Baxter prediction - Labour majority 134
December 11th, 2004
Will Michael Howard only get four more seats?
Converting this into seats Martin’s formula produces LAB 390 CON 169 LD 57.
So even though Labour would be 5% down on their 2001 vote share they lose just 13 seats on the notional old House of Commons with the Scottish changes factored in.
The current Labour spreads of 345-353 seats offer great value bets for those who believe -
In 2001 the pollsters were largely spared embrrassment by the size of the Labour lead, tactical voting and higher than average swings by incumbent MPs.
Can we be certain that in a changed political climate that it will work in the same way? If it doesn’t then seat changes could be disproportionate.
Mike Smithson
MessageSpace Advertising
I’m still going to argue that “tactical voting” as folks here often perceive it does not exist.
Rather, independent voters in 1997 and 2001 who wanted to defeat the Tories voted for either the LD or Labour based on the notion that both where progressive parties which where the opposite of the conservatives. I do not believe that a great number of hardened Labour Party loyalist or LibDem supporters switched to one or the other in a tactical move in as great numbers as many here like to think, sure some party loyalist may have “voted to get the Tories out” but for the most part it was independent voters who chose to plum for the party most likely to win in their own seat and its been overplayed ever since. For a start there are not sufficient “hardened LibDem supporters” to have had the effect some claim, and many voters who voted LD or SDP prior to 1997 can’t necessarily be seen a firm LD voters unlike the Conservatives or Labour the number of “tribal LDs” are negligible and they would have to be significant if they where to have the kind of effect that some like to accredit them with.
This time around the notion of anti-Labour tactical voting is also puzzling because those alienated by the government can generally be characterised as on the “left” of the political spectrum and so would not vote Conservative, so there would not be anti-Labour tactical voting rather a swing to the LDs from those alienated by the governments polices, and then the traditional erosion of some support for a government from independents who might take into account who would be best placed to win in their seat, but there is no reason to think that the LD and SDP voters or 87 and 92 where loyalist of the kind found for both Tories and Labour and will suddenly come back to the LDs.
That said I’m not expecting a Labour majority of 130+… as Baxter’s predictor suggests, but somewhere between 70 and 100, Mike and many others will disagree but just wait
As has been said before this election will be won and lost by specific seat targetting - Baxter says the LDs will lose seats and the Tories only gain 4? Don\’t think so, the LDs will (net) gain a couple and the Tories will be significantly up. I\’ve seen their on-the-ground campaign and they\’ll do better than the polls suggest.
Ben,
I don\’t think that there is the disagreement over definitions that you think. \”Tactical voting\” is a short hand for a series of trends & shifts in voting patterns.
IMHO things really started to shift after Black Wednesday in 1992, when roughly a quarter of the Thatcher\’s 1980s coalition defected from the Tories. These may or may not have literally been the people who voted Tory in 79, or 83 or 87 - but they were broadly the same sorts of people in the same sorts of seats. They were the legendary C2s who Margeret Thatcher hasd such a hold over & who became so disillusioned with the perceived, incompetence & weakness of the Major years.
Those Tory defectors, comprised about 10% of the population as a whole. Almost on mass that group defected to Labour, and broadly speaking that wandering 10% are still the pivotal vote in deciding the next election.
I think these are your \”independent voters\” in \’97 & \’01 they were anti-Tory, mostly they voted Labour but where the LDs had a strong local presence & campaign and could establish themselves as the main challengers they voted LD. In that sense they voted tacctically to get the Tories out but they weren\’t Labour or LD voters - if anything they were dissafected Tories.
Since \’01 the wandering 10% has begun to drift. I suspect they are now pretty evenly split between all 3 major parties, which explains why Labour are hanging around in the mid/high 30s in the polls.
The situation is however complicated by the other shifts between the parties - Labour bleeding its traditonal GMW & Muslim vote to the LDs, Greens & RESPECT; the Tories losing ground on the right to UKIP.
Despite these interesting distractions off-stage the real battle next year will once again be for the \”wandering 10%\”. For Labour to win with the big majority you\’re predicting they will have to scoop up 3/4 of that vote & minimise the losses amoung their core. They might just do that - especially if te economy looks good & Iraq quitens down but thats quite a big if.
Bullseye, Then i think we pretty much agree
The Tories will do better than there vote percentage suggests because they seem to have woken up to targeting but do not forget that Labour as well as the LDs is use to targeting, my own predication is that the Tories will fall short of 200 seats by a noticeable margin.
While LD targeting is often mentioned it should be remembered that targeting by Labour in 2001 was very important in allowing them to hold on to a number of their marginal seats, particularly in Northamptonshire. The fact that the Tories have begun to target seats (namely the 160 or so they need to win, so their getting there
with this targeting lark) has also been mentioned but as I say they have yet to probably understand and get to grips with this system of campaigning, so I don’t think it will produce spectacular results but improvements in the east midlands and west midlands in particular may be accentuated by it. In the end the targeting by all the main partys may cancel each other out.
Has anybody got a link to the swingometer as published before the 1997 election? How many extra gains did Labour get from the Tories in that election over and above what the swingometer/UNS would have predicted for a 10.7% swing? That might give us a better idea.
It\’s actually 57 LD, not 51.
James, thank you, it has been amended. I\’m away and the article was created on my mobile phone.
Well, if targetting is going to be practised by all the parties, maybe it\’ll all cancel out and UNS will work after all!
Is Ben denying the case for Tory gains in London? Sean Fear has pretty much convinced me…
If everyone targets then those who have the most to defend have the most to lose…
IA, The Tories will in all likelihood win more seats in London, Seats like Finchley and Golders Green, Enfield North, Ilford North, Hornchurch, Bexleyheath and Crayford, Wimbledon and a few others are all winnable for the Conservatives this time round, that said they won’t win all of them I don’t think, targeting seats on that scale would require lots of people on the ground and for the conservatives their lack of young activists is a problem, though not the problem it once was.
I\’m sorry to have to disagree with commentators, above, but the idea that there is a wandering 10% of \’independent\’ electors who will decide the next election is not borne out by the evidence, to put it mildly. For a start, the turnout alone dropped by 12% in 2001. In addition to this section of the electorate which changed its behaviour (and turnout may rise or fall further next time) there is the question of how many voters change their mind between elections.
Do not be fooled by net swing figures. An enormous amount of \’churning\’ takes place. The British Election Study suggests that around half of electors behave differently, or change their mind, between one election and the next. For example, between 1983 and 1987 (two elections with relatively similar outcomes, 53% vored the same way in both elections. This includes those entering and leaving the fray, ie voting in only one of the elections for whatever reason, but even of those who voted in both 1983 and 1987 37% changed their vote between the two contests. And this is in a panel survey where those responding both times are both more likely to be interested in politics and not to admit changing their minds!
Subsequent academic evidence all points the same way. How large is the \’floating vote\’? I would say - at least half the electorate. This volatility has increased dramatically since around 1960 with the decline of mroe rigid voting along class lines. The reasons for this are extremely interesting and rather complex to go into here, but the impact of television as well as fundamental changes in the economic base are certainly involved. In any case it actually offers the opportunity for all parties to beneift from large net swings than those previously customary. The evidence for the volume of committed or strong supporters is now that Labour have just over 20%, the Conservatives around 20%, Liberal Democrats around 5%.
Another thing that users of this forum may care to bear in mind is just how untypical they themselves are of the electorate as a whole.
I know I am - and I\’m sure I spend less time mulling over electoral minutiae than some! The knowledge of the pattern of competition and candidature within constituencies is far more limited among the bulk of voters than may be thought. The impact of targetting, of how photogenic a candidate is, how active an MP is and so on is severely affectded when you consider that less than half of voters are aware of any local campaigning during a general election or when only two (now I guess three!) cabinet ministers can be identified from a photograph by more than half the electorate.
I was once asked by one of the major parties\’ European group of MEPs to ask a national sample of voters about qualified majority voting and about the democratic deficit. I would be interested to have guesses from this panel of exremely well informed and impressive contributors what percentage of average voters had ever heard of either of these vitally important issues…..
Unsurprisingly I also disagree with Ben
As an LD, I have never argued that the number of tribal LDs has gone up but that the number of tribal Labs in particular has fallen and less so for the Cons. Those who were tribal Labs may not vote for anyone else but they sure won\’t vote for the Blairistas - see reduced turnout in 2001 partly caused by this.
Robert,
Again I\’m not sure that there is the difference in opinion that you think. I quite accept your point. I live in an inner London ward that produces fairly stable election results at a local level but have a 25-30% turnover in the electorate between votes, quite apart from churn between the parties.
As I clearly didn\’t explain very well my \”wandering 10%\” are not literally the same voters at each election - but I suspect they do come from broadly the same demographic slice of the electorate.
The core party votes are clearly very much smaller than their actual vote - and as a LibDem I agree that the core LD vote is very much smaller than the others. That\’s why I agree with Ben that the idea of a battalion of Labour or LibDem voters shifting en mass from one party to the other is wrong.
However, I do beleive that the wandering 10% are distinct part of the inter-party churn, whose voting pattern is driven by concerns over economic competence, leadership quality & trust.
Lastly, I think we\’d all agree that anyone posting on this site is pretty much a political anorak!
WRT London, I think the Conservatives will achieve a better than average swing in the suburbs, not necessarily across the capital as a whole. I\’m not expecting any Conservative revival in Haringey, or Islington, or Hackney. OTOH, I would expect quite large swings in Barnet, Enfield, Croydon etc.
I\’d say less than 10% Robert. Your point about the average voter knowing or caring little about the candidates and parties is one I very much agree with. However under our present system the difference between winning and getting everything and losing and getting nothing can be less than 1% of the vote - so having a candidate who is local, looks like a supermodel, may have a few hundred personal votes etc etc can all make that difference.
Big Tall Tim
I’m afraid you’re wrong to ascribe the reduced turnout in 2001 to tribal Labourites not voting for a Blair lead Labour party, low turnout was largely because the election was seen as being a forgone conclusion a notably higher turnout might well have meant a much larger Labour majority (!).
You should also distinguish between the Labour Party base and the Labour Party membership, true the Muslim population has traditionally been part of the Labour base and the halving of Labour’s support amongst Muslim voters will hurt the party in some traditionally safe seats, but the only ideological voters turned off by the Blair government in a notable way are the GMW voters who’s electoral impact is negligible in all but a handful of seats (as I have said in the past). Aside from Muslim voters and GMW the Labour base is still very solid, low turnout is the only threat and at the same time Labour’s support from moderate independent voters who first backed them in 97 and 2001 (and to a much lesser extent in 92) also remains fairly robust. The notion that the Labour base in being eroded is not true; two groups are alienated by the current party leadership, the most tribal elements of the Labour party remain so for the most part.
Robert
I agree a great deal, the majority of voters have what can best be described as a mental caricature of each party and that together with the immediate events around the campaign and past voting habits is largely what makes up their minds over who to vote for. In some seats the nature of the MPs performance however does have an effect however this is only going to be notable in marginal seats such as Enfield Southgate where Twig is a particularly diligent MP or Leicester West where Pat Hewitt is also very active (although hers is hardly a marginal seat
)
Most posters on this board are very well informed and have a clear understand of current events and the electoral geography of the country, however our understanding is only typical of around 10-15% of those who vote, if that. And I am in no way being condescending when I say that.
Ben,
\”Aside from Muslim voters and GMW the Labour base is still very solid.\”
This goes to the heart of the matter - the 10 or so LibDem gains I expect from Labour are largely in Muslim/GMW areas. None-the-less there are some urban areas, where Labour\’s core white working class vote is also under threat. Before the Iraq War the areas where the LDs have made progress against Labour were in white working class Labour places, from Liverpool, to Islington, to Manchester, to Lambeth. One of the interesting side questions of the election wil be whether the LibDems can convert some of their local success in white working class constituencies into national votes. I suspect they will though not enought to win any of those seats - at least not this time round
I dont understand your other comment - \”Labour’s support from moderate independent voters who first backed them in 97 and 2001 (and to a much lesser extent in 92) also remains fairly robust\”. If there is only a small bleed in the core vote and the moderate swing vote remains robust - why is Labour polling 5% lower than its election result in 2001 & 10% lower than it was at this stage in the last Parliament? Surely this loss of support has come from somewhere? As I argued before I think that although Labour has retained a lot of that support, the Tories have begun to claw that support back and for the first time since the 80\’s the LDs are making inroads into theat support aswell
I would argue that for the most part LD gains in white working class areas have been a product of low turnout, in Leicester for example on the outer estates which are predominantly white and working class, voted very solidly for Labour in the by-election however in the local council elections both returned two LD councillors each, the reason for this IMHO is that for the most part those who have not voted are Labour voters while the Liberal Democrats have been able to maximise their support and win over the community activist (people who sit on tenant associations) or those who perpetually have some kind of local issue they want to address, I do not think this foreshadows any threat to this crucial section of Labour’s vote.
For villan at 6 -
Martin Baxter has his historical record on his site. For 97, based on the polls at the time:
Predicted Con 30.6%/191 seats (actual 31.4%/165)
Predicted Lab 47.2%/415 seats (actual 44.4%/419)
Predicted LD 15.4%/20 seats (actual 17.2%/46 seats)
So on a UNS of the actual swing, the Tories should have had about 210-220, Labour about 380-390, LD about 20 (negative swing from 1992 but not enough to lose a seat).
These are coarse estimates - I\’m trying to tighten them up
I\’m glad that my response produced such a high-quality debate of a Sunday morning (though I wonder how many other citizens spent this traditionally \’lazy\’ time in such a manner - I feel that 10-15% is something of an overestimate of how many \’understand\’ things in a similar way to us!)
A few thoughts in way of reply: from all survey evidence the Labour core vote is clearly about half of what they polled in the last two general elections, so it\’s not just \’Guardian readers\’ and Muslims who are vulnerable, though I agree these are two groupings worth monitoring (and if the Sunday Times had not postponed their articles including my recent research this may have been partly published today - perhaps next week!). If the economic feelgood factor were to fade substantially they\’ll have much greater problems.
The reduced turnout in 2001 is partly to be ascribed both to Old Labour thinking the result was foregone and disappointment that they did \”govern as New Labour\” as Blair promised. But it\’s also due to longer-term factors such as the decline in the sens of civic responsibility and duty apparent in all western democracies, to there not being a sense of crisis, and of the ideological convergence of the main parties - these last two factors accounting for the lower drop in N Ireland.
On personal votes, I recommend reading the authoritative analysis of the 2001 results by John Curtice and Michael Steed (both Liberal Democrats by the way!) in David Butler and Dennis Kavanagh\’s book on that election. All incumbent groups, for all parties, did better than new candidates. All parties did on average better in seats they were defending. There is clearly therefore an advantage in the publicity gained by an active MP, applying most to LDs but also to all other parties.
Incidentally, this reflects on the measurement of targetting by parties. The Conservatives did worse in the seats they were targetting (obviously these would be aspirational gains) than in those they didn\’t. Labour held on better in their key seats, one of the reason why uniform swing projections and the notional results on an even split make such grim reading for other partisans. Presumably they did target these, but it is difficult to separate this factor from that of incumbency (especially as Labour in general benefited from \’double incumbency\’, ie in their key seats they had an MP in 2001 whereas the Tories had had that advantage in 1997). The LDs clearly did benefit from special targetting in certain seats such as North Norfolk, as one would expect of a party of more limited national resources. Overall there is little evidence that individual constituency targets make much difference overall in what is predominatly a national campaign.
Finally, Curtice & Steed note that as well as the LD vote declining most in favor of Labour where Lab were defending marginals, indicating the famout LD-Lab tactical vote, it is also true that the LD vote decined almost equally in Conservative-held marginals - where the Conservatives got a higher than average swing! This suggests that if the phenomenon of tactical voting (clearly even higher in 2001 than 1997, by the way) \”unwinds\” - and this remains pure speculation -it could equally damage Labour and Conservative in the key battles between those two.
Fascinating thread… 3 points:
(i) the de-alignment of 2-party politics, evidents since the 1960s, will continue at this election. I think it highly likely the cumulative Lab-Con votes will be <70% of voters. Is there a betting market for this?
This process will continue, regardless of PR being implemented or not, though obviously a pure form such as STV will accelerate it. This will undermine the effectiveness of uniform national swing as a predictor (as we\’re already seeing, hence Mike\’s \’wet finger in the air\’ predictions based on previous opinion polls\’ failure to predict accurately).
(ii) it\’a always dangerous to assume national trends. It\’s easy to characterise the \’97 mood as \’time for a change\’, and \’01 as \’give Lab more time\’. The fact that both elections were the result of low turn-outs tends to get lost. That said, and something I think will hit Lab this time, is that many people would actually like a Lab/LD government - so the prospect of a hung parliament is not something that will frighten voters. Which is why Lab\’s \’vote LD and get a Tory\’ strategy will just not wash.
(iii) finally, Lab will be hit by complacency, which, despite all the warnings, activists such as Ben are compounding. Very few people - beyond this site! - believe Lab won\’t win the next election, so trying to motivate a disenchanted core vote will be that much harder. I think Lab will get a workable majority, say 20-30. The warning signs are there: look at 1970… Can Blair survive that? Doubtful. And when Blair succeeds him, well - I\’ve long believed Brown\’s destiny is to lose Lab the one election it fights under his leadership!
I think Ben\’s gone off message!
Traditionally, Labour support in its \”white working class\” heartland was crucially organised around community opinion formers - shop stewards and Tenants\’ Association leaders were seen as absolutely crucial to getting the vote out in the days of old Labour, and I rather suspect that there are people at the heart of Government - John Precott comes to mind - who would still hold to that. Surely the government\’s local regenration strategies, which set considerable store by such people, is expected to have an electoral pay-off!
On the other hand, I sense a strain in New Labour - most obviously represented by Peter Mandelson, I would suggest - which is quite happy with the political disengagement of great swathes of the \”working class\” (for want of a better term). In this view, policies which would appeal to them, particuarly in the economic arena, are simply a liability in terms of what \”middle England\” wants - and the advantage of their disengagement from civic life is that it prevents the creation of an effective left wing party.
Does anyone have any figures for electoral participation by people under, say, 40, who left school as soon as they could? I\’d be staggered if it was over 20%, or indeed if as many as 75% of them were even on the electoral roll. In Canada, \”Elections Canada\” canvass the entire country before a General Election to get the roll up to date, and the same should be done here.
Points 22 (i) and (ii) appear to contradict each other. If the turnout had been higher then the two-party % would have been higher, as so many of the non-voters would otherwise have voted Labour.
However, I think Stephen T is right about partisan dealignment continuing. Back in the 1950s, for example, the two-party share of over 90% was underpinned both by the highest ever level of voting along class lines, and by a clear public/private sector economic ideological divide, albeit within the consensus of a mixed economy.
With the long and irreversible decline of Britain\’s manufacturing industrial base and the cultural blending enabled and stimulated by both rising levels of affluence and the impact of mass access to television from the late 1950s, the incidence of electoral volatility and the opportunities for both partisan dealignment and realignment started to expand. The major electoral story within the lifetime of the present Parliament has been \’a plague on both your old houses\’ and on now not-so-New Labour\’s as well.
Although I doubt if either Respect or UKIP will win a single seat in 2005, or that the Liberal Democrats will make a major breakthrough, neither \’major\’ party can take anything like 70% of the vote for granted or win with core voters. The prospects of a majority government with the mandate of only around 20% of the eligible electorate looms.
On the Martin Baxter substantive electoral \’prediction\’ in Mike\’s article, I myself would currently guess that Labour looks around 30 high and the Conservatives about 30 low. I accept the arguments about the over-estimation of Labour in polls, though am less convinced by tactical unwind, targetting or other further adjustments.
In fact, quite independently I estimated a Labour overall majority of 72 in the PSA-EPOP prespohologists\’ \’guess the result\’ contest a couple of months ago. I have managed to win this particular event of earth-shattering import on both the last two occasions!
However, all this polling is predicated on \”if there were a General Election tomorrow\”, which there will not be. There is still much to play for in national party strategy, tactics and positioning - and even more out of their control.
Stephen I am not complacent at all, I think Labour will win a third term and hold a majority of 70-100, but this is not automatic and I am certainly not complacent, hard work by activists and party workers will be essential in achieving this just as it was in both 1997 and 2001, what many here like to think is that the membership of the Labour party has effectively collapsed and is now simply not functioning such is not the case and come the general election the Labour party will be just as active in promoting its candidates as in the past.
The themes of crime and the economy seem to be targeted to appeal to both the Labour’s core white working class vote as well as those moderate swing voters who have plumed for the party in the last two elections in overwhelming numbers.
Ben - I wasn\’t being personal, I mean complacency among the electorate based on a belief that no-one but Lab can win. Almost no-one thinks the Conservatives can get a majority; very few that a hung parliament is at all likely (Alan Watkins, Mike et al excepted). That has got to make it harder to motivate voters, core and especially floating voters alike. And that alone will cost Lab votes which in some constituencies will make the difference between re-electing a Lab MP and not. That will only partially be off-set by Con voters being de-motivated by their party\’s seeming inability to look like a potential government.
Robert - I accept the de-alignment is principally the result of the fragmentation of the class system, mainly as the manufacturing base declines. But 2 related questions: (i) how much impact has rising educational standards had (or is this just a function of rising affluence?); (ii) how far was 2-party politics itself an aberration even within universal suffrage, and confined only to the \’50s?
Finally - has anyone ever done any analysis to what extent even the best, most impartial opinion polls are, in effect, \’push polls\’? Ie, they\’re an easy way for those polled to signal their dissatisfaction (eg) with a government, without actually making a decision with any reponsibility… Of course that doens\’t fit in with Mike\’s analysis, with which I agree, that opinion polls conistently over-state Lab\’s lead. But I do wonder whether even reputable opinion polls can fall victim to the public wanting to give politicians a kick without consequences.
No offence taken Steve, I think motivating the core vote in weighting heavily on the leaderships mind and the concentration on crime and the economy will no doubt be aimed at getting these voters to vote.
Stephen, Mike\’s analysis is that opinion polls taken just before an election over-state Labour\’s position. But it is also true that polls mid-term, just like by-elections, local government elections, and other contests have generally seen an anti-government protest or \’push\’ as you put it. 1997-2001 was an exception, but the Oposition were ahead in 1995, say - and also in 1990, 1985, 1981, all before elections the government came back to win.
This was partly behind my point that you cannot predict what will happen in May 2005 from the polls now. The mood could go either way, before you start applying Mike\’s formula (or anybody else\’s as an actual prediction).
Like by-elections etc, polling responses are free of charge. Voting in a General Election is not. And I use the financial metaphor advisedly.
Incidentally, I don\’t see why all this is a problem for opinion polls unless you think that mid-term or pre-election polls are in some way meant to be predictions of the result.
On your two other questions, some might dispute your suggestion that educational standards have risen, although as one working in that field I would not be so ungenerous and certainly the educational value of TV, in the sense that people have come to see so many other possibilities (how many in the Rhondda would have seen the intimate side of a Lord Brocket, to trivialise the point on a not-so-grand scale!?), and travel.
The 1950s was the high-point of class and two-party voting, as people had started ceasing being politically socialised by the pre-class cleavage, predominantly religion (see Kenneth D.Wald\’s remarkable analysis, appropriately entitled \”Crosses on the Ballot\”) around the First World War, and this had taken time to seep into full effect; but the Liberals had been in steep decline from the mid-twenties, and were pretty week from 1931 through the thirties and forties. Admittedly there were lots of Independents elected in 1945, but that was in most peculiar circumstances, and some of it was the result of the continuation of \’Common Wealth\’ success during the war, itself largely made possible by the major party truce.
Robert, just as a matter of interest how did you come up with a Labour majority of 72 given how you feel about the volitility of the electorate and the fact that you say polls at this stage are not a reliable indicator of the election outcome this far from May.
Max: there was a deadline on the entry guesses, and that\’s just what mine was. I tried to take the best available information at the time, from polls, econometric forecasting, local elections, local byelections, historical extrapolations of last-6-month behaviour, eye of newt and toe of frog and I suppose my own view of each individual constituency based on research for the Almanac of British Politics.
Of course this is all quite likely be redundant due to unforeseen \”events, dear boy\” .. but wouldn\’t life be boring if we actually could read the future?
My prediction formula is meant to be a running thing and is designed to give an idea of how things currently stand.
It is also a counterpoint to the UNS numbers from Martin Baxter. I have to admit, as well, that it helps to stimulate great debates on the site from which I for one certainly learn.
The anti-Government poll shares that Robert describes actually only apply to Labour. Taking out the 2000 fuel protest effect Labour whether in Government or opposition sees its poll shares drop compared with the real thing on election day. In my view it is all part of the Labour over-statement which I bore everybody with!
Mike - In European Elections 2004(ignoring UKIP)Greens polled 6%,BNP 5%,Respect 2% and others 5% making \”others\”18%
In Local Elections for all 2003 \”others\” polled 16.4%(see U.K Election Statistics 1918-2004 Research Paper 04/61)yet the pollsters consistantly only arrive at 10% for \”others\”
Is this because some \”others\”(thinking especially of BNP)say labour when asked their voting preference.
Would this explain the overpolling of Labour and the underpolling of \”others\”?
Robert, does the Almanac take the new Scottish constituencies into account. If so does your research throw any light on how the election may progress in Scotland. For instance can you see the kind of \’partisan dealignment\’ so evident with the advancement of the Greens and SSP replicated in a UK general election.
\”Others\” will poll better in the Euros because of PR, and better in the locals because local factors are at stake. That said, I do believe that the minor parties will poll their best vote share for decades in the next general election (even 10% would be very high by historical standards.)
Andy C, thanks for those figures.
Robert W, I\’m amazed at those figures showing so many people changed their vote betwwen 83 and 87. Within the context of the time, that would have included a lot of to-ing and fro-ing betwwen Labour and the SDP. Does it also apply between Labour and the Conservatives on the same scale?
I\’m not sure I understand Mike\’s comment. Surely it has long been accepted wisdom (that doesn\’t make it always true) that governments usually suffer midterm slumps then recover at least somewhat at the subsequent General Election. Labour governments improved from their positions in say 1976/7 and 1968/9 even though as \”Old Labour\” they lost the general …in mid-term they were double figures behind in polling as well as losing seats like Ashfield (or Birmingham Stechford, the predecessor of Hodge Hill - to the Tories), or Dudley on a 20% swing, during the 1974-9 parliament; and Oldham West and Walthamstow East (and Leicester South West - to the Tories!)in 1966-70. In 1985 the local elections were disastrous for the Conservatives. There are many more examples. How can this be a matter of the polls consistently overstating Labour\’s position? Midterm results as well as polls have usually been bad for governments, 1997-2001 excepted. The reasons for this are clear: protest when the choice of government is not at stake, the manipulation of economy and election date by administrations, and so on. I\’m not saying that Labour will strengthen their position in the next 6 months; but it\’s a possibility which might counteract their present unconvincing lead in unconvincing polls. The Conservatives might do surprisingly well, but it would be surprising given their performance midterm in elections never mind polls compared with previous oppositions.
Max - will have to wait for next edition of the Almanac for Scottish seats, but you may be aware of the notional \’results\’ for 2001 on the new boundaries in the Media Guide to the New Scottish Seats by Colin Rallings, Michael Thrasher and David Denver. For prospects, you could apply uniform swings for those - if you dare!
Villan, just seen your post. Transfers between 1983 and 1987, for example did not show much direct switching between Labour and Conservative.
20% of the eligible electorate voted Conservative on both occasions, 13% voted Labour both times, 9% Alliance (which is what I presume you menat by SDP) consistently.
11% abstained both times.
10% either entered or left the electorate.
Of the remaining 37%, only 1% switched between Conservative and Labour. About 6% each switched between Alliance and Labour (in both directions of course) and 6% between Conservative and Alliance. But over 20% switched between voting (for any party) and non-voting, or vice versa. There were small values for Others to add to 100%.
I\’m not sure if you can make any sense of that - but if you want the source, see Anthony Heath, John Curtice, Roger Jowell, Geoff Evans, Julia Field and Sharon Witherspoon, Understanding Political Change (Pergamon 1991).
Thanks for the figures
Robert, it is Mikes contention that whereas Conservative incumbents recover from mid-term poll lows, that wont necessarily happen with this Labour government. Intuitively I disagree with him, but I admit the statistics from 2001 bear out his view.
(Really I should leave it to him to answer because I dont want to misrepresent his argument but it does go to the heart of the various GE predictions and explains why he believes Labour will do so badly.)
Anyway the basic point is that whilst Labour did indeed go up in the polls in the run-up to the 2001 election, their real result in 2001 did not reflect that. In fact their real score was similar to their WORST ICM poll figure in the previous twelve months (leaving out those months distorted by the fuel protest).
Worst poll 2000 = (May) 41% (actual result 42%)
Which was broadly the same pattern as before the 97 election (without any incumbency considerations)
Worst ICM poll 96 = (four times) 45% (actual result 44%)
In other word\’s Labours climb to 49% in the March 2001 ICM poll was illusory. The best predictor of their General Election performance was their earlier low point.
As a purist you will say that one should never use polls as predictors but this is a betting site after all and one has to take what pointers are available beforehand.
Anyway, if he is right and Labour\’s current recovery in the polls is not \”real\” then Labour will get 34/35% next May* and Mike can buy up Bermuda.
*worst Labour ICM poll 2004 = (June) 34%
Villan, thanks for summing up my position so lucidly. I\’m away from my computer until this evening and don\’t have access to my normal research materials. I had been trying to establish whether what has happened to Labour poll shares in 2000/01 was reflected during the 1974-79 Labour Government. That would give further substance to my contention that Govenment recoveries in election run-ups are Tory specific.
Mike, Labours poll ratings in prior to the 1979 election did indeed improve, Thatcher at one time held huge leads… but at the same time i think that the political landscape was so different at that time that simply because that government was a Labour government and this government is a Labour government does not mean there will be any correlation.
One reason that the pollsters may have seemed off in 2001 is that thanks to it being such a foregone conclusion for many voters who would have voted Labour and in the past had voted regularly just didn\’t vote. Had turnout been higher then it is very likely that Labour would have been returned with an increased majority and a much larger share of the vote as the polls had forecast. What the polls failed to do IMHO was to see that many people who would normally vote and responded to polls as such just did not bother to go out and vote because the contest seemed to be a foregone conclusion, the conservative party was only saved a far worse result than 1997 thanks to the this complacency of many ordinary voters. So in conclusion I don’t think the polls discrepancy with the actual result where a measure of failed methodology but instead a measure of just how few people thought their vote would be “needed”… but hey that’s just my opinion.
OK, I think this debate is a simple one:
Proposition one: governments typically experience a run-up in support ahead of the election. Evidence for: 79, 83, 87, and 92. And against: 97 and 02.
Proposition two: it is more that the Labour vote goes *down* into the election, than that the government recieves a boost. Evidence: 97 and 02. Not really relevent, every previous election.
Personally, I find myself agreeing more intuitively with Proposition one. Opinion polls are - as Robert Waller pointed out - \”cheap\”. You will register your own personal (free) protest vote if you have something to complain about. Most people vote with their wallets; if they believe that the government will make them richer (or at least, less poor) then they will tend to revert around polling day. The Conservatives have still not regained the mantle of economic credibility they lost under Major. I cannot see this changing under Howard/Letwin. An inability to decide what tax rates should be or do, demonstrates this.
Whether New Labour would still have the economic halo if Brown were in charge is another matter altogether.
My forecast therefore is for a small Labour recovery into the election.
Regards,
Robret
ps - the server seems deathly slow at the moment. I don\’t know why this is but am working to solve the problem (which may involve moving the whole site onto its own machine).
There will also be another major over-haul to the politicalbetting site in the coming week. Hopefully disruption will be small/non-existant.
Ben,
I\’m not sure I agree with that at all - the drop in turnout was not caused by the Labour core vote staying at home. The turnout was down across the board in Tory & LibDem seats as well as Labour ones. Richmond Park in London for instance a classic LD/Tory marginal, the Labour stayed remarkably stable falling just 1% to 11.3% but the turnout in this overwhelming middle class seat dropped by 11.4%.
On the issue of overstating Labour. IMHO Mike is right the polls are likely to be overstating Labour, however the same thing happened to Thatcher\’s Govts in the 1980\’s. I think the problem is that some people want to be seen as supporting the party they think is most likely to win the nesxt election. when a party has a massive lead then theat lead is exagerated as people try to associate with it when ask how they will vote by pollsters even though in reality they will not vote for them. I suspect that the closer the parties are the less shy people get & this phenomenum declines - so while I thik Mike is essentially right about the polls tending to overstate Labour\’s lead - the further under 40% they show Labour\’s total the less the overstatement.
I agree with Ben that the non-voting of a substantial number of naturally Labour voters saved the Conservatives from a much bigger disaster in June 2001.
In the 2001 election I was closely involved in the Conservative campaign in Beckenham, which having been a very safe Conservative seat had seen the Tory majority fall substantially in May 1997, and to just 1,200 over Labour in the November 1997 by-election caused by the resignation of Piers Merchant. In 2001, the national polls suggested an even bigger win for Labour than in 1997, therefore Conservative Central Office was very worried about losing Beckenham to Labour.
What makes Beckenham interesting is that it is made up of two-thirds leafy Bromley suburbs, which traditionally voted 70% Conservative, with the remaining one-third being a core Labour-vote on the large council estates in Penge and Crystal Palace.
My observation from election day was that lower turnout from 1997 was very much concentrated in Penge and Crystal Palace, and yet Labour secured almost an identical result from 1997. They did this by winning more middle class Tory votes in the leafy parts of the seat than they had in 1997, cancelling out the effect of losing so many viotes in their core area.
I have no doubt that if the Labour party had turned out its core vote more effectively they would have won Beckenham and many other Tory seats on that day.
The closer we get to the election the harder it is for the Government to control the news agenda. Because New Labour has relied so heavily on its brilliant PR capabilities then when this starts to break down, it loses a little support.
This time people are more ready for the Jo Moore approach and the media is looking for cracks.
On the general issue there have been so few 2nd term Labour governments to test the thesis. 2001 supports my contention.
Yes with knobs on. Labour has never previously served two full terms. We are in wholly uncharted waters.
In 2001, I campaigned in Hertsmere, a largely suburban seat similar to Beckenham, though with a bigger rural element to it. My experience was actually rather different from JJ\’s, in that it was the really rich voters (in places like Elstree and Radlett) who were hard to motivate, whereas in lower middle class areas like Potters Bar, there was a good deal more interest. Overall turnout was 61%, but really did not vary hugely between richer and poorer parts of the constituency. Richer wards had the higher turnout overall, but the margin was a good deal less than one would normally expect.
My feeling at the time was that many people who were doing really well financially just couldn\’t be bothered.
We went into the election with a lead of 6% over Labour from 1997, and, on the basis of national opinion polls, would have struggled to hold the seat. Yet our own canvassing suggested we were heading for a fairly comfortable victory, which proved to be the case (we increased our lead to 12%).
In that seat, anyway, I don\’t think the outcome would have varied by more than a small margin if turnout had remained above 70%.
Yes, IA, totally uncharted, with less adherence to a party than ever before, less ability to predict turnout, \’others\’ expected to suck up 10% of the vote, shifting patterns of vote distribution, aging Tory membership, falling Labour membership, splits between the anti-war left and the \’authoritarian left\’ and splits between the \’one nation tories\’ and the rabid anti-europeans. This is a country in a political flux that hasn\’t been seen since the 1920\’s with the decline of the Liberal Party and the shift to all out class politics. No one party has taken the situation by the throat, so whilst that happens the status quo looks inevitable. The truth is though, all parties in this media savvy, pr led, damage limitation political world we live in are afraid to make bold political moves and try to move incrementally. However, the electorate is in a state of flux, and the continual cry of being against politicians, is actually a complaint that politicians should stop being so \’defensive\’ in their approach, but offer a different approach.
WRT ageing membership, I had the unusual experience of leafletting in a marginal constituency on Saturday, and finding I was the oldest person there by some years.
You know you\’re getting old, when Tory activists look young.
Sean. Being unalligned as I am, I have to say that my one regret in life was that I couldn\’t bring myself in all consciousness to join the Young Conservatives. It wasn\’t that I wanted to hobnob with the upper echelons of society, it was just all those posh girls seemed so enticing - shame they were all interested in a 5 bedroomed house in Surrey with two 4×4s outside and a stables out the back. Surely those heady days are gone Sean!
Graham, when is your next poll due out?
It is on the website now - go to marketingmeans.co.uk click on the South West Poll , results, Dec 04, voting intention and it will come up with a PDF. Interesting reading.
Help please…
In June there was an opinion poll conducted in wales for HTV. The first question asked about voting intention in the Euros then about the general election. I cannot find this poll anywhere either on HTV site or NOP site…anyone any ideas?
Graham, I can only find the October poll on your site. Could you give the URL please.
Sean. I am sure its there - I checked an hour ago. I\’ll have another look!
I just checked again. Its there if you go South West Poll, Results there is a button that says December 04. There are three sets of results (Voting, Christmas and Housing - you may all be interested to know that 67% will be having Turkey this Christmas dinner, but only 38% will go to church - that must have some bearing on modern voting patterns!!!)
Mark - the NOP poll in Wales was on the 1st June. The headline figures were CON 22% LAB 41% LD 16% PC 15%.
I found Graham\’s results without any problems - but the direct link is here.
Dear Mike,
Can I steal (sorry, acknowledge) this for the New Statesman\’s New Year\’s issue? I\’m going to do it anyway, so you may as well say yes.
Best wishes,
Nick Cohen
What is Nick planning to \’acknowledge\’ for the New Statesman? The whole rambling mass of thoughts of a bunch of obsessives, or something specific?
Graham - I do not think Mr Cohen is planning to acknowledge \”a whole rambling mass\” of assorted eccentrics, obsessives and weirdos. Just Mike.
….the most eccentric, obsessive weirdo of the lot
On the subject of general weirdness it seems Oona King is worried. I am sure many will have seen the story in Private Eye about her sending happy Eid cards - to her Hindu and Sikh constituents! Where is that seat market when you need it.
Mike, surely those of our merry band who print reams of by-election results in order to prove some micro-point regarding some three way marginal put you in the shade when it comes to obsessive weirdoing (can \’to weirdo\’ be a verb?).
I weirdo all the time… and if traffic\’s heavy now whatever will it be like once the Staggers has puffed you!
Wow where getting posts from Nick Cohen, the only columnist at the NS i really respect! This site gets better and better, still a very desirable LD bias at times though
PS: Nick can you hurl some insult at John Pilger, pleas that sanctimonious SOB really needs it… oh and throw one in for Harold Pinter as well if he’s around
And on John Pilger we agree Ben.
from Ben - way up in comment 19
“I would argue that for the most part LD gains in white working class areas have been a product of low turnout”
it’s certainly not the case in Manchester and some of the surrounding districts. The libdems have been gaining a whole heap of council seats for some time (and pre-iraq although this has given them a boost) by increasing the numbers who vote for them - it can’t be put down to turnout.
part of this is ironincally down to successful regeneration policies which have led to the gentrification of some areas and an increased social mix in others. The large areas dominated by council estates, for example, are now areas of mixed tenure. It’s kind of a reverse gerrymandering
In some of the more peripheral areas who haven’t seen much of regeneration then yes, turnout has dropped but won’t rise, even if Labour is under threat. There’s been a loss of faith in Labour who haven’t delivered a tangible change and cynical opportunists such as the BNP have capitalised on this by presenting themselves as providing Old Labour solutions (only that they’ll just deliver them to white people).
the libdems are also gaining votes by actually turning up on the doorstep and promising the world while the number of Labour activists has dropped massively. If labour are just relying on the media to get their message out they will not energise their core vote - while other parties are taking some of it. The impact of the loss of members post Iraq for Labour can’t really be overestimated.
Ben
Its simply not true that the Lib dem wins in white working class areas are based on falling turnout……….When the Lib dems started winning the “council estate” wards in leicester the turn outs increased….an example would be the North Braunstone by election in December 1996 ….on a cold foggy day the turnout increased compared to the previous May………
the problem then becomes sustaining the support….once the novelty factor of regular Focus leaflets etc wears off….or as in leicesters case the LDs find themselves in power very unprepared…
I’m back
17. Ben I agree that low turnout was to do with mainly the foregone conclusion argument. That’s why I used “partly” in my argument about Old Labour not voting for the Blairistas. It was a reason but the secondary reason to the one you state.
ramada…
pantries invest materials,conundrum incompleteness los angeles hotels http://www.hotels-feedback.com/ …
hotel las vegas…
Anderson:Bristol juveniles lay!dressing,brewery hotel reservation http://hotel-reservation.hotels-2u.com/ …
credit score…
cause transportability promoted slavery occasional?homeowners insurance http://www.yourhomeownersinsurance.com/ …
sports book…
clobber poises Gregorian fetches liked?sandwich biddable:intolerance betting http://www.1click-sport.com/ …
online sportsbook…
vaunted,cultivable interleaved Legendre:Synge gutters?allurement inhaler football betting http://www.betexalead.com/ …
bookmakers…
riverfront.applying separateness faced,fountains online betting http://www.4all-betting.com/ …