An anatomy of Labour’s defeat

Now that we are a couple of weeks past the initial shock it is time to start looking at the detail of the General Election results, as in the anatomy of our defeat we will find useful pointers about how to recover.

Here are some initial useful snippets to consider:

Geography

The House of Commons library had published its usual excellent statistical breakdown by region:

Here is the change in Labour’s vote share by region:

London                                  +7.1% (3.4% swing from Con to Lab)

North West                            +5.2% (2.9% swing from Con to Lab)

Yorkshire & Humber            +4.8% (2.5% swing from Con to Lab)

North East                             +3.3% (0.9% swing from Con to Lab)

East                                        +2.4% (0.3% swing from Con to Lab)

West Midlands                      +2.3% (0.1% swing from Con to Lab)

South West                           +2.3% (0.7% swing from Lab to Con)

South East                            +2.1% (0.3% swing from Con to Lab)

East Midlands                       +1.9% (0.2% swing from Lab to Con)

Wales                                     +0.6% (0.3% swing from Lab to Con)

Scotland                                -17.7% (23.9% swing from Lab to SNP)

The pattern here seems to be that in the most urban regions the Lib Dem collapse and our political positioning since 2010 helped Labour, in the most rural it helped the Tories, and where there was not much Lib Dem vote to start with the rise of UKIP hurt Labour more than it did in other regions. We seem to have a Scottish problem, a Midlands problem, and a Welsh problem to add to the “Southern Discomfort” we had from 1979-1992, but at least we no longer have the 1987 “London Effect”.

Change since 2005

I also thought it was worth looking not just at the comparison with an election we also lost (2010) but at the comparison by region with the last election we won (2005).

Here is the change in Labour’s vote share by region since Blair’s final victory:

London                                  +4.8%

North West                            -0.5%

Yorkshire & Humber            -4.5%

South West                           -5.1%

Wales                                     -5.8%

West Midlands                      -5.8%

North East                             -6.0%

South East                            -6.1%

East Midlands                       -7.4%

East                                        -7.8%

Scotland                                -14.6%

And here is the impact since 2005 in terms of seats:

London                                  +1

North East                             -2

Wales                                     -4

East                                        -9

South West                           -9

North West                            -10

East Midlands                       -11

Yorkshire & Humber            -11

West Midlands                      -14

South East                            -15

Scotland                                -39

What leaps out is that in the North West we are on almost the same vote share as we won the 2005 General Election with, but winning 10 fewer seats. I.e. we have swapped a load of ex-Lib Dem votes in seats where it doesn’t affect the outcome for Con/Lab swing voters in the marginal seats that really matter. And similarly in London whilst the overall pattern is far better than in the rest of the country, the 4.8% increase in our vote share has been at the expense of the Lib Dems, mainly in seats we already hold, and has only delivered us a net one additional seat. 

Demography 

Ipsos Mori has published tables showing the demography of the support for each party, helpfully posted here by Mike Smithson:

These show that in the context of an overall small swing of 0.35% from Con to Lab, the following happened:

• People aged 18-24 (only 43% of whom turn out) swung to Labour by 7.5% but the crucial over 65 demographic (78% of whom turn out) swung to the Tories by 5.5%. Beyond our stance on the NHS I don’t think we had a clear offer to older voters. Labour almost came third behind UKIP among OAPs.
• We got a small swing to Labour amongst working class DE social group voters and to a greater extent the skilled working class C2s but the Tories got a swing in their direction among high turnout professional ABs and lower middle class C1s. Again we didn’t really have a clear offer to these groups.
• The clarity of our policy stance on private renting seems to have helped lift our vote among voters with this kind of housing tenure by 10% but we actually went backwards among home owners (perhaps because of the Tory pledge on inheritance tax?). Again our vote share went up in the groups of people with the lowest propensity to actually bother voting.
• There was a small net movement towards the Tories among BAME voters – the Tories have made no secret of their efforts to target the increasingly suburban and prosperous Hindu and Sikh communities.

Marginality

Lewis Baston has plotted the Tory lead over Labour compared to 2010 in every seat in England and Wales.

This seems to prove that there was a swing to Labour from the Tories in the seats where it didn’t matter (safe Labour or safe Tory seats) but the one group of seats that mainly swung to the Tories were Tory-held marginals, i.e. the places where it mattered most. This is partly explained by the double-incumbency effect (new Tory MPs with five years of casework and working for the area replacing Labour MPs who in 2010 had had a similar incumbency bonus) but the scale of this pattern suggests something more i.e. even though we had a very vigorous key seat campaign the Tories out-targeted and resourced us in those seats (presumably with paid-for calls, advertising and direct mail to compensate for Labour’s bigger activist base) and had a policy offer that resonated with swing voters in swing seats. 

Best and worst performances

This Twitter user has been helpfully compiling ranked tables of individual constituency results.

Here is his ranked table of the change in Labour’s vote share: first part and second part.

Our 20 best performances can be easily summarised as 13 seats where there is a very large Muslim vote some of whom would have gone Lib Dem or Respect in 2005 and 2010 because of the Iraq War and have now swung back to Labour, and 7 seats that are in big cities and had previously had a very strong Lib Dem presence.

As might be expected, the 20 worst performances are all in Scotland, so it is worth looking at the 20 worst changes in Labour’s vote share in England and Wales:

Burton

Derbyshire South

Morecambe & Lunesdale

Rhondda

Kettering

Cynon Valley

Sittingbourne & Sheppey

Kingswood

Basildon South & Thurrock East

Barrow & Furness

Wellingborough

Brigg & Goole

Dorset South

Tamworth

Somerset North East

Hartlepool

Thanet South

Rochester & Strood

Chatham & Aylesford

Clacton

16 of these are seats that had been Labour-held marginals gained from the Tories in 1997 or 2001 and lost in 2010.

Six are in areas of Kent or Essex where UKIP has clearly made big inroads into the Labour vote.

I would urge everyone considering the way forward for Labour to study the results in as much detail as possible and draw objective lessons rather than looking for patterns that confirm our preconceived views about how we might win in future.

And finally, thanks also to @AndyJSajs here is Labour’s target list for forming an overall majority in 2020, which requires 94 gains and a swing of 8.7% even before any adverse boundary changes. It makes sobering reading