yes, it’s that Jesse Rae

My adjustment prediction again uses the least favourable E Forecast 2 prediction and adjusts the Labour and Lib Dem vote as I have done for the others.

Berwickshire, Roxburgh and Selkirk is the only one of the three constituencies to have an independent candidate. Election Forecasting are predicting Jesse Rae will get a 2% vote share. He is a well-known local character in St. Boswells, but he is unlikely to improve on the 0.3% share the Scottish Jacobite candidate achieved in 2010.

Now if we were to allocate this 1.7% difference to the SNP vote this constituency would be an SNP Tory tie. But I did say I would keep the methodology consistent throughout.

Late Update Election Forecasting now have Tories down 1 and SNP up 1


These three constituencies are still three way marginals, but the overwhelming evidence is that the front-runners in all three are the SNP and the Conservatives. My cautious ‘Adjustment’ calculations are based on changes to the least favourable prediction in all three. Electoral Calculus may be over estimating the drop in the Labour and Lib Dem vote based on other evidence in the tables. Equally, Election Forecasting may be inflating these same parties vote shares.

My prediction is the SNP are most likely to win in Dumfries and Galloway, are strongly placed to take Mundell’s seat and SNP candidate Calum Kerr is well placed to defeat Tory candidate John Lamont in a very tight fight likely to go to the wire.

The Tories have poured money into all these constituencies so they are very serious challengers to SNP ambitions. The only thing the SNP and the Tories agree on is – these are all two horse races. Whether the Borderlands start to turn blue or yellow next week is as likely to be down to turnout and the weather on the day as the polls. It is that close.

Finally I wish to be clear that I have a high regard for the work of the teams behind both Electoral Calculus and Election Forecasting. They have sophisticated modelling techniques and systems that allow new data to be uploaded quickly to amend their predictions.

I have found Election Forecasting’s Dr Chris Hanretty of the University of East Anglia helpful, but have not been able to identify the reason for some sudden changes last weekend when the polls continued to be positive for the SNP.


*Ashcroft does not show a separate vote for UKIP but lumps them in with any other party or candidate.

E Forecast 1 Friday 24th April.  E Forecast 2 Sunday 26th April

Author’s Adjustment – Assumes Electoral Calculus underestimates Labour and Lib/Dem vote shares and E Forecast 2 is overestimating the vote of these two parties. The average results in a combined reduction that is then added to the E Forecast 2 SNP share as the party most likely to benefit. Predictions for other parties are unaltered in this methodology that still produces a more cautious prediction that most other data columns in the tables.

(Photograph and tables Russell Bruce)