stat

Wednesday 26 December 2012

Parachuting Boris?

One of the odd ideas floating around, as demonstrated in this Daily Telegraph article.

The idea is that if Prime Minister David Cameron fails to win the May 2015 general election, then Conservative MPs oust him and Mayor of London Boris Johnson is parachuted in to a safe Conservative seat and then wins and becomes Conservative leader.

So, what is wrong with it? Like all careful plotting, it relies on a cast of tens of thousands knowing their role- constituents and Conservative party members alike.

Firstly, what does it mean to lose a general election? Or to fail to win?

Once upon a time, a single party would win over half the seats in the House of Commons and - incredible as this sounds - Governments would be made up of politicians from just one party. It was immediately obvious who had won and who had lost.

But now? Did Cameron lost the May 2010 general election? He did form a Government comprised of his own Conservatives and the Liberal Democrats. But what if Labour had won just a few more seats and the Conservatives were the largest party and a Labour/Liberal Democrat government had been formed? Would Cameron have lost?

There was a narrative that was common among the Conservatives in the wilderness years. And it can be summed up as:

Margaret Thatcher led us to landslide victories in June 1983 and June 1987. Then MPs knifed her in the back and John Major became Prime Minister and simply threw away a massive majority and led us to a huge defeat in May 1997.

The problem with that narrative is that there was an election between 1987 and 1997- that of April 1992. And look at the result of that:

  • Conservatives - 41.93% (down 0.37%)
  • Labour - 34.39% (up 3.56%)
  • Liberal Democrats* - 17.93% (dowm 4.64%^)
  • * In August 1987 the Liberals and Social Democrats voted to merge, and the Liberal Democrats were formed in March 1988. 3 Social Democrat MPs - Rosie Barnes (Greenwich), John Cartwright (Woolwich) and David Owen (Plymouth Devonport) - chose not to join the Liberal Democrats. Owen retired at the 1992 election, while Barnes and Cartwright sought re-election, and neither faced a Liberal Democrat challenger. I have included them in the Liberal Democrat total

    ^ Compared to the combined Liberal and Social Democrat vote in 1987

    A common explanation is that the Conservative vote held up fairly well, but Liberal and Social Democrat voters switching to Labour pushed plenty of Conservative seats into the Labour column.

    Now consider if the changes in share of the vote were matched at the 2015 election, and we get Labour on 292, Conservatives on 290*, and the Liberal Democrats on 40.

    [* The calculator gives 291 Conservative MPs, but this includes John Bercow, the Speaker of the House of Commons and MP for Buckingham]

    When you remove the Deputy Speakers (2 Labour, 1 Conservative), you have the two main parties tie-ing on 290 MPs. A Conservative/Liberal Demcrat or a Labour/Liberal Democrat government are both possible.

    So, the Conservative vote has held up and the sitting Government can continue. Has Cameron won or lost?

    And what about if the Conservative vote increases, but a large voter switch from the Liberal Democrats to Labour means a Labour Government? Has Cameron lost then?

    In hung parliament and coalition territory, the terms "win" and "lose" are a bit harder to define.

    Secondly, it assumes that Johnson would, in the cold light of day, want to become Conservative leader. Just assume that Cameron has, by a commonly-accepted definition, lost, and Labour leader Ted Miliband is then Prime Minister. Would Johnson really want to throw away a third term in power for opposition? What would he really want to be from May 2016 to May 2020- Mayor of London or Leader of the Opposition?

    Thirdly, to be a candidate for the Conservative leadership he would have to be an MP. There is a story I've heard - and maybe any Labour people reading this can confirm - that one resson Michael Foot resigned as Labour leader after the 1983 election so a leadership election was completed in October 1983 was that Tony Benn had lost his seat in 1983 and Foot wanted the leadership election to be in motion before Benn had the chance to get back into the House of Commons, so that Benn could not be a candidate.

    A quick resignation by Cameron after the 2015 election could lead to the nominations closing while Johnson is in the middle of a by-election campaign.

    Fourthly, the Conservative MPs might not be compliant. In the last 2 leadership elections, MPs have ensured that an obvious winner has not made the top 2 (Michael Portillo in 2001 and the Minister without Portfolio, Ken Clarke, in 2005) from whom the members chose the leader.

    Fifthly, even if Johnson manages to be a candidate, and Conservative MPs let him through to the final round, there is no guarantee that the members would elect him.

    Sixthly, in every constituency there is a group of awkward people. The Parliamentary Voting System & Constituencies Act 2011 has at its heart the idea that in every seat there should be a roughly equal number of awkward people. They are called voters and have an annoying habit of not doing what they are told. While powerful men and women can draw up great schemes, it's the voters who can - armed with no more than a pencil - destroy such a person's dreams and plans. The pen is indeed mightier than the sword.

    The result in Blaenau Gwent at the May 2005 general election is a useful reminder of just how awkward these people can be.

    Voters don't like being taken for granted. Just after the 1983 election, Thatcher appointed William Whitelaw to be Lord President of the Council and Leader of the House of Lords. Naturally enough, he had to leave the House of Commons and there was a by-election in his Penrith & the Border seat in July 1983, which saw David Maclean (later to be Conservative Chief Whip) win, but narrowly, as the Liberals nearly overturned a 30% majority.

    Yes, they might find Johnson a safe seat. But Penrith & the Border was a safe seat. The voters don't like being used. Is someone really going to stand for re-election in 2015, promising to do the full 5 years, while deep down just keeping the seat warm until Johnson decides to stand? That would be treating their constituents with contempt.

    Tuesday 25 December 2012

    Why I Wanted To Be A Deer

    Yesterday late afternoon, after a church service, I walked up to North Baddesley to deliver some Christmas cards. An interesting journey as pavements vanish and one is left walking along grass verges. Wet, slippy grass verges.

    Cards delivered, and time to walk back. You don't have to get very far from North Baddesley before it become quite rural. And, walking down Rownhams Lane, to the left of me there was a rustling noise and the sight of a deer leaping off away from me.

    Thankfully there was a fence, as at least that meant it couldn't leap in the direction of traffic.

    I was a bit annoyed with myself for unintentionally scaring it.

    As I was walking along I was thinking that if only I could have become a deer, then I could have reassured it that it didn't need to be scared of me.

    Of course, if I had simply become a deer, then I would no longer be human, and not able to share my human knowledge with it. What would be needed would be for me to find a way to be fully deer and fully human.

    Sunday 16 December 2012

    If A Quarter Of Voters Are Unrepresented

    There has been a lot of attention to a ComRes poll that puts the UK Independence Party in third place on 14%, ahead of the Liberal Democrats on 9%.

    What if this were repeated at a general election?

    Rather than look at how each party's vote has changed, look at how votes have moved between parties.

    Labour would gain 105 seats from the Conservatives, 33 from the Liberal Democrats and 1 from the Greens - a net gain of 139.

    The Conservatives would gain 23 seats from the Liberal Democrats, while losing 105 to Labour and 1 to Independent Community & Health Concern - a net loss of 83.

    The Liberal Democrats would lose 33 seats to Labour, 23 to the Conservatives and 1 to Plaid Cymru - a net loss of 57.

    Plaid Cymru would gain 1 seat from the Liberal Democrats.

    Independent Community & Health Concern would gain 1 seat from the Conservatives.

    The Greens would lose 1 seat to Labour.

    This gives an election result of:

  • Labour - 397 (which will include 2 Deputy Speakers)
  • Conservatives - 223 (which will include 1 Deputy Speaker)
  • Northern Ireland parties - 18
  • Scottish National Party - 6
  • Plaid Cymru - 4
  • Independent Community & Health Concern - 1
  • The Speaker - 1
  • At the May 2010 general election, UKIP set the record for the highest share of the vote for a party which didn't win any seats. On this opinion poll it will dramatically break that record!

    And the Liberal Democrats would achieve the second highest ever share of the vote for a party which didn't win any seats.

    This would create a bizarre result. We are used to the idea that First Past The Post can leave people feeling unrepresented, but this has usually meant constituents having an MP from a party they didn't vote for, or a party getting a small number of seats compared to its share of the vote.

    But this would be different. UKIP, the Liberal Democrats and the Greens would between them command around a quarter of the vote- but not have a single MP between them.

    Among the current Cabinet, the Liberal Democrats would naturally all lose their seats. Lord President of the Council Nick Clegg would lose Sheffield Hallam to Labour; Business & Innovation Secretary Vince Cable would lose Twickenham to Labour; Energy & Climate Change Secretary Ed Davey would lose Kingston & Surbiton to the Conservatives; Scottish Secretary Michael Moore would lose Berwickshire, Roxburgh & Selkirk to Labour; and Chief Secretary to the Treasury Danny Alexander would lose Inverness, Nairn, Badenoch & Strathspey to Labour.

    The only Conservative Cabinet casualty would be Welsh Secretary David Jones, losing Clwyd West narrowly to Labour (this would be the second closest Labour/Conservative marginal).

    In Scotland, of course, Moore and Alexander would not be the only Liberal Democrats to lose their seats. Indeedn all 11 Liberal Democrats representing Scottish seats - as well as Minister for Scotland, David Mundell, who is Conservative MP for Dumfriesshire, Clydesdale & Tweeddale - would lose to Labour, leading Scotland to have 53 Labour MPs and 6 Scottish National Party ones.

    The Welsh story is a bit different. The Liberal Democrats would lose their 3 seats to different parties- Brecon & Radnorshire to the Conservatives, Cardiff Central to Labour, and Ceredigion to Plaid Cymru.

    Jones would be the most senior Welsh Conservative to be defeated, but not the only one. Aberconwy, Cardiff North, Carmarthen West & South Pembrokeshire,Preseli Pembrokeshire and Vale of Glamorgan would all fall to Labour, giving Labour 33 seats, Plaid Cymru 4 and the Conservatives 3.

    The highest Liberal Democrat share of the vote would be in Orkney & Shetland, where they would be in second place (Labour winning the seat for the first time, and the first time since the July 1945 general election that a non-Liberal candidate won) with 25.49% of the vote. The only other seat where the Liberal Democrats would get over a quarter of the votes is Westmorland & Lonsdale, where the Conservatives would win, with Labour second and the Liberal Democrats third on 25.12%.

    However, the top target seats would be:

  • Yeovil- a 2.56% swing from the Conservatives [Labour in second place in the closest Conservative/Labour marginal]
  • Norfolk North- a 2.70% swing from Labour [Conservatives in second place]
  • Westmorland & Lonsdale- a 3.05% swing from the Conservatives [Labour in second place]
  • Bath- a 3.17% swing from Labour [Conservatives in second place]
  • Twickenham - a 3.60% swing from Labour [Conservatives in second place]
  • Lewes - a 4.37% swing from the Conservatives [Labour in second place]
  • Orkney & Shetland - a 4.49% swing from Labour
  • Ceredigion - a 4.52% swing from Plaid Cymru [Labour in second place]
  • Kingston & Surbiton - a 4.73% swing from the Conservatives [Labour in second place]
  • Southport - a 4.74% swing from Labour [Conservatives in second place in the closest Labour/Conservative marginal]
  • Thornbury & Yate - a 4.80% swing from the Conservatives [Labour in second place]
  • Devon North - a 4.88% swing from the Conservatives [Labour in second place, UKIP in third place] in a seat wher the Liberal Democrats would fall from first to fourth place
  • The Liberal Democrat problem is obvious. On that opinion poll, they are looking at obliteration. Not only that, but their top targets are generally seats where they are third (or even fourth).

    UKIP would have 500 seats where they get over 10% of the vote, and in 39 of those they would get over 20%. The highest share of the vote is 23.87% in Christchurch.

    However, the targets which they would win with a swing of less than 5% are:

  • Devon North - a 4.37% swing from the Conservatives [Labour in second place]
  • Norfolk North - a 4.68% swing from Labour [Conservatives in second place, Liberal Democrats in third place]
  • Something is obvious here. It seems that UKIP target seats would also be Liberal Democrat ones, and indeed, in every seat where UKIP needs a swing of less than 10% to win, the Liberal Democrats also need a swing of less than 10%.

    While on the subject of UKIP target seats, one of the most interesting is Aberdeenshire West & Kincardine, which would be a Labour seat. UKIP would need a swing of 8.62% swing to win it from Labour, but would have to leapfrog the second-placed Conservatives, the third-placed Scottish National Party and the fourth-placed Liberal Democrats. This would be the seat with the smallest gap between the first and fifth parties.

    For the Greens, there are only 3 seats where they would get over 10% of the vote:

  • Brighton Pavilion - 33.01% of the vote, and the only seat where they need a swing of less than 10% to win (0.14% swing from Labour)
  • Norwich South - 17.67% in a seat where Labour is first and the Conservatives second [the Liberal Democrats falling from first to fourth place]
  • Cambridge - 11.13% in a seat where Labour is first, the Conservatives second, Liberal Democrats third and UKIP fourth
  • There are things to note. Firstly, this is a middle of the mid-term poll.

    Seocndly, shifts in votes are consistent, and every seat will buck the trend in one way or another (although the poll indicates a Liberal Democrat wipe-out there would be some who would buck the trend). This creates an element of volatility - a small bucking of the trend can change who the winner is, and with this repeated across loads of seats, any result is going to be a surprise.

    Thirdly, and connected with this, there will be a rise in three- or four-way marginals. And there is the potential for some seats to be ones where during the election campaign you cannot be sure which one of five parties is going to win it. The first genuine five-way marginal cannot be too far away.

    Fourthly, we should expect more seats to be won on a majority, but not a plurality, of votes.

    Fifthly, the overall result will be less proportional.

    Tuesday 11 December 2012

    2020 Vision On Electoral Reform

    One date to put in your diary now is Thursday, 7 May 2020, for something will happen that has never happened before and will not happen again until Thursday, 3 May 2040.

    The Fixed Term Parliaments Act 2011 set the date of the next general election (to the House of Commons only) to be Thursday, 7 May 2015 - thanks to the failure to introduce legislation for House of Lords reform, the Commons will be the only chamber at Westminster we'll be voting for, although I have looked at how House of Lords reform can be salvaged in line with previous proposals. And then every 5 years on the first Thursday in May, with certain exceptions.

    The Scotland Act 1998 sets the elections to the Scottish Parliament to be every 4 years on the first Thursday in May from 1999 onwards. However, the Fixed Term Parliaments Act 2011 postpones the 2015 ones to Thursday, 5 May 2016.

    The Government of Wales Act 1998 (now superseded by the Government of Wales Act 2006) sets the elections to the Welsh Assembly to be every 4 years on the first Thursday in May from 1999 onwards, just like the Scottish Parliament. And, just like the Scottish Parliament, the Fixed Term Parliaments Act 2011 postpones the 2015 ones to Thursday, 5 May 2016.

    The Northern Ireland Act 1998 sets - yes you've guessed it - elections to the Northern Ireland Assembly to be every 4 years on the first Thursday in May, with the first elections being Thursday, 25 June 1998, and the cycle would then be 2002. 2006, 2010, 2014 etc.. However, the Northern Ireland Act 2000 allowed the Northern Ireland Secretary to suspend the Assembly, and under the terms of the Northern Ireland Assembly Elections Act 2003 the next election was on Wednesday, 26 November 2003. The Northern Ireland (St Andrews Agreement) Act 2006 set the next elections to be Wednesday, 7 March 2007, and since then the 4-year cycle in the Northern Ireland Act 1998 has re-commenced. That put the last elections on Thursday, 5 May 2011, coinciding with the Scottish and Welsh ones, so we would have expected the three non-English devolved legislatures to be elected at the same time.

    But the Fixed Term Parliaments Act 2011 did not apply to the Northern Ireland Assembly, so in May 2015 the good people of Northern Ireland will be voting (using First Past The Post) for the House of Commons and (using the Single Transferable Vote) for the Northern Ireland Assembly, with the following Stormont elections being on Thursday, 2 May 2019.

    The Greater London Authority Act 1999 set up the two parts of the Greater London Authority- the Mayor of London and the Greater London Assembly. It set the first election to be Thursday, 4 May 2000, and then the first Thursday in May every 4 years. So, the next one occurs at the same time as the Scottish Parliament and Welsh Assembly.

    And this brings us to the Police & Crime Commissioners across England (outside Greater London) and Wales, elected under the Police Reform & Social Responsibility Act 2011, with the first elections having been on Thursday, 15 November 2012, and then on the first Thursday in May starting with 2016.

    So, in May 2016 there will be 4 simultaneous elections covering all of Great Britain:

  • Police & Crime Commissioners
  • Greater London Authority
  • Scottish Parliament
  • Welsh Assembly
  • And these will be on the same 4-year cycle.

    Not only that, but:

  • 89% of the United Kingdom electorate would be using the Supplementary Vote (for the Mayor in Greater London; for the Police & Crime Commissioners in the rest of England and in Wales)
  • 25% of the United Kingdom electorate would be using the Additional Members System (for the Scottish Parliament, Greater London Assembly and the Welsh Assembly)
  • Now, if you have a 4-year cycle and a 5-year cycle, then every 20 years they coincide. And, when these are up for re-election in 2020, it will be alongside elections to the House of Commons.

    This will be a general election where the whole electorate of Great Britain will not only be voting (using FPTP) for their MPs, but alongside that will be voting using SV and/or AMS.

    As people enter the polling booths in May 2020, FPTP won't be the way we do elections here. It will be just one of the ways we do elections here.

    Saturday 1 December 2012

    Notional Result- Brighton Pavilion

    I have been having a look at notional results and in my last post considered how they could be unreliable, and suggested one way forward, by projecting the May 2005 general election results onto the proposed constituencies for the May 2015 general election.

    I touched on the issues around the proposed Brighton Pavilion

    However, start by turning to the west of Brighton Pavilion, to the proposed Hove seat, which is actually the current Hove with the Brighton & Hove City Council ward of Regency picked up from the current Brighton Pavilion.

    Colin Rallings and Michael Thrasher drew up the notional results of the 2005 general election based on the real 2010 constituencies. For Hove they got:

  • Labour- 16,829
  • Conservatives- 16,381
  • Liberal Democrats- 8,077
  • Greens- 2,593
  • Others- 1,072
  • and for Brighton Pavilion they got:

  • Labour- 16,283
  • Conservatives- 10,416
  • Greens- 9,804
  • Liberal Democrats- 7,174
  • Others- 1,064
  • The real results for Brighton Pavilion in the May 2010 general election are:

  • Greens- 16,238
  • Labour- 14,986
  • Conservatives- 12,275
  • Liberal Democrats- 7,159
  • Others- 1,176
  • The real results for Hove in the 2010 general election were actually:

  • Conservatives- 18,294
  • Labour- 16,462
  • Liberal Democrats- 11,240
  • Greens- 2,568
  • Others- 1,291
  • The notional reults for the proposed Hove are:

  • Conservatives- 19,213
  • Labour- 17,934
  • Liberal Democrats- 12,024
  • Greens- 3,712
  • Others- 1,895
  • If we take the difference, we get:

  • Labour- 1,472
  • Greens- 1,114
  • Conservatives- 919
  • Liberal Democrats- 784
  • Others- 604
  • This is, as far as we can calculate it, how the good people of Regency voted in 2010. We then divide each party's votes by the number of votes they got across Brighton Pavilion as a whole and multiply that by the notional 2005 result calculated by Rallings & Thrasher, to work out how Regency voted in 2005:

  • Labour: 16,283 x (1,472 / 14,986) = 1,599
  • Liberal Democrats: 7,174 x (784 / 7,159) = 786
  • Conservatives: 10,416 x (919 / 12,275) = 780
  • Greens: 9,804 x (1,114 / 16,238) = 673
  • Others: 1,064 x (604 / 1,176) = 546
  • We now add those to the notional 2005 results that Rallings & Thrasher calculated:

  • Labour: 16,829 + 1,599 = 18,428 (37.35%)
  • Conservatives: 16,381 + 780 = 17,161 (34.78%)
  • Liberal Democrats: 8,077 + 786 = 8,863 (17.96%)
  • Greens: 2,593 + 673 = 3,266 (6.62%)
  • Others: 1,072 + 546 = 1,618 (3.29%)
  • So, this is the notional 2005 Hove result based on the 2015 boundaries. If the 2010 election had been fought on the 2015 boundaries, then these are the results that Labour and the Conservatives would have been using to indicate it was a 2-horse race.

    The next step is to take how we have calculated the Regency ward to have voted in 2010 and subtract that from the real 2010 result for Brighton Pavilion:

  • Greens: 16,238 - 1,114 = 15,124
  • Labour: 14,986 - 1,472 = 13,514
  • Conservatives: 12,275 - 919 = 11,356
  • Liberal Democrats: 7,159 - 784 = 6,375
  • Others: 1,176 - 604 = 572
  • These figures are how that part of the current Brighton Pavilion that remains in the proposed Brighton Pavilion voted in 2010.

    The motional results we have for Brighton Pavilion are:

  • Labour: 16,718
  • Greens: 15,351
  • Conservatives: 13,266
  • Liberal Democrats: 7,751
  • Others: 1,350
  • If we take the difference, we get:

  • Labour: 3,204
  • Conservatives: 1,910
  • Liberal Democrats: 1,376
  • Greens: 227
  • Others: 778
  • What's causing the difference? It's because the proposed Brighton Pavilion takes in a Brighton & Hove City Council ward - Moulsecoomb & Bevendean - from the current Brighton Kemptown. The low Green vote in that ward could mean that the voters don't give a stuff about polar bears. Or it might mean they were living in a constituency which the Greens were not focussing on as they concentrated on (successfully) getting Caroline Lucas elected as MP for Brighton Pavilion.

    And if the 2010 general election had been fought on the proposed 2015 boundaries then this would be a ward that the Greens would have extensively campaigned in to get Lucas elected.

    Rallings & Thrasher calculated the notional 2005 results for Brighton Kemptown to be:

  • Labour- 14,939
  • Conservatives- 13,086
  • Liberal Democrats- 6,482
  • Greens- 2,506
  • Others- 1,325
  • The real 2010 results were:

  • Conservatives- 16,217
  • Labour- 14,889
  • Liberal Democrats- 7,691
  • Greens- 2,330
  • Others- 1,578
  • As we have done before, we can take the 2010 results for Moulsecoomb & Bevendean and divide them by the 2010 results across the whole of Brighton Kemptown, and then multiply them by the notional 2005 results that Rallings & Thrasher calculated. What we then get is how the good people of Moulsecoomb & Bevendean voted at the 2005 election:

  • Labour: 14,939 x (3,204 / 14,889) = 3,215
  • Conservatives: 13,086 x (1,910 / 16,217) = 1,541
  • Liberal Democrats: 6,482 x (1,376 / 7,691) = 1,160
  • Greens: 2,506 x (227 / 2,330) = 244
  • Others: 1,325 x (778 / 1,578) = 653
  • The final step in this lengthy calculation is to take the notional 2005 results for Brighton Pavilion that Rallings & Thrasher calculated, subtract the 2005 Regency vote and add in the 2005 Moulsecoomb & Bevendean vote:

  • Labour: 16,283 - 1,599 + 3,215 = 17,899 (37.95%)
  • Conservatives: 10,416 - 780 + 1,541 = 11,177 (23.70%)
  • Greens: 9,804 - 673 + 244 = 9,375 (19.87%)
  • Liberal Democrats: 7,174 - 786 + 1,160 = 7,548 (16.00%)
  • Others: 1,064 - 546 + 653 = 1,171 (2.48%)
  • This is the notional 2005 result on the 2015 boundaries. If the 2010 general election had been fought on the 2015 boundaries then these are the results the parties would have been using as the results of the last election. And it gives the Greens a higher mountain to climb.

    At thr 2010 general election, the Greens increased their share of the vote by 9.42%, the Labour vote decreased by 7.48%, the Conservative vote increased by 0.40% and the Liberal Democrat vote decreased by 2.22%. If we assume that the notional Brighton Pavilion would have seen the same shift in vote, then we get the Greens on 29.29%, but Labour winning on 30.47%.