Time to dust down a post from last December looking at scenarios for the 2015 election based on current polling – two of which pointed to the Conservatives being likely to take a poll lead in the next year.
(NB: as then, please note my huge caveat – “the extent of the polling science on display here is me playing around on an Excel spreadsheet.”)
In particular, I was curious what might have happen to my third scenario in the meantime. So pasted below is what I wrote in December, but I’ve updated the graph to add the last four months’ average poll leads to see how they fit the trend-line. (The answer is pretty well.)
Scenario 3
File this under the heading “a bit of fun… probably” – let’s look at the whole parliament and insert a polynomial trendline to take us through to May 2015. Here’s what happens:
Under Scenario 3, then, the Conservatives bounce back from their mid-term slump to lead Labour by 6% come the next general election. It couldn’t happen – could it?
By the way, if you recalculate the trend-line to include the last four months’ average poll leads, it still suggests a 6% Tory lead by May 2015.
Will the monthly average poll lead data points continue to follow this trend-line? It seems unlikely. As I’ve said before, I expect the polls to jump around a fair bit this year, especially around the Euro elections. But poll-watchers look at trends – and currently there’s only one trend apparent in the polls: a narrowing Labour lead.
Apart from my amateur polynomial trend-lines, there’s another criticism that can be flung at this quick and dirty look at the polls – that it’s wrong to focus on poll leads, far better to focus on the party poll shares. True enough.
But history won’t be much consolation to Labour here. The graph below is from this article – 3 early warning signs that Labour’s poll-lead drama is about to become a full blown crisis – by LabourUncut’s editor, Atul Hatwal.
It shows that opposition parties lose an average of 5.7% in support in the final year before a general election. The only time in the last half-century the opposition didn’t lose support in the final year before an election was 1983, as Labour had already reached their nadir in the immediate aftermath of the Falklands war.
Watch out Labour. Opposition parties lose av 5.7% in final year before election notes @atulh > http://t.co/TDV2rvsRYV pic.twitter.com/xV9Kpbpu0q
— Stephen Tall (@stephentall) March 24, 2014
Oh, and yes… for the record, the Lib Dems poll share remains 10%, where it’s been since December 2010. Consistency, that’s the watchword!
* Stephen was Editor (and Co-Editor) of Liberal Democrat Voice from 2007 to 2015, and writes at The Collected Stephen Tall.
Follow @LibDemVoice on Twitter
Like us on Facebook
Subscribe to our feed
Sign-up for our daily email digest
Manuela Perteghella, Stratford-on-Avon and John Profumo
The "Just War" tradition
Our kids need a lorra, lorra, larffs!!
West End Ward Tuesday Surgeries #dundeewestend
Roland
Michael BG
Simon R
David Sheppard
20 Comments
Here’s a real trend that matters – in seven of the last eight general elections the ruling party has ended up with less seats than before. The idea that the Tories could win the next election outright is improbable from a historic perspective, but even more improbable given they are behind in the polls and unpopular.
With regards to poll leads, polling methods have changed significantly over the last few decades. They are trying to predict what would happen in a general election held tomorrow (asking people how they are going to vote in 13 months time would be absurd – nobody’s going to say that they would vote for party A tomorrow but for party B in the future) and use various methodologies that re-allocate votes accordingly – they are, in effect, making a prediction. It isn’t, therefore, sensible to compare today’s polls with those from the 80s and it isn’t sensible to draw lines and extrapolate ‘trends’ as there is no rational reason to assume that there is a direction that will be followed. There is no reason why the Labour lead won’t grow again and there’s no reason why it won’t contract either. At the moment the best information we have is that Labour have a narrow lead, but one that would give them a clear majority of seats.
I’d say your scenario is a waste of time.
First off, an equally valid trend would be a drop from +5 at May 10 to -5 at May 11, followed by a slower drop from -5 at May 11 to about -7 at May 15. In short, you don’t have enough data to identify a trend. Worse, the data you do have is a difference between two much larger numbers, and anyone who knows about statistics knows that such differences are subject to much greater fractional uncertainty that the original large numbers.
Second, it is utterly utterly foolish for any political party to put its faith in statistics like this. Voters don’t actually vote for trends, they vote for policies. Is there something that explains the 5% change between Jan 12 and May 12? If not, then 5% is a measure of your inaccuracy. If so, it’s far more important to identify the reason for the change, and it won’t really need fancy statistics to know what it is.
Stephen as you know I have been predicting a greater lead for the Tories even than your 6% for some time and have used the 1983 election as my model for how a Tory recovery would shape following an economic recovery from a deep and/or long period of slump – rather than from say the less severe situation prior to 1992.
The point about the situation leading up to the ’83 election was that the recovery in the Tory vote preceded the Falklands with the turn beginning in December 1981 … and in fact the Tories polling support tailed off in late 1982 as the Falklands’ effect subsided so that their eventual vote share of 42.5% was mostly down to a perception of economic success.
Labour was always weak through this period. It was the Alliance that was the depository of anti-government sentiment – the voting intention at one point rising to 50% . Although UKIP support is not and will not ever get to those levels, we should expect its highpoint after May 14 to dissolve as many of its supporters return to the Tories. Some of these will be ex-Tories returning and some will be ex-Lab 2010s a number of whom will go Tory on the economy.
For this to happen wages have to rise, so expect the Tories to make that happen no matter the cost.
So it would be interesting to see where your curve would end up if you reallocated 30 or 40% of today’s UKIP intenders to the Ts.
My guess remains that the Ts have a good chance of reaching 42%. We should have been pushing them far more and far more vocally on the economy.
I remember in 1983 we kept all our guns on Labour long after they were already sunk. (I suspect because between the SDP and Labour it was personal). You can always leave the Tories to do that – they don’t need our help and it just stops our vote recovering.
Richard it is always the economy! Not policies. And what was happening to the economy between Jan and May 2012?
“Richard it is always the economy! ”
Really? So John Major was returned with a majority in 1997 after the huge improvement in the economy?
Keynesian economics is based around the idea that the government behaves in a counter-cyclical manner, reigning in spending during the good times to counter the over-exuberance of the population and getting the cheque-book out during the bad times when the herd mentality of the population becomes too cautionary. However, it is the herd that elects the government and signs of a booming economy are likely to be matched by a vote for a government with looser purse strings rather than a government wanting to reign in spending further. I’m not sure that voters attribute the performance of the economy to the government, but they are likely to vote for a government that responds to the economy with pro-cyclical promises. We had a vote for austerity in 2010 when things looked bad, despite the fact that austerity is the wrong response to recession. I doubt there will be much appetite for austerity with an improving economy.
Bill,
I don’t buy it. The tories failed to get a majority in 2010 because the boundaries and voter trends were against them. The labour voters opting for UKIP are doing so precisely because they are not the Conservative Party. It’s more ingrained and tribal than you seem to be allowing for. The Tories are dead in Scotland and dead in most of the North. Further more the collapse of the Lib Dem vote in those areas will strengthen labour votes. Plus a fair percentage of UKIP’s vote will stay with them. On of which Labour are very well organised and are always on message which they never were in the 80s.
To me its too early to predict anything, but I suspect another hung election or a small labour majority coz they have less work to do .
@Steve I’m not a Keynesian – and especially not as you define Keynesian policy. Interest rates have been the counter cyclical policy tool for a long time now among descendents of the Keynesians you describe. If you try to stimulate through fiscal policy the MPC will tighten monetary policy to offset it – all you get is increased debt and no increase in aggregate demand. (In reverse that is why the US cuts to spending in 2013 were effectively offset by FOMC monetary easing.) Now if you were arguing against monetary tightening from here on for two or three years, I’d be with you.
The Major Government lost its economic credibility when having gone into the ERM it had to crash out on Black Wednesday – They never recovered from that, even though the economy did.
As the saying goes “Don’t give the keys back to the people who crashed the car in the first place” – it worked for Blair and I consider it will work for Cameron – especially if/when that +40% feel they’ve got surplus cash to spend.
In the end – it’s my opinion and your is yours. I’m just saying bashing Labour and also bashing UKIP is doing the Tories’ job and worse it is counter-productive.
At the moment our share of the ‘centre – to- left’ vote is about a quarter 9:36. Our future depends on upping that share.
Stephen,
I note your caveat “the extent of the polling science on display here is me playing around on an Excel spreadsheet.”
The outcome does however appear to marry with the current forecast of Stephen Fisher, an Oxford University pepsologist who uses a new polls based methodology. The approach is broadly to predict the next election based on current opinion polls and the track record of polls in previous electoral cycles, allowing for change in opinion in the run up to the election. http://users.ox.ac.uk/~nuff0084/ge15forecast/.
Forecast Election Day Shares with 95% Prediction Intervals
Con : 37.5 plus or minus 8.4 i.e. between 29 and 46
Lab : 31.7 plus or minus 6.3 i.e. between 25 and 38
LD : 14.3 plus or minus 10 i.e. between 4 and 24
Implied point estimate for Others combined: 16.5
Forecast Election Day Seats
Con : 313
Lab : 279
LD : 30
Con largest party, but short of a majority by 13
Bill le Breton has no credibility and tries to say that demand is measured by transaction values, rather than consumption. Of course you can increase transaction values if you print money, but this is not demand. Everyone should bear this in mind.
@ Glenn surely the crash in the Lib Dem vote has already happened – the majority of the effect of that is already in the Labour vote share (36%). The tussle is between Labour trying to keep them and the Lib Dems trying to get them back (especially in our held seats and Tory marginals).
If the cost of living argument holds sway in May 2015, the Labour Kippers will return to the fold, but if it doesn’t opt for nurse rather than someone worse, quite a few of them will vote Tory.
As Kinnock said in a slightly different argument, “Brothers let us save you from your poverty.” No poverty, red Tory???
Eddie, being attacked by you does wonders for my credibility.
We dont have to re-invent the wheel here, there already is a rigorous academic study of Polling trends, a team led by Steve Fisher of Oxford, their central prediction is for a 5% Tory lead at the Election in 2015. Of course the Past isnt a perfect guide to the future but its the only one we have.
Polls arent there for us, they exist for the people who pay for them, Newspapers mostly. What the Papers want is a cheap generator of News, conflict, drama, change.They dont want a predictive tool .
PS Theres an article talking about this on the “Labour Uncut” site. The article is headlined “Why the surprise ?…..
“… there already is a rigorous academic study of Polling trends, a team led by Steve Fisher of Oxford, their central prediction is for a 5% Tory lead at the Election in 2015.”
The simple fact is that no one knows what’s going to happen. Fisher’s predictions are actually qualified by huge “plus or minus” figures which render them valueless in practical terms (and I’m not at all convinced they are big enough in some respects, huge though they are).
The only extrapolation that has a certain believability is a straight-line extrapolation of the Lib Dem voting intention figure, because it has already been lying on a dead straight line for the last three and a half years. But in reality the Lib Dem vote is unpredictable, just like everything else,
Paul and other admirers of Fisher’s intoxication with stats (sorry this is long – hope some find it interesting)
I read a great piece recently – an interview with Bill James who came up with Sabermetrics – “the search for objective knowledge about baseball.” Stephen Tall will be especially interested to know that he was an inspiration to Nate Silver.
James used stats and knowledge of baseball to work out where the market hadn’t seen true value and his statistically derived player evaluations were used by the Red Sox to buy under valued players.
In the article he was asked where he thought there were other fields where knowledge was not particularly well structured (as baseball evaluation had been before him)? Politics was his answer. “People who are perceived as learned experts go on television and say stupid sh*t, and nobody says, “Boy, that’s really stupid.””
And that is what I think of an Oxford Dons attempt at prediction. My concern with Professor Fisher’s method it that it “allows for three main historical tendencies” these include the tendency that “parties move back towards their long-run average level of support and/or the level of support at the previous election”.
That might be ok in most elections but because the economy is so central to election performance and deep recessions are so rare (small recessions come and go on the cycle, but only get turned into deep recessions by monetary/central bank incompetence) elections following recoveries from significant elections must a regarded as a sub-set.
The Fixed Parliament Act and the Bank of England 1998 Act also disturb the vision in the rear view mirror: Prior to these there are elections in which the economics are benign throughout the term – Governments front load the pain and then there’s an inevitable recovery in their support. And in the Great Moderation, it was all plain sailing and hardly any dip.
Then there are those cycle-based recessions which Governments, pre-97, and having control of monetary policy could bring forward or delay depending on their preferred election timetable –which until 2010 were within the gift of the Prime Minister.
The averages relying on the past performances of elections in all economic situations and therefore the predictions Fisher makes to my mind will be unlikely to shed light on 2015 and may mislead.
The coalition has been very successful in blaming Labour for the Great Recession, so its narrative is that it has shown a very special competence in bringing about recovery, the job is not yet finished and ‘now’ is not the time to give the keys to the people who crashed the car.
This is therefore a crowded open outcry market place in which for the Lib Dems to be heard. So, we won’t get much of our vote from that direction. And I think that almost all of the UKIP vote will transfer to this narrative in 2015. At present this totals 12% + 34% = 46%.
Labour’s response is ‘Cost of Living’ and ‘they only look after their friends’. The higher consumer confidence rises, the higher wages rise, the higher employment rises, the higher spare cash in the pocket rises, the weaker both those narratives argument. And they will. The Labour vote (presently 37%) will decline by 2015.
Some of that decline will go to the Tories. Some could be induced to keep good Lib Dem MPs in place. As indeed some of the 46% centre and right vote (see above) may remain loyal to UKIP. But a six point lead is a major underestimate.
The Fisher prediction in a way endorses the present Lib Dem strategy, which is why some may find comfort in it. My prediction requires a reconsideration of that strategy.
Bill point taken,
But I still think it’s simply too big a hill for the Tories to climb and that you are overestimating how many Labour voters will switch from UKIP to the Tories and how far UKIP’s vote will collapse. And the thing about living standards is that no one feels better off even when they are. It’s a very subjective..
This isn’t “a bit of fun,” it’s terribly misleading. Actual voting behaviour doesn’t follow neat little curves like that; in fact, making allowances for statistical noise, the polls have been largely static with occasional instances of movement. Labour overtook the Conservatives in early 2011; the Tories closed the gap briefly in late 2011/early 2012, but then the gap widened (to about 10%) in the rest of 2012; in the first half of 2013, both Labour and the Conservatives lost support, this correlating to the rise of UKIP in the polls; in mid-2013, the Tories regained about 4-5% to reach the low 30s, and Labour has remained steady in the high 30s since then. If there’s a trend, it’s a straight line. No doubt things will look different by May 2015, but it could just as well be an increased Labour lead as a Tory closure; there’s no way to project past polling trends into the future *except* in periods of overall opinion shift, and then only for limited periods (e.g., if the mid-2013 Tory rebound had been projected indefinitely into the future, they would be at 45% or better now); and such periods of opinion shift are much more easily detected in retrospect than at the time they occur. They typically (though not always) start with small shifts that are indistinguishable from statistical churn, and only start to look like a trend after a few months.
Glenn, I do see your point and it’s perspective too. And the past would appear to support it, but I have strong memories of campaigning up to and during the 1983 election which is therefore ‘colouring’ my view.
As David 1 says after you, significant shifts are more easily seen in retrospect. This was very true of the turn that took place from December 81 to March 82, which was then buried by the Falkland’s effect.
That is why I keep my eyes on consumer and producer confidence surveys and especially surveys of whether people think their economic fortunes will worsen slightly or improve slightly over the next year. This is the group which when it begins to feel better trip into changing their tactical voting behaviour . Tactical in the sense of voting on the economic security of themselves and their family. That only has to be 5 or 6 percent of the electorate to get the significant shift … That can happen over 4 months easily.
We shall see, but if I am right the present national strategy is that of a dog barking up the wrong tree.
The latest YouGov poll puts Labour 6% ahead, the previous one had them 4% ahead and the one before that had them at 3% ahead. Clearly an exponentially incresaing trend. I predict they will be 13473% ahead by May 2015.
I don’t think enough has been made of boundary changes. As I blogged back in 2012 (https://www.libdemvoice.org/with-all-constituencies-declared-labour-has-an-overall-majority-29753.html) Labour has a fantastic inbuilt advantage on the current parliamentary boundaries. And then you have the UKIP factor which costs all the main parties, but costs the Conservatives the most.