More election stats

Second places
We know the Liberal Democrats didn’t end up increasing the number of seats they hold, but they did substantially increase the number of second places they have, and have more winnable marginals. The notional 2005 figures had the Lib Dems holding 62 seats and in second place in 188. Following the 2010 election the Lib Dems hold 57 seats, but are in second place in 242. On the 2005 notional figures the Lib Dems were within 10% of the winning party in 31 seats, now they are within 10% in 45 seats.

Conversely, Labour held 348 seats and were in second place in 151 – a total of 499. Those figures are now 258 and 160 (assuming they retain second place in Thirsk and Malton), a total of 418 and suggesting they have dropped to third place in an additional 81 seats.

Swings needed
Based on the 2010 results, the Conservatives would need a swing of 2% in order to gain an overall majority (meaning they would still need a lead of roughly 11 points over Labour to win an overall majority). In short, any effect from unwinding tactical voting or shifting voting patterns has not made the system kinder to the Conservatives.

However, it has got less kind to Labour. On the notional 2005 result, Labour could have got an overall majority by getting an equal number of votes to the Conservatives. From the 2010 results, Labour would require a swing of 5% in order to gain an overall majority, the equivalent of being 3 psoints ahead of the Conservatives. For Labour to become the largest party in a hung Parliament they would need a swing of 1.7%.

Regional differences
We have past instances of Scotland behaving somewhat differently to the rest of Great Britain (most obviously 1992, when England and Wales swung towards Labour, but Scotland swung to the Tories). This election produced an extreme contrast – in England and Wales there was a swing of between 5-6% from Labour to the Conservatives, in Scotland there was a 1% swing towards Labour, mostly at the expense of the Liberal Democrats, whose vote rose in England and Wales.

More unusually there was a significant difference between London and the rest of England. In London the swing to the Conservatives was only 2.5%, compared to 6.1% in the rest of England. Labour’s vote fell by 2.3% in London, but 8.2% elsewhere in England. Perhaps some of it is a Boris effect, but some will also be the high ethnic minority population in London. Labour’s vote seemed to hold up better in seats with a high ethnic minority population, and in some seats with a high proportion of Muslim voters Labour’s vote share increased as the Iraq effect from 2001 faded.

Marginal swing, and a puzzling question
The Conservatives performed only slightly better in marginal seats. In the country as a whole they had a swing of 5.03% from Lab=>Con, in Lab held marginal seats with a majority of under 10% they got a swing of 6%, in Labour held marginal seats with a majority between 10% and 20% they got a swing of 5.13%. This does raise the question of why they got so many seats – they managed 305 seats, when on a uniform swing of 5% they should have got only 289. If they didn’t do better in the marginals, how come they won more seats than they should have?

There are two reasons. Firstly, while the mean average swing in Conservative marginals where they needed a swing of between 5% and 10% was only 5.13%, the median swing in those seats was 5.84%. The mean was dragged down by some Scottish and London marginals where the Tories went backwards, but in most seats in that range the Conservatives did slightly better than their average performance across the country. The other reason is sheer, dumb luck. There were 11 seats where the Conservatives and Labour were within 1% of each other and the majorities were under 500 votes, the Conservatives won 8 of them.

1,255 Responses to “More election stats”

1 2 3 26
  1. Good homework, AW thanks.

    I think more analysis on Lib dem Con seats will reveal astonishing variances.

  2. Fascinating stuff Anthony. Thanks

  3. So the marginal premium was .84%
    I’d love to know if Labour performed the best in marginals 50 – 100 as I believe that is where they focussed their resources.

  4. And in London there was actually a minus premium of 2.5%

  5. @Sue,

    Yes, the marginals premium was relatively small. But the Tories still outperformed UNS considerably. It will take a fair old while and a lot of tired calculator fingers before we work it all out I think.

    Of course, if we get PR we’ll never have to worry about it again!!

  6. Neil A – Does PR in any or all its forms mean the end of marginals? I see profit in devising multiple swingometers and ever more money being spent on targetting of the voters

    I’ve been away, so a[pologies if it’s ben covered, but the uncanny accuracy of the exit poll (which depended partly on the canvassers’subjective view of the physique of voters) doesn’t mean that the whole process of spending money in marginals will end.

    The funders of all sides’ campaigns will spend more, not less in the marginals, only they’ll demand more influence on the exposure of leaders on the TV together. That more than anything limited the effect of highly resourced, targetted campaigning because everyone has a telly.

    Buying power requires more money to be spent in the same way that the first breach of the sound barrier required more speed.

    PR will raise the stakes – it will cost more to influence the electorate, and no-one’s going to say “the game’s up, might as well spend it onsomething else”

  7. Both your pieces on Election Stats are most interesting and as has already been said, we all need to analyse all sorts of things before drawing conclusions ( if we ever can draw them).

    It would be interesting to look at:
    1. Regional differences in swing (not just England to Wales to Scotland)
    2. Incumbency
    3. Was turnout higher in marginals as appears at first sight?

    Many thanks for your research.

    [I can answer question 1 now. Apart from London there were only small differences between English regions, nothing worth writing home about – AW]

  8. @AW
    Labour’s vote fell by 2.3% in London, but 8.2% elsewhere in London

    Could you clarify please?

    [Er… elsewhere in England :) – AW]

  9. Anyone care to hazard a guess how the seats would have turned out if we had been voting using an AV system – preferred to a true Proportional system by Labour and IMHO the only possibility for Tory acceptance, since it retains the direct link to an MP for a (relatively) small constituency?

  10. wow, the sheer luck part of the 11 seats is interesting.

    I do maintain, if Labour had had over 265 seats, I think a Lab/Lib rainbow coalition might be favourite.

    are these the sort of margins we are dealing with?!


  11. So am I right in thinking that Lib Dem support has spread out even more thinly, even further?

  12. Very interesting AW. Thanks.
    Obviously very close, and difficult for polls to have predicted. Amazing , in fact that the exit polls were so close.
    Of course we would expect concentration of funds/active electioneering in marginal seats. Are there any statistics available as to how postal votes were distributed, and whether this has any significance? Or is this data not public?

  13. PeterMK – I think the Independent on Sunday did a calculation and it produced predictably results that had the Tories still the largest with fewer, Labour with fewer and LDs with a lot more (but still very much fewest).

    The Av, AV+ and STV had effects that were progressively worse and worse for Blue, but all left them largest (and presumably in a position to dictate negotations in the same way that they are now)

  14. @John TT,

    Well of course it all depends on the system. There will be marginals under any kind of “top up” system, but spending a fortune to take a FPTP marginal only to have that victory offset by triggering a top up seat to another party will often be pointless. Under STV more money and energy would be poured into securing second-place votes from your rivals supporters than anything else.

    A pure list system (which I’d utterly object to) would remove the concept of “marginals” in terms of constituencies, although it might result in some regional targeting. For example if the vote shares in the SW region meant that it was a tossup between the final seat going to one of two parties, it would be worth pouring resources into that region.

    All in all it would severely reduce the scope for marginals “meaning” anything.

  15. Fascinating,

    Onedid feel that London performed worse for the Cons and better for Labour. A lot has been said this is because of the recession uncertanties.

    My view is that housing values in the capital haven’t fallen in the past year and unemployment doesn’t seem to be as bad in the capital. The financial services industry has also been recovering.Certainly the ethnic minorities on the whole supported Labour.

    I’d be interested in how many seats the Tories took where Labour and Lib/Dem duked it out. Hampstead and Kilburn nearly went that way.

  16. Neil A- you make sense and I’m assuming a full list system is rightly rejected by all

    my main point is that the more apparently futile the spending, the more will be spent, simply because of the prevailing notion that heavy spending works, and that in PR system, you need to spend even more to overcome the off-setting and topping-offs and have the same effeect as under FPTP

    I find it ironic that the Conservatives subscribe to idea that throwing money at targets works – I’d have thought the opposite would be more their bag (ie get volunteers to do stuff for no payment – more for less, community spare time action etc) Perhaps they do that as well , but it doews make me smile to see people happily signing huge cheques in the belief it will solve anything (while at the same time railing against the Govt’s doiung the same thing)

  17. Unfortunately, even London is divided into the very rich boroughs (where virtually every single house is worth over £1 million) and the poor boroughs (i.e. those with lots of Council Houses). There is no reason whatsoever for the poor boroughs to vote Conservative – this is something the Tories have to address.

  18. Dumb luck indeed and I don’t I know it – 332 votes in Stockton South, probably a slight Cous effect as in travel to work area.
    Just glad it was not less than 50 as we would have been wondering if we had given up one more night etc.

    Imagine what the defeated candidate and her/his team in NI are feeling like – 4 votes!!
    Re London – Anthony, am I right that London swung a litle more to the cons in 2005, so perhaps some of the potential swing was already embedded?

  19. @Sherwick,

    Hmm, no I’m afraid I don’t agree with your characterisation of London. There are one or two boroughs that are basically big council estates (Barking and Dagenham is the most obvious) but the vast majority of London boroughs have a split between poverty and affluence. If anything, the lack of space means that in London those areas are cheek-by-jowl with one another. I can remember policing Islington, where there was a lovely townhouse worth about £1.5m and its garden wall seperated it from the New River estate – a cesspit of drugs, despair and death. Hackney is much the same.

    I haven’t reached a view about the Labour strength in London yet. It may be a diversity issue, with immigration and LGBT issues suppressing the Tory swing (given what happened in the campaign). But something definitely went on there.

  20. Jim Jam – and what about the local election in Yarmouth North? A dead heat, with Labour winning on the draw of a playing card!

  21. @John TT,

    I think the Tories throw money at things because they have money to throw at things.

    To me, the best use of cash isn’t in swanky poster sites and slick commercials, it’s in providing salaried staff for the local Conservative Association, subsidising their events and buildings etc.

    When I was a Young Conservative (in Enfield North) there was a local rock band whose guitarist was a fellow Tory. They played every Friday and Saturday night at the Conservative Association and it drew scores (sometimes hundreds) of local youths. It may not have directly recruited any voters, but those kids at least didn’t develop a “don’t have anything to do with Tories” mindset.

    If a local Association was struggling and was in danger of losing it’s premises, the best use of Ashcroft money would be to bail it out. In areas where there are no premises, Ashcroft should be donating money to acquire one.

    I doubt that’s how he thinks though. Tory success in 20 years time is probably not high up on his list of concerns.

  22. UKIP 917,175 votes but no seats (4th largest party in terms of votes)
    BNP 564,321 but no seats (5th largest party in terms of votes)

    Then compare that with..

    SNP 491,376 with 6 seats
    Green 285,616 with 1 seat
    Sinn Féin 171,942 with 5 seats
    DUP 168,216 with 8 seats
    Plaid Cymru 165,394 with 3 seats
    SDLP 110,970 with 3 seats

  23. Neil A – I agree and Union organisational experience similarly is very beneficial to Labour.

    Money useful of course.

  24. These posts are really good information, cheers Anthony. I’m afraid I am going to be a bit pedantic about 2 errors though:

    1. As mentioned above, it should read “Labour’s vote fell by 2.3% in London, but 8.2% elsewhere in England.”

    2. “and in some seats with a high proportion of Muslim voters Labour’s vote share increased as the Iraq effect from 2001 faded.” Should presumably read ‘2005’ since the 2001 election was not only before the Iraq war but even 9/11.

  25. Neil A – all healthy – hopefully lessons will be learned (by all sides) about rich patrons with ethically challengeable motives and levels of validity. What a contrast between means of acquiring support that you describe.

    All up for grabs? <aybe the "big society" starts at home in the local offices of the parties?

  26. Ash:

    Surely you missed the best example, Alliance. 1 seat with 42K votes, over a quarter of which were in that constituency.

  27. @Jim Jam,

    Yup, money is useful where it facilitates the recruitment, retention and development of people.

    Ashcroft should set up a kind of bursary scheme where Tory associations can get grants to pay for agents and admin staff. I know this happens for the constituency offices of MPs but it could be broadened out.

  28. “I doubt that’s how he thinks though. Tory success in 20 years time is probably not high up on his list of concerns.”

    Perhaps it should be though, inheritance tax and all…

  29. @John TT,

    I never really subscribed to the notion that Ashcroft had “bought” the Tory party. He is clearly very interested in politics and wanted to do what he could. Given that he’s extremely rich “what he could” included spending a lot of cash. The fact that he is now fulminating against Cameron’s coalition plan gives me a certain sense of satisfaction.

  30. Neil A

    Am a great admirer of the good and often thankless work done by the police, but am shocked by your characterisation of the New River Estate. Are there some good things you could say about any of the people who live there?

  31. @Quincel,

    Do they have inheritance tax in Belize? ;)

  32. @Billy Bob,

    In fairness it was “under refurbishment” when I left that posting, but it was awful. Obviously as the police we don’t generally get called when anything nice is happening, but that estate, along with the Andover estate in Holloway, was soul destroying.

    As with anything, I am sure a lot of the people living there were lovely. It’s just that as a copper you never meet those people.

  33. “@Quincel,

    Do they have inheritance tax in Belize? ;)”

    Touche, though this is straying into partisan chat, so I’ll go to bed before it gets any further.

  34. Neil A

    Thanks for your reply. I dare say the nice people were glad you made frequent visits :-)

  35. I still think AV in the Brown sense would have only a limited impact (in 3 way marginals) and that Labour could have more to lose from it than the Conservatives, I have posted on this before.

    I note that it has been pointed out that there are two (at least) variations on list-based PR:
    (1) national list. One national list, all votes counted together and split pro rata i.e. 36% of the national vote means 36% of the seats in the Commons.
    (2) regional lists. In this version each region would have its own separate count and lists, e.g. south east, central southern, south west etc. The impact of this would depend partly on the size of the regions. The smaller the regions, the less impact it would have.

  36. Whether we get Proportional Representation in the short term is one thing. I can’t help thinking it won’t be many elections away if fourth party votes continue to increase. If 3 way party politics puts FPTP under severve strain, 4+ party politics breaks it. I mean, a pat on the back to Caroline Lucas winning Brighton Pavillion, but with 31.33% of the vote????? This election produced 11 Mps elected on less than one third of the vote.

  37. What about the Conservative argument that ordinary voters are just not bothered about electoral reform?

  38. Neil, large swathes of South East London and (i think) North East London are dominated by Council Housing.

    This differs markedly from South West London. (North West London is fairly mixed though).

    I think the results correlate more with affluence than with ethicity, though that’s just a gut feel.

  39. @Ash

    Such discrepancies also exist in countries with have some variant of PR, but on a regional and not on an national basis. For example in Spain, the nationalist center-right party Convergencia i Unio, that exists only in Catalonia, has 10 seats with 3,2% of the votes, whilst the leftist Izquierda Unida, with 3,8%, has only 2 sets, because its votes are almost evenly spread all over Spain. The same applies to other nationalists and regional parties of Spain, such as the PNV (Basque Country) and the BNG (Galicia). A similar (albeit not identical phenomenon) occurs in Italy with the Lega Nord.

  40. @Sherwick,

    The proportions vary, but in every area I have direct experience of (North, NorthEast, East, SouthEast London) there is a mix. As I say, Barking and Dagenham is probably the exception, but that is because it was essentially a rural area converted to an industrial dormitory in a short space of time – a “new town” effectively.

    As you no doubt know, London’s growth from Middlesex/Essex/Surry/Kent towns and villages to massive conurbation has left “nice bits” all over the place. These tend to still be pockets of affluence. The infill varies, from pleasant treelined 30s semis to massive concrete wastelands. Lots of good work was done in the past 20 years to replace the worst of the concrete with attractive terraced housing, but the demographic pattern was already set.

  41. Since this is a polling site, I’m going to set up a poll. And set it up with my preferred system: AV.

    The poll is simple; rank voting systems in order of preference.

    To set the ball rolling, here’s mine:
    1 – AV
    2 – STV
    3 – AV+
    4 – FPTP
    5 – List

    Anyone else?

  42. You’ve missed out AMS, as in Scottish and Welsh devolved assemblies, GLA – and Germany.

  43. @Robert

    1) AMS is the same as AV+
    2) I doubtless missed a large number of other systems used around the world. I merely listed the 5 most quoted for potential use here.
    3) I was asking for peoples views, not their cleverness.

  44. @ Richard O

    “wow, the sheer luck part of the 11 seats is interesting.

    I do maintain, if Labour had had over 265 seats, I think a Lab/Lib rainbow coalition might be favourite. ”

    But if the 11 had split ‘evenly’ [6/5] it would have only added 2 or 3 to Lab taking then to 260/261. It would have needed a super-lucky 10/1 split to get them to 265!

  45. i’d go,

    1. STV
    5. List

  46. Comparisons between the northern irish parties, the nationalist parties and UKIP/BNP are not fair as BNP and UKIP fought in most seats across the country. A more interesting stat might be the average vote per seat fought.

  47. I wonder how much of the variable swings can be explained by postal vote fraud? I seem to remember that a couple of days before the election the police were investigating 50 complaints, but it’s gone very quiet since.

    [I think they were 50 complaints…in Tower Hamlets. So even if those complaints were well founded, it could only affect a couple of seats! – AW]

  48. @Steve Coberman: Unless got it all wrong, AV+ and AMS are NOT the same. Just vaguely similar. There is no ranking by the voter in AMS.

    @Robert Waller: Not QUITE. Some states in Germany have single-vote-AMS, but the Bundestag and most state parliaments use a system of two votes: one for a list of candidates for the whole state, the “Landesstimme”, and one vote for the candidates in the respective constituency, the “Wahlkreisstimme”. The distribution of seats follows the relation of “Landesstimmen”, although for political reasons, the “Landesstimme” is refered to as “Zweitstimme” [“secondary vote”], especially by smaller parties who have no chances to win constituencies, but can increase their share of seats by winning lots of those votes as so called “lended votes” (“Leihstimmen”), when the followers of a big party graciously vote for a potential junior coalition partner’s list, not knowing that they deprive their own party of the one vote that counts the most… ;-)

  49. “What about the Conservative argument that ordinary voters are just not bothered about electoral reform?”

    They aren’t bothered. When I was out canvassing not one person mentioned electoral reform as such. What does concern people is the notion that politicians are out-of-touch, corrupt, untrustworthy, don’t listen, etc. Now, if you can convince people that electoral reform will fix these problems, you’re golden, although that would presumably depend on the reform you are suggesting (a list system, for instance, would almost certainly make things worse).

    Most voters don’t care about 23% of the vote not getting 23% of the seats: they see themselves as electing a government, not a legislature, so there needs to be a winner. I think the Lib Dems are making a huge mistake if they let negotiations break down over PR. If you ask people their voting priorities, it’s the economy, the deficit, schools, hospitals, crime, immigration, schools… if a new system for choosing the government means we’re better able to tackle these problems, then all well and good. But the system is not an end in itself.

  50. Ordinary voters aren’t bothered about electoral reform? What does an “ordinary voter” look like? Lib Dem voters are clearly very bothered about it for a start, never mind supporters of the minor parties.

1 2 3 26