Robert Syms is MP for Poole.
If after the 2010 election I had been told that in 2015 we would have zero inflation, one of the fastest growing economies, and two million new jobs – coupled with rising living standards – I would have said the Conservatives would win.
Over the past five years, apart from a brief period during the so-called “Omnishambles” Budget, the Conservative Party has had a big lead on leadership and economic management, both key ingredients for victorious parties in every general election since the war. In short, the fundamentals heavily favoured a Conservative victory this May.
So were the opinion polls wrong? Yes – they understated the Conservatives and overstated Labour. This point was made six months ago by Rob Hayward, who gave a presentation pointing out that all the electoral contests in 2014 had seen the Conservative Party outperform the polls. These problems were compounded by the fact that election coverage is now much more dominated by hundreds of polls, and many websites have sprung up trying to turn possible votes into seats. If predicting votes is difficult, predicting seats is a far more risky enterprise.
So what went wrong? The first thing to realise is that if one is polling for a two-way or three-way choice, such as a Boris versus Ken contest, it is easier to measure political support. When pollsters are trying to find levels of support of several political parties, then polling becomes less accurate. Even rounding up a percentage for two minor parties can throw off a major party lead.
Secondly, pollsters’ skill is not finding out how people would vote, but guessing how people would vote if they expressed no preference and are don’t knows. In the past, pollsters have tended to reallocate these votes on the basis of what these people have done in the past. In a world of fixed two-party opinions this works, but in a world where people’s political loyalties have become much more fragile and where people operate pick-and-mix politics, it is much more difficult to distribute these votes in the right way, if for no other reason than that people forget how they voted in a previous general election.
Thirdly, turnout: it was quite clear in the campaign that the Conservative vote was rock-solid, but that support for the Labour Party from their supporters was soft. The pollsters overestimated the turnout – and indeed the turnout of Labour voters. The question asked by pollsters about whether people intend to vote is one of those that tends to attract a higher degree of liars.
Fourthly, tactical voting: the constituency polls undertaken by Lord Ashcroft have shown in the case of UKIP that their support falls when voters are asked to focus on their own individual seat. In a multiparty scenario in which the choice of prime minister is essentially David Cameron or Ed Miliband, it is inevitable that some minor party supporters will vote tactically.
For me, I always presumed that if the Conservative Party hit 34 or 35 per cent in the polls prior to polling day, and that we would be able to squeeze at least a couple of percent from UKIP supporters worried about Miliband and possibly the SNP. I think that is what happened, and that pollsters should have had follow-up questions to UKIP and Green supporters to assess their likelihood of switching to their preferred main party choice.
On the question of seats, the 2015 election showed a small swing to Labour. The Conservative lead of 7.1 per cent, which gave us 38 more seats than Labour in 2010, reduced to a 6.5 per cent lead but gave us 99 more seats than Labour. The Conservative annihilation of the Liberal Democrats and the SNP annihilation of the Scottish Labour party has taken out virtually all the disadvantage that the Conservative party faces under current boundaries. The Boundary Commission, whose recommendations will now be implemented this parliament, will be the icing on the cake in further helping the Conservative party. Which pre-election model would have predicted a swing to Labour but a Conservative overall majority? The answer is none.
The outlook for 2020 already looks good for the Conservatives. The Labour Partys need a national swing of 5.3 per cent to create a hung parliament but 8.75 per cent to win a majority of one. In short, to beat us they need to take out seats with Conservative majorities of up to 8,000, the sort of task the Conservative party faced when Blair was in power, which left us in opposition for over 13 years. Something for putative Labour candidates for leadership to think about!
Robert Syms is MP for Poole.
If after the 2010 election I had been told that in 2015 we would have zero inflation, one of the fastest growing economies, and two million new jobs – coupled with rising living standards – I would have said the Conservatives would win.
Over the past five years, apart from a brief period during the so-called “Omnishambles” Budget, the Conservative Party has had a big lead on leadership and economic management, both key ingredients for victorious parties in every general election since the war. In short, the fundamentals heavily favoured a Conservative victory this May.
So were the opinion polls wrong? Yes – they understated the Conservatives and overstated Labour. This point was made six months ago by Rob Hayward, who gave a presentation pointing out that all the electoral contests in 2014 had seen the Conservative Party outperform the polls. These problems were compounded by the fact that election coverage is now much more dominated by hundreds of polls, and many websites have sprung up trying to turn possible votes into seats. If predicting votes is difficult, predicting seats is a far more risky enterprise.
So what went wrong? The first thing to realise is that if one is polling for a two-way or three-way choice, such as a Boris versus Ken contest, it is easier to measure political support. When pollsters are trying to find levels of support of several political parties, then polling becomes less accurate. Even rounding up a percentage for two minor parties can throw off a major party lead.
Secondly, pollsters’ skill is not finding out how people would vote, but guessing how people would vote if they expressed no preference and are don’t knows. In the past, pollsters have tended to reallocate these votes on the basis of what these people have done in the past. In a world of fixed two-party opinions this works, but in a world where people’s political loyalties have become much more fragile and where people operate pick-and-mix politics, it is much more difficult to distribute these votes in the right way, if for no other reason than that people forget how they voted in a previous general election.
Thirdly, turnout: it was quite clear in the campaign that the Conservative vote was rock-solid, but that support for the Labour Party from their supporters was soft. The pollsters overestimated the turnout – and indeed the turnout of Labour voters. The question asked by pollsters about whether people intend to vote is one of those that tends to attract a higher degree of liars.
Fourthly, tactical voting: the constituency polls undertaken by Lord Ashcroft have shown in the case of UKIP that their support falls when voters are asked to focus on their own individual seat. In a multiparty scenario in which the choice of prime minister is essentially David Cameron or Ed Miliband, it is inevitable that some minor party supporters will vote tactically.
For me, I always presumed that if the Conservative Party hit 34 or 35 per cent in the polls prior to polling day, and that we would be able to squeeze at least a couple of percent from UKIP supporters worried about Miliband and possibly the SNP. I think that is what happened, and that pollsters should have had follow-up questions to UKIP and Green supporters to assess their likelihood of switching to their preferred main party choice.
On the question of seats, the 2015 election showed a small swing to Labour. The Conservative lead of 7.1 per cent, which gave us 38 more seats than Labour in 2010, reduced to a 6.5 per cent lead but gave us 99 more seats than Labour. The Conservative annihilation of the Liberal Democrats and the SNP annihilation of the Scottish Labour party has taken out virtually all the disadvantage that the Conservative party faces under current boundaries. The Boundary Commission, whose recommendations will now be implemented this parliament, will be the icing on the cake in further helping the Conservative party. Which pre-election model would have predicted a swing to Labour but a Conservative overall majority? The answer is none.
The outlook for 2020 already looks good for the Conservatives. The Labour Partys need a national swing of 5.3 per cent to create a hung parliament but 8.75 per cent to win a majority of one. In short, to beat us they need to take out seats with Conservative majorities of up to 8,000, the sort of task the Conservative party faced when Blair was in power, which left us in opposition for over 13 years. Something for putative Labour candidates for leadership to think about!