Where Labour can gain seats from the SNP

The other day I blogged about the decline of the SNP and I’ve done a little more analysis to see where Labour can make inroads.

The two graphs below show quite clearly that there is fertile ground for Labour in Scotland, but there are also some possible gains for the Tories.

The first graph shows the SNP majorities in ascending order, grouped by second party.  Many of these majorities are very slim and the Tories are second in only eight of the 33 SNP held seats, with Labour second in the other 25.

The second graph plots the number of votes for each party in the general election and shows that in some seats the second and third parties are very close.

All in all, it shows Scotland is very much in play and on the current results it looks hopeful for Labour.

Edited to add: one of the commenters on the blog and Glen O’Hara spotted that there were some seats missing.  I checked my work and saw I’d got a gremlin in the system that filtered a few seats out of the graph. The correct updated graphs are now displayed. Thank to those who have helped me make this accurate.

09 SNP majorities

09 SNP majorities by party

The decline of the SNP and why Scotland matters

I blogged here the other day, comparing the Labour and Conservative “surges” in the recent general election. I’ve now had the chance to play with the data some more,  and Scotland has jumped out at me.

First, see below for my visualisation of the 2015 and 2017 turnout broken down in GB regions. I realise that Scotland and Wales are actually a nations and not a regions, but for the purpose of this post and to make it easier to write, I will only refer to the term “region” to mean Scotland and Wales too – but the unwritten part is that I accept they are both nations.

I’m quite pleased with the visualisation, as although it takes a little effort to read, it does give a lot of info in a small space. Points to note:

  1. Each constituency has a vertical dot plot of (red) 2015 and (green) 2017 turnout.
  2. In each region the constituency (green) 2017 turnout dots are in descending order.
  3. Each region has a red and green line which is the average regional turnout for 2015 and 2017 respectively.
  4. The regions are ordered in descending order of 2017 turnout.
  5. Because each constituency is shown as a vertical dot plot, the width of each region is proportional to the number of seats in it. So you can compare region sizes.

Scotland has gone from having the highest turnout in 2015, to the third lowest turnout in 2017, and more importantly it is the only region that had a fall in turnout.

06 Turnout by Region

So turnout in Scotland fell. But even more importantly, the SNP’s fortunes declined from their quite stunning zenith in 2015. They’ve fallen back down to earth with a bump, losing 21 seats.

The horizontal dot plot below shows the change in 2015 (light grey) majority and 2017 (black) majority. It is descending order of 2017 majority. As well as losing 21 seats, they now have several hyper marginals, and they now have no seats with a majority of more than 7,000. (Data visualisation point: this type of plot is often referred to as a barbell or dumbbell plot).

Given the fall in majorities seen in 2017 it essentially means every SNP is now a marginal in play at the next election. This is something I missed, and has been rather under reported by the political commentators. Given how close this election was overall, Scotland could be the deciding factor in the next general election.

06 SNP majorities


Comparing the Labour and Conservative “surges” 

There’s been a lot of talk about the “Labour surge” and I blogged about it yesterday. In an attempt to avoid a narrow view of the election I decided to look at this from the other side: the “Conservative surge”. 

Yes, there was one, and it has been lost in the chatter about the very impressive Labour surge.

First some facts. The Labour vote share went up an impressive 9.5 points. But what many people, especially Labour supporters missed is that the Conservative vote share went up a not insignificant 5.5 points. 

The most interesting thing to do is to compare the surges for each party. To do this I’ve produced two graphs of the top 50 surge seats for each party (I modified the Labour surge graph I did yesterday). Note both are to the same scale to make comparisons more meaningful. They are shown below.

What do they tell us? 

The top 50 Labour surge seats had a total of 579,000 extra votes whereas for the Conservatives the top 50 surge seats generated 421,000 extra votes which is more than a quarter fewer than the Labour surge generated.

Despite generating fewer surge votes, the Conservatives made them count more as they produced more gains for the party and they ate into more Labour majorities. Conversely, the Labour surge votes predominantly increased existing Labour majorities, made fewer gains and ate into fewer Conservative seats. 

With fewer “surge” votes the Conservatives have made them achieve more. 

The Conservative surge graph should be up on the wall at Labour high command as it shows the seats where the Conservatives have either won seats from Labour or made a major dent in the majorities. 

The common factor with these seats is that they are nearly all “post industrial communities”. No doubt the Labour message enthused young people and those living in metropolitan cities but Labour need to ensure that they can enthuse the post industrial communities too. This incidentally was point made by John Mann MP yesterday with his “Bolsover question.”

Some analysis of the “Labour surge” 

There’s been a lot of talk about the “Labour surge” so I’ve had a dig in the numbers to see what I can find. This is just a quick analysis and if I get time I will post more. In the meantime here’s what I’ve found. 

Before we look at the Labour surge I want to look at the frequency distribution of the majorities for the two main parties in 2015 and 2017.

The two graphs below compare the 2015 and 2017 majorities. Note, to enable like-for-like comparisons of the trends the graphs only show the seats that were held in 2017 – they do not include the seats that were won or lost by each party in 2017. The graph uses histogram bins of 2,500.

Three things strike me:

1. In 2017 Labour increased the majorities of safe seats and made them even safer. This can be seen by the dark red line shifting significantly to the right. Much of the Labour surge means Labour piling up extra votes where they are of little use other than to flatter the MP.

2. In 2017 the number of Labour seats with a majority of 2,500 or less decreased. These are what would be seen as classic marginal seats. Form a Labour perspective this is good as they will have fewer marginals to defend.

3. Conversely for the Conservatives in 2017 the number of seats with majorities of of 2,500 or less increased. This means they have more marginals to defend. Again this is good for Labour.

Below is the 2017 comparative frequency distribution of the seats held by each party. You can see Labour has far more ultra-safe seats where the votes have piled up.

Finally, turning to the biggest part of the surge, the graph below shows the 46 seats where Labour added on more than 10,000 extra votes. The colour coding shows which seats were Labour holds, Conservative holds and Labour gains. It shows quite convincingly that the biggest part of the Labour surge only (where Labour put on 10,000 or more votes) only helped Labour gain four seats out of the 46 seat where the biggest surges happened.

Any future “mass Labour surge” needs to be make a bigger difference to maximise the seats gained rather than making already big majorities bigger.

Pie chart to show how Brexit will weaken our trading power

Those in favour of hard Brexit, like Secretary of State for International Trade, Dr Liam Fox seem to think once we leave the EU we needn’t bother about trade with the EU so much, as we will have the freedom to trade with the rest of the world in a free and easy manner.

After taking a look at GDP data from the World Bank here, I don’t share the good doctor’s prognosis.

The US economy is currently the largest economy on the planet with 24.7 per cent of global GDP. The EU (including the UK) is second largest with 22.4 per cent of global GDP.

Currently, as a member of the EU, the UK is part of the second largest global trading block. That makes the EU (and by extension the UK) a very lucrative trading partner.

But when we leave the EU, the EU will still be the second largest trading block (albeit with a reduced share of global GDP down to 18.5 per cent). The UK will fifth largest trading block with 3.9 per cent of global GDP.

Post Brexit the EU are not going to give the UK an easy trade deal, and the US and China will still see the EU as their primary trading partners rather than the UK.

Now, imagine you are India, Brazil, Canada, or any of the other 160 countries that are smaller than the UK but still make up one-third of world’s GDP. Now ask yourself would you rather focus your time on getting good trade deals with the US, the EU and China, or would you want to get spend your time on thrashing out a trade deal with the UK?

I’m generally not a fan of pie charts for the reasons given here, but in this case I think a pie chart illustrates my point well – see below.

The UK economy is significantly smaller than the big three trading blocks and is broadly comparable to Canada, Brazil and India. We will not be one of big players.


Those Labour Purge Missing Ballot Papers – the real numbers

Yesterday I posted about the “Labour Purge” and the suspensions that had been carried out by the NEC. The (relatively) very low figures surprised me and they surprised a lot of other people. Two quite reasonable questions were asked by a few people.

First, my data source was questioned. So I will restate my figures came from Labour Party NEC member Christine Shawcroft. She is resolutely on the left of the party and has been so for many years. She was on the Centre Left Grassroots Alliance (CLGA) slate in the recent NEC elections. The CLGA were very clearly pro-Corbyn, and Christine continues to be a vocal supporter of Corbyn. Christine has no motivation whatsoever to publish fake figures to downplay the number of suspensions. It simply is not reasonable to seek to discredit these figures because of her political motivation.

The second question that was raised was “what about the people who had missing ballot papers?” That is a very good question and I will turn to that now.

In a Facebook post here Christine says “The figures show that 5,000 people asked for replacement ballots (ie they hadn’t got the first one for some reason).” According to a post on Richard Burgon’s Facebook page there were 654,006 ballot papers distributed. So the 5,000 ballot papers amounts to 0.8 per cent of all papers issued. And remember nearly all of these people will have had a new paper reissued.

So, 3,963 people (or 0.6 per cent of the selectorate) were suspended and  5,000 people (or 0.8 per cent of the selectorate) did not get a ballot paper initially. Has there been a deliberate and systematic purge by the NEC that denied masses of Corbyn supporters their vote? I’d suggest the evidence says no because:

  1. Of the 3,963 people (or 0.6 per cent of the selectorate) were suspended we know many of them will have been for valid reasons. You only have to look at the pro-Corbyn Labour Abuse and anti-Corbyn Gentler Politics twitter accounts to see the reports of abuse. Abusive behaviour is definitely happening. Also some people have been found to be members of other parties. So the number of “wrong” suspensions is going to be fewer than 3,963.
  2. Some of the people being suspended will invariably have been Smith supporters.
  3. Of the 5,000 people who reported that they had not got a ballot paper nearly all will have been reissued with a new ballot. Of course, some may not have received the replacement ballot paper, but the actual numbers who, in the end, did not get a ballot paper is much, much less than 5,000.
  4. Some of the people not getting their ballot papers will invariably have been Smith supporters
  5. Even if you make the highly implausible suggestion that all suspensions were “deliberate and wrong suspensions of Corbyn supporters” and all 5,000 missing ballot papers never got replaced and they were all “deliberate exclusions of Corbyn supporters” it amounts to 8,963 people or 1.4 per cent of the selectorate. It will have had negligible impact on the result.

Given the huge amount of outrage on social media about the NEC’s wilful and deliberate exclusion of Corbyn supporters through suspensions and missing ballot papers I was genuinely surprised by these figures, and maybe you are too. But if you are want to dispute them you either need to provide your own figures (and data source) or explain why an undoubted prominent Corbyn supporter would supply the wrong figures. If you answer either of these two questions then please do comment as if my conclusions are wrong I’m happy to be corrected.

Although I’m not really a fan of pie charts (see here for why) I think I pie chart illustrates my point well (I will admit they do have their  very occasionally – usually when there are only two or three categories).


Those Labour Purge Suspensions – the real numbers

Jeremy Corbyn won the Labour leadership election yesterday with 62 per cent of the vote. Congratulations to him on a convincing win. On social media there has been a lot of chatter about how that his margin of victory would have been much larger if the deliberate purge of his supporters had not happened.

I’ve seen some very high figures bandied about, some vent suggesting well over 170,000 people were prevented from voting. This seems like a very high figure so I thought I’d dig around to see what numbers I could find. I am not going to get into the rights and wrongs of decisions of the NEC to suspend or expel people – I solely want to make an objective assessment of the scale of those prevented from voting.

First, I want to start by quoting my sources for this analysis. Given that it is Corbynites who are claiming there has been a massive purge I will quote data I have sourced by two prominent Corbyn supporters. One is NEC member Christine Shawcroft who has a long track record of being on the left of the party and she is a prominent Corbyn supporter. Using Christine as a source has the added benefit because as an NEC member she has access to data other people may not get and she has kindly published it on Facebook. The other source is MP Richard Burgon who is in the Shadow Cabinet and other prominent Corbyn supporter.

Now let’s look at the suspensions. For reasons of brevity when I refer to “suspensions” I actually mean “suspensions and expulsions.”

Suspension from the party because of complaints about abusive behaviour, membership of a proscribed organisation, support for another party etc. A final reason for suspension is (and I paraphrase many Corbynites) “suspension for some ridiculous reason to deliberately purge Corbyn supporters.”

According to a post on Richard Burgon’s Facebook page there were 654,006 ballot papers distributed.  And according to a post on Christine Shawcroft’s Facebook page 11,250 complaints were put in. Over half of them were discounted. The 11,250 complaints represents 1.7 per cent of the selectorate for this election.

As Christine herself notes “If I was trying to conduct a purge of Party members or supporters, I wouldn’t dismiss half the complaints.” In fact two out of every three complaints were dismissed.

She goes on to say “of all the complaints, 3,963 led to suspension and expulsion.”

The total number of suspensions and expulsions amounts to 0.6 per cent of the selectorate for this election.

If some of the reports of individual accounts of suspensions are correct then it certainly looks like the wrong decision was made and they are likely to be successfully appealed. But there can be no doubt many of them will be reasonable. You only have to look at the pro-Corbyn Labour Abuse and anti-Corbyn Gentler Politics twitter accounts to see the reports of abuse. Abusive behaviour is definitely happening.

There is no way determining if the suspensions were disproportionately aimed at Corbyn supporters or how many will be successfully appealed but we can say for sure they only amount to 0.6 per cent of the selectorate and not all of this 0.6 per cent were Corbyn supporters, and not all of them will be successfully appealed.

Given this evidence, the idea that there has been a systematic purge of the selectorate using suspensions that disproportionately impacted on the result is plain wrong. So, if after reading this blog, you are still promoting this idea you are wilfully  spreading disinformation and being divisive.

Although I’m not really a fan of pie charts (see here for why) I think I pie chart illustrates my point well (I will admit they do have their  very occasionally – usually when there are only two categories).


Match outcomes for Premier League referees for 2010/11 – 2015/16

Referees outcomes

The graph above shows the match outcomes (home win, draw and away win) as a percentage of the total number of Premier League matches officiated by that ref. The graph is for every ref in the seasons from 2010/10 to 201o/15/16. That is 2,660 matches. I got the data from here.

The dashed lines show the home win, draw and away win percentages across the 2,660 match sample so you can compare each ref with the average across the time period.

The number in brackets after each ref’s name is the number of games tat they have refereed.  I’d say that it’s only worth trying to draw firm conclusions from those refs who have refereed more than 50 matches as that gives enough data to smooth out any outliers.

Those refs near the top who have officiated at over 100 matches seem quite consistent in their results profile, apart from:

  • A Taylor who seems to give significantly fewer home wins, and signficantly more away wins.
  • L Mason who gives more away wins and fewer draws than average.
  • N Swarbrick who appears to give a lot fewer home wins than average.

To really understand how significant these deviations are I’d need to do a standard deviation calculation weighted by the number of matches officiated at. But I didn’t have time for that.

I was also surprised to see that more matches end up as an away win than as a draw. I thought it would be the other way around.

Why have I done this? Mainly because I can to hone my Excel skills. Also, if you are prone to the odd wager on the football this is information that might help you.

An Excel nerd point: the use of the horizontal dot plot required some significant wrangling as Excel does not have this type of chart as an option. Look here for the instructions on how to do it.

Fact checking Paul Mason

Today on Radio 4 Paul Mason claimed, that if there had been a snap general election after the EU referendum, Jeremy Corbyn could have been in a position to form a minority government because it was his belief that:

Eighteen of the top twenty marginals in England and Wales could have been won by Labour if one to three thousand Green voters would come back to Labour

You can hear his very words on the Daily Mirror website here. When I heard his claim it startled me as I very distinctly remembered posting a while back that there were ten seats where the Green vote was bigger than the Tory majority. So I thought I’d do a proper fact check on Paul’s claim.

I make no comment on his claim about the motivations of the resignations of the shadow cabinet ,when he says they “feared” it was their “last chance” to unseat Jeremy Corbyn before a “winnable” general election.  I am solely fact checking the claim that 18 of the top 20 marginals  made in the quote above.

Below is a list of the top 25 marginal seats where Labour came second in the 2015 general election. It is in order of smallest majority to overturn. The seats in red are the ones where the Green vote is more than the majority.

Pau Mason top 20 odd marginals

Lets fact check Paul’s claim…

First, 0f the top 20 marginals only 10 of them have a Green vote larger than the majority of the winner. So his claim is wrong.

Second, of all the seats in England and Wales there are only 11 that could be won if all the Green voters switched to Labour. So, even if you gave Paul the benefit of the doubt about the top 20 seats, even looking at all seats, his claim that there are 18 seats that could be won to Labour by Green voters switching still does not stack up.

So, Paul has been fact checked and sadly his claim has been found to be wrong. I don’t know if he misspoke, or just misunderstood the numbers. If you want to take a look at my data go here.

Edited to add:

With Sinn Fein not taking their 4 seats in Westminster, 324 seats are required for a working majority. The Tories got 331.

If those 11 seats went Labour of the seats above were won by Labour the Tories would have 321 seats (remember one of 11 the seats is a Lib Dem seats).

This would mean the Tories would not have a majority in Westminster but remember there are eight DUP MPs who would almost certainly vote with the Tories as would the one UKIP MP (who was a Tory). So the Tories could easily form a government.