New Constituency Boundaries
As promised many months ago, I have produced an updated estimate of the effect the new constituency boundary changes will have on the political make up of the country. As most readers will know, every 12 to 14 years the counstituency boundaries on which MPs are elected are updated to take account of changes in the population. The latest changes for England and Wales have just been completed and, baring legal challenges, will be in place in time for the next election.
The main effect will be to make the boundaries more favourable to the Conservative party. If the 2005 election had been fought on the new boundaries the Conservatives would have won 214 seats, rather than the 197 they actually won, and Labour would have had a majority of 36, rather than 64. It also slightly reduces the swing the Conservatives need to get an overall majority - down from 7.4% to 7.1%.
Since the last version of my notional results I have corrected some errors than people pointed out to me, and have factored in some feedback from people with local knowledge of how particular wards voted at a Parliamentary level.
The updated figures are available at UKPolling Report’s new Election Guide. This includes notional results for every seat in Great Britain, lists of target seats for each of the three main parties and lits of candidates selected so far. Apart from any blindingly obvious errors, I will not be updating the figures again - it will probably not be long until Ralling and Thrasher’s “official” figures are out anyway. However, every seat has a separate page for comments allowing you to add feedback if you think the seat will end up voting differently from how the notional figures suggest or just ot comment on the politics of the seat in general and what you think the result will be at the next election. Obviously the chances are that the next election will not be for several years, but hopefully putting up the guide now will let it build into a more useful resource by the time the election does come around.
There are still a few errors in the guide pages (I had to put Wyre Forest into the target lists by hand so some of the “target numbers” are still skew-iff, and there is one Welsh seat that claims to have two Plaid Cymru candidates and no Lib Dems), but they should be ironed out in the next few days.
UPDATE: Comments are now working properly as well!
Filed under: Uncategorized
if the conservatives gain power they should insist on boundary changes that do not favour labour.this marginal gain does not rectify the problem of the tories needing at least 8% more votes to achieve a similar seat result.
An excellent job Anthony , probably be some minor quibbles but these things can never be done with 100% accuracy . Will your swingometer be updated to include any changes since your original forecast ?
John , the reason for the bias in the sysyem is not so much caused by the way constituency boundaries are drawn and will not therefore be solved by insisting on more . Baxter has a report on the reasons for the bias and Lewis Baston did a similar analysis on the Electoral Reform Society Site and I would refer you to these studies .
Mark - yep. Hopefully on Monday, but sometime next week anyway.
Don’t suppose we could get an SNP target list added.
Peter.
Peter - one of the things I plan to do (though obviously the boundary changes don’t make any difference to it).
It should be noted that every year the Conservative vote is eroded by around 0.25% , 1% over a 4 year parliament as older votes who die are replaced by younger voters less likely to vote Conservative . It is true that there is a tendency for voters to become more Conservative as they get older but this change will be masked in the general swing between elections .
Anthony’s calculation of the swing required for a Conservative overall majority is therefore around 8.1% some of which will be achieved by a natural tendency for people to change to the Conservatives as they become older .
Excellent work, Anthony.
Mark, have you plucked those figures out of something related to thin air, or have you got a link which explains them?
Mar,
But we have an aging population.
If we have 55 more likely to vote and more likely to vote Tory and the over 55’s fast becoming larger than the low turnout pro labour under 25’s, then in actual fact it may well be that time is actually on the Tories side.
A simple experiment would be to take the breakdown of the population today using the YouGov age breaks and work out how people would vote, and then do the same using the age predictions for ten years time using the same voting intentions.
Another factor in demographic trend is household size.
With more and more single or twin occupancy homes and a rise in middleclass city living and “Yuppyisation” it may be that the Labour dominance of some inner cities will be erroded over time.
Compared to twenty years ago I’d suspect that there are some parts of London where traditional Labour voters have all but been replaced by affluent middle class people far more willing to vote Tory.
Peter.
Sorry Mark,
Missed a “k”, I should prooof read my posts, but I’d probably end up never posting.
Peter.
Sorry - but I had to laugh at Mark Senior’s posting
Rik W - Well you would laugh at simple Maths wouldn’t you .
Richard - the Maths are fairly simple , Take the Conservative vote share of the Over 55/65 age group from one of the last GE surveys , calculate the loss in no of Conservative votes per year through death . Take vote share of 18/25 age group and calculate no of new Conservative voters per year and subtract from loss previous calculated . . The Conservatives loss is around 90/100,00 votes per year equating to a loss of vote share of around 0.25% a year .
Peter yes I did say that older people are more likely to vote Conservative but at some particular stage someone getting older may take the decision to change to voying Conservative but that will be part of the swing required for Conservatives to win the next election not only as Anthony has calculated to be 7.1% but also the extra loss through natural wastage . Your mention of the propensity of older people to be more likely to vote could be valid if they become more more likely to vote than they were at the last election or the Under 25’s become less likely .
I’m pretty sure that’s incorrect Mark.
Anthony pointed out sometime ago that the Tories came third among the 18-34 age group in October 1974, yet are now in the lead among the same group of voters 32 years on.
Sean see the detailed data for % vote share from their final 2005 election polls which are on their website :-
ICM 18-24 yearolds Lab 36 LibDem 26 Con 25
Populus “” “” Lab 39 LibDem 30 Con 23
This is a good piece of work, Anthony, but I was wondering whether you would explain the reasoning for using the GLA constituency vote when calculating the notionals in London seats rather than the most recent council elections as you have done in the rest of the country?
Henry - the problems with local election results are wards that are uncontested where you have to make up results, wards where a particular party doesn’t contest them where you have to make up a notional result, wards where a strong independent or residents candidate skews the vote and wards where the personal vote of a long standing councillor skews the vote. It’s also normally impossible to come up with notional figures for fringe parties like Respect or the BNP because of a lack of council candidates.
Using the GLA figures takes all those problems away, gives you a vote for every party in every ward without any real skew from independent candidates in particular wards. It also let me do proper notionals for the BNP and Respect in the areas where they are strong in London.
Mark, and how have you accounted for the possibility that Labour voters die younger? The deaths of those every year aged 70+ will harm the Conservatives, but as you start moving into those dying in their fifties and sixties deaths are likely to start moving in their favour.
Richard - Please quote your source for the evidence that Labour voters die younger .
Mark, according to UK National statistics comparing 2003 with 2031, the percentage of people under 16 will drop from 20% to 17% and over 65 from 16% to 23%.
Roughly that has younger voters dropping by 10% and older voters rising by close to 40%.
Now lets take a simple version.
We say that we have under 30, 30 to 60 and over 60.
If the ratios are,
6, 7, 7 now, in age ( total 20) but the voting rates are 50%, 60% and 70 %, then we get. 3, 4.2, and 4.9 giving us an average turnout of 60.5%.
But if the ratios change to 5, 7, 8 then we get 2.5, 4.2, and 5.6, which is 61.5%, not much of achange overall, but the young to old ratio has changed from about 3 to 5 to more than 1 to 2, and that makes a big difference if the old vote differently from the young.
If we said the vote was young 70/30, middle 50/50 and old 40/60, then on 6, 7, 7 we would get 4.2/1.8, 3.5/3.5 and 2.8/4.2, giving a final result of 10.5 to 9.5 0r 55% to 45%.
Now if we add voting intention to that we would get 2.1/0.9, 2.1/2.1 and 2.0/2.9 or 6.2 to 5.9 about 53% to 47%.
But if we change the ratio from 6, 6, 7 to 5,7,8 then we get 1.75/0.75, 1.75/1.75, 2.2/3.4 which gives 5.7 to 5.9 so the oldies carry the day because there are more of them and they are more likely to vot.
Sure they will die but as people live longer and there are fewer young people as family size diminishes the odds are still in the conservatives favour, if and it’s a big if, the parties stick to traditional positions.
As I noted once before if you look at attitudes to comprehensive education, most young people are happy with it and don’t have a problem, but it’s the over 55’s who see it as poor and need ing change.
Who’s oppinions are most evident in the policies of both Labour and Tory… The over 55’s, because they are the ones that vote, along with motivated middle class parents.
So education policy is driven more by electoral demographics than the needs of children or the education system.
Peter.
You don’t believe it?
I thought everyone accepted that mortality was related to social class. Obviously things aren’t as clearcut as they used to be as party affiliations have loosened, the link between voting and social class remains strong.
Probably some stuff here, here or here for starters. But that’s just using Google.
Peter all you say is correct and the number of older people will increase but the Conservatives at some stage must change the voting intention from someone who was voting non Conservative to Conservative and that change will be relected in the swing between General Elections .
What would be fair to say is that the swing needed for a Conservative majority which Anthony states is 7.1% because of death rates is actually 8.1% but the Conservatives can pretty much count on x% of that being achieved by the propensity of voters to vote Conservative as they get older . I am not capable of putting a value to x though .
Richard - Sorry to keep on this morbid topic . Around 85% of annual deaths are people aged 65 and over and 65% aged 75 and over . Yes OK those younger to die may well be more likely to be Labour voters but the older ones by inference are more likely to be Conservative and the approximate figure I have given for the effect on voting remains valid .
Mark Senior - you are quoting out of date polls for your age related voting stats. MOre recent polls show the Conservatives doing much better.
Henry, just to add to Anthony’s response: the ward figures use, essentially, only there to work out the relative contributions each party in each ward makes to the overall constituency vote: they aren’t actually used in the notional forecasts.
In other words, they should show that party X draws most of its strength in a particular constituency from wards A, D, F and H, and does much less well in wards B, C, E and G (this should by and large be the case whatever the overall performance of the party is).
If wards A and D are transferred to a new constituency, it is likely that a greater proportion of party X’s strength is transferred across than if wards C and E are moved.
Hence the reason why it can be seen that, for example, Staffordshire Moorlands or Northampton South changes political allegiance; whereas just extrapolating the overall result from the old seat to the new one can produce inaccurate results.
That said, I’m not as convinced as Anthony that the GLA results are as sound a basis for measuring the party shares than the 2006 local elections are (because the vote was far more fragmented than in 2006) - but Anthony would have started his analysis before the 2006 results were available.
As I’m finding out currently, it’s a massive job!
Rik W - Average of last 2 polls by ICM has Conservative support at only 15% for 18-24 year olds
Ummmm … the Wrekin constituency is missing.
It’s here -
http://ukpollingreport.co.uk/guide/seat-profiles/thewrekin
I see that what I wrote was ambiguous Mark. The 18-34 year olds who placed us third in October 1974, are now the 50-66 year olds among whom we lead.
I don’t think it helps at all to try and add in estimated demographic change to the swing needed by any party to win an overall majority. There will always be demographic factors working both for and against the parties.
May 2005 suggests that the Conservatives are more than replacing lost voters. If your estimate is correct, then c.400,000 people who voted Conservative in 2001 had died by 2005. Yet the overall Conservative vote rose by 400,000, and, at the same time, suggesting that the Conservatives had obtained 800,000 new voters.
It shouldn’t make that much difference whether you use the London Assembly ward results, or those for 2006, as the basis for calculating notional results on the new boundaries.
But I think Anthony is right to use the former, for the reasons he has given.
Sean yes you are correct that the Conservatives more than replaced their voters who had died between 2001 and 2005 but not by anywhere near the factor required to give an overall majority .
I think Mark Senior is right. Older voters are so overwhelmingly Tory compared to younger voters that it is inevitable that over time the Tory vote will fall slightly, all other things being equal.
I believe this will occur despite the fact that the population and electorate are ageing.
The best example of this is perhaps working class women: those born before about 1945 have a relatively strong tendency to vote Conservative despite their relatively poor status. Working class women born since 1945 are massively less likely to do so. As these older working class women die, it cannot but have the effect of depressing the Conservative vote to a certain degree.
I can’t access the Richmond (Yorkshire) figures. It comes up with ‘Error 404 - Not Found’.
If the 2005 General Election had been fought on the new boundaries how many seats would the Liberal Democrats have won, and why did you not say so in your article?
John, I think the brackets in the seat name must have buggered things up somewhere, because the page doesn’t seem to exist! I’ll add it at some point.
Bruce, this is just a pointer to the actual figures. Look at the guide itself and the overview of the changes.
Sean, except that:
a) there were no uncontested London wards in 2006 so that doesn’t apply (and very few wards not contested by Tories, Labour or at least one Lib Dem)
b) the only borough where independents skewed the result notably in 2006 was Havering; whereas there were several independents standing everywhere in the GLA elections
c) which GLA results do you use: Mayoral (best for Labour), Assembly (worst) or Euro (intermediate). The Mayoral results in somewhere like Wandsworth most closely replicate General Election voting, but in other boroughs it overstates Labour’s vote - and is disproportionately spread
d) the Respect vote in Tower Hamlets and Newham in 2004 were stronger than their 2005 results in those boroughs; 2006 is closer
Adam,
I tried it out using both the 2006 results and the 2004 results for Enfield North, Enfield Southgate, and Ealing Acton. You have to use a different multiplier in each case, but I found it made scarcely any difference.
But which 2004 results did you use, Sean - Assembly, Mayoral or Euro?
[...] October MatGB01:55 pmAdd comment The Boundary Commission has almost completed their review of constituency boundaries. Anthony Wells of UK Polling Report has run an analysis of the new seats and completed a very excellent guide to the next election, with target seat list for the three main parties and a breakdown of the notional 2005 results and swings needed for each seat in the country. Scary amount of work there, but very impressive. There’s a comment box for every seat, and he welcomes contribution from anyone with local knowledge or facts he may have missed. An excelent resource that I suspect will get a lot of use over the next few years. As for Tooting? From the 214 seats the Conservatives notionally hold on the new boundaries they would need to win an extra 112 to get the 326 seats necessary for a majority. In order of marginality, the 112th most winnable seat for the Tories is Tooting. [...]