Everything we can learn about the Labour Party from 2020 NEC results

NEC elections are very useful for giving us accurate membership numbers, as under the rules only paid up members get to vote. NEC election electorate numbers give us the most clear picture of committed membership at each point they are held. Recent numbers are as follows:

• July 2017 – 538,606
• November 2017 – 525,779
• June 2018 – 506,320
• November 2019 – 430,359
• January 2020 – 552,835
• August 2020 – 495,961

In publishing these numbers, I can imagine someone will say “look, paid membership dropped 108,000 under Jeremy Corbyn between July 2017 and Nov 2019” and it will no doubt lead to some online argument. My response would be that membership totals rise and fall, and we should all just get used to that. They rose by 140,000 under Tony Blair, then fell by 220,000 in the aftermath of the Iraq War. At the same time, paid-up membership rose 350,000 under Jeremy Corbyn, then fell from their peak by around 120,000 to 430,000 paid up by the 2019 general election. This is inevitable as a consequence of decisions they as leaders take, which will surely not please various members all the time.

In the case of the latter fall, I think the numbers do explain the Brexit policy change debate in 2018/19. If you read Left Out and Owen Jones books, a lot of the focus will be on the action of senior shadow cabinet individuals from all wings of the party. However when you consider together a 2017 to 2019 decline in membership by 108,000 along with polling showing 79% of the remaining members backing a People’s Vote whilst 90% of members voted Remain along with 66% of Momentum members, the numbers do indicate that a change of policy was going to be inevitable eventually. This was especially the case when the only internal countervailing number within the party to fight for the 2017 policy was the subsequent polling that showed Unite members voted 39% Tory in 2019.

This should not be a surprise bearing in mind that Unite’s executive council was worrying about member voting trends in 2013 and a big chunk of the union is not just the historically Labour-voting TGWU but also former unions AEEU, EETPU, TASS and ASTMS, all of which would have had very large Tory votes in the 1980s. When you see that Tory vote share, you can totally understand why Unite’s leadership was prominent in leading the defence of the 2017 Brexit policy to try to stop that Tory vote in their union rising.

But taking all the numbers together, you might also see why it would have likely been overwhelmed in the eyes of the various parts of the party leadership described in those two books, committed as they were historically to a “member-led party” so by temperament responsive to those changing numbers. All that was even before we got to external factors such as the Euro elections results and third quarter 2019 by-election results, which would have also worried many MPs on top of those numbers above. Perhaps if we express this sort of debate in numbers, we might even get a better understanding from both sides as to the inevitability of why things developed as they did.

We will now have a strong debate over the drop in paid-up member numbers by 57,000 in the first six months of the current leadership, and so we should. I cover membership composition in the CLP vote analysis below. Since I have shown membership does rise and fall under various leaders, perhaps we should all seek more honesty and accuracy when conducting this debate. My consistent view over the years is that the party should publish quarterly membership numbers, including paid-up as well as part-arrears/arrears and pre-lapse numbers so there is no confusion as to what we are talking about.

Electorate and voting

In the first day after the NEC elections, there was a claim that “6,000 members were excluded” from the CLP vote. There are good grounds for saying this is a misunderstanding of blank ballots and not evidence of any vote rigging. Blank ballots happen when a person votes in one ballot and not another, e.g. a disabled member in Wales voting in the Welsh and disabled rep ballots but not for NEC teasurer or the CLP section. Bearing in mind over 8,000 members voted in the Young Labour election, it would not have surprised me if many voted in that one while being less interested in some of the other ballots. There are other indications that most of the “Invalid and Blank” votes were actually blank.

• If you look at the same report with the 6,000 number of Invalid and Blank ballots for the CLP section, it also shows there were 26,000 Invalid and Blank for the treasurer election. Bearing in mind how one sided a result that was (74% to 8%), is anyone really going to argue that more members were excluded from the treasurer ballot than the CLP ballot?
• If you look at the 2020 NEC by-elections, it showed the split in Invalid and Blank ballots as 214 invalid and 96,291 blank, showing what one might expect in this recent ballot.
• However, let’s not even just trust the Labour Party. Civica Election Services (CES) has its own reputation to guard, as it conduct elections for most Labour affiliates including Unite. How did the recent Unite EC elections Invalid and Blank ballots split? The results from Unite’s own election conducted by the same organisation again show most ballots are blank rather than invalid
• Have any of the campaigns, such as Grassroots Voice, submitted a complaint? I am not aware of any complaints over excluded voters a week after the elections.

In the unlikely event that a majority of the 6,000 were invalid and not blank, the still small numbers would not have changed the political composition of the NEC CLP section. Assuming any invalid votes were evenly nationally spread, it would have been unlikely to have led to Ann Henderson overtaking Mish Rahman – both from the same slate – for the final place, bearing in mind his 802-vote lead over her.

The one place the Invalid and Blank vote might have theoretically made a difference was in Wales, where the 309 Invalid and Blank votes exceeded the 272 majority by 37 votes. This has even received some news coverage in Wales. However, it would:

• Require over 88% of all the votes to be Invalid rather than Blank, which certainly goes against the evidence of the numbers above. Whilst the Welsh result only says “Invalid” votes, it was the only result that indicates this – yet we know from the Welsh Young Labour results that there were some Blank ballots recorded and those Welsh Young Labour voters were a subset of the wider Welsh member vote.
• Require only 37 members in Wales (0.35%) – a much lower figure than any other election – to have decided not to vote in the Welsh election. This seems very low, when for example the 354 young members voting in the Young Labour election in Wales may have not cast a vote in all elections. Indeed 5% of them voting look as if they did not vote for their own Welsh Young Labour rep, so it seems unlikely the number would drop so low compared to the Young Labour Invalid and Blank numbers in an election that nobody is disputing.
• After both the first two points, then require enough Invalid votes to have some reason for validity. The party emailed all candidates to say all members who voted and resigned (a perfectly understandable and even logical behaviour) had their votes accepted and there was a much smaller number of people who had resigned then tried to vote (pretty strange behaviour in comparison) and these would have been declared invalid. This seems reasonable invalidity in view of the actual behaviour, which is analogous to attempting to cast a vote in a general election when you have moved away from the area and do not have a valid vote there any more.
• Require there to be an objection from the losing side. In this case, losing candidate Mick Antoniw says he accepts the result.

Thus it would be useful to have the Invalid and Blank ballots separated, but the evidence so far is that most were actually blank and not invalid.

Turnout

Turnout was 27%, a little bit down on the 30%-35% shares in recent years. It was 19% for Young members, 28% in Wales, 42% for disabled rep and 54% for councillors. Only the disabled place was lower than I expected, as the previous BAME NEC rep election had a 61% turnout. I have listed past turnouts here for comparison.

CLP section

What conclusions can we draw from the CLP NEC results?

The new single transferable vote (STV) system had a big impact on the outcome. If this had been a first-past-the-post election, and assuming a 10,000 candidate spread in slates (a cautious smaller number than the 15-18,000 candidate spread in slates last time), the result might have been eight for Grassroots Voice (GV) and one for Labour to Win (LTW) with Luke Akehurst narrowly getting on. Some might argue that GV would have won all nine, but all would probably agree it would be eight or nine. Thus even though GV got a very impressive five seats, as a result in the change of voting system, it is still four down from 2018 and two down from just before the most recent election.

Did the events around the Equalities and Human Rights Commission report have an impact on the election? My view is that a lot of what would be “core vote” voters for the main slates would have voted early, similar to postal vote elections when up to half vote in the first third of an election period and two thirds by about half-way through. It’s possible that it may have increased the Labour Left Alliance (LLA) vote slightly, and may also have brought out a few more votes in support of LTW. That probably squeezed Open Labour’s (OL) vote a little. But I suspect that most voters motivated by the issue voted early, so GV success looks to me to be more about good organisation than as a result of political events that took place towards the end of the balloting period.

Vote share was GV 37%, LTW 31%, OL 9% LLA 5%, Tribune 4%, Independents 14%. I have posted previous results going back to 2012 here. It should be noted that despite being presented with the chance of voting for the two traditional slates, 32% chose to use their first preferences beyond them reflecting a broader diversity in voting. I cover the implications of that in more detail below.

Grassroots Voice did very well with its regional vote management strategy, and they should be extremely pleased with that. They got a very good 55% of seats for 37% of the vote. LTW and OL are likely to adopt this strategy next time, so that may make it just a bit harder for GV to hold the fifth seat. They will be pretty pleased with five seats as their organisers did say three or four was a likely expectation. However GV’s vote share was down 9% on earlier this year and 18% down on the last full election in 2018, which is a significant drop. That was partly due to a splintering of their past votes with the LLA and a number of Left Independents standing. The left coalition of 2015-16 is more splintered now and will continue to be with at least two strong left candidates in both the current UNISON and future Unite general secretary elections too.

Labour to Win plateaued at 31%, where it has stood since 2012. My understanding is that their people were hoping for perhaps a higher vote share. Their votes for candidates were heavily skewed by Luke Akehurst’s very large personal result, which did have an impact on vote transfers and their competitiveness all through the counting rounds. Nevertheless, they will be pleased with a gain and Luke Akehurst back on the NEC. The three of them are also very experienced. They fought a high-profile campaign and will no doubt want to have a strong fourth candidate (presumably a woman) and regional vote strategy next time.

Open Labour got 9% of the vote, which was a clear advance by 5% on their 2018 result, but perhaps a little lower than they might have hoped. 12% was their share in a sample of those CLPs that held all-member meetings (AMMs) with STV, which may have encouraged predictions of them adding to the gain by Ann Black. They very narrowly missed two extra seats overall, in the CLP section and disabled section. However their 9% does now clearly establish them as the party’s third slate – above Tribune.

Tribune had a poor result at 4%, which was 4% down on the 2020 NEC by-elections. They ran three candidates when they should have probably run one, possibly Paula Sherriff who ran in the earlier by-election.

Labour Left Alliance did better than I expected, though it was still 4% below their vote in a sample of STV-AMM results. Their 6,000 votes to the left of GV is a significant number, which shows a constituency is out there that may be least in favour of some of the EHRC report changes.

Most of the top nominated Independents polled poorly, except for David Anderson, and were possibly squeezed by GV. Crispin Flintoff might have expected to do better with his profile and the 12,500 votes he polled as an independent in 2014, but his voters were probably most likely to put a first preference for a GV candidate, which squeezed him out.

Nomination numbers were, as usual, a very good indicator of the outcome. Eight of the top nine nominated candidates were elected, with only Luke Akehurst with fewer nominations breaking into that top nine and he himself was tenth in nominations. Ann Henderson was the only one in the top nine nominations who missed out, perhaps illustrating that GV could have gone for five candidates, and she still came tenth in the ballot. Jermain Jackman who came 12th in nominations came 11th in the ballot. He was seen in some predictions to be the possible ninth candidate, and it shows that the predictions were not far out from the result.

Due to the length of the ballot paper, there was some evidence that being higher up the alphabet helped some candidates, with Akehurst, Amin, Anderson, Baxter, Bayanu, Black and Bolton probably doing best from this factor. I suspect this was a one-off with 42 names to choose from. A more likely ballot of 22-24 next time may see that benefit decline, and I explain the reasons for that shorter ballot paper below.

When it came to individual candidates, Luke Akehurst surprised everyone, including I suspect Laura Pidcock, who many assumed was likely to have come top in view of her nomination numbers and profile. The LTW numbers were very skewed, and they tried to correct it a bit with their campaign advertising, but their voters clearly wanted Luke on the NEC above others. Gemma Bolton also did very well, perhaps showing a London and SE base may have an impact as it is likely to be 30% of the membership.

Nadia Jama did very well as a new candidate battling with a sitting NEC member on the GV slate. As well as the GV regional vote management strategy, she benefited from being on the “social media activists slate”. Ann Henderson, as I expected, trailed not just due to the nation vote in Scotland being much smaller than many of the party English regions. It’s also possible she may have lost votes due to controversies within the party where some GV members said they were not going to vote for her on the grounds of alleged transphobia.

The full vote transfer data is here and here, and analysis is still being done by various people at present. Here are a few early bits of number-crunching:

• • Stats for Lefties flagged up that if you continue transfers forward to exclude all but the main three slates you get GV 55%, LTW 34%, OL 11%. This goes some way to explain why GV did so well to get 55% of seats on what in STV terms would have been a low vote share of 37%. He also points out that GV more efficiently and evenly distributed their vote (12% top candidates to 4% bottom candidate) than LTW where it was 16% for the top candidate to just 1% for the bottom one. This inevitably helped GV as the count proceeded through all its rounds of vote redistribution. LTW had their candidates eliminated early, and as a result had no candidates competing for the 9th place when the candidates left were two GV and one OL.
  • Andy Lawton has done some very good estimates of vote spread and vote transfers basing his political assessment of the Independent vote not on their politics but how voters saw their politics and as a result how it transferred their votes to others. He comes up with a slightly lower numbers – GV 48%, LTW 30%, OL 10% – than Stats for Lefties but makes essentially the same point that transfers from LLA and Independents benefited GV but their firmly aligned vote was just under 50%. Overall GV held on to 78% of their vote whilst LTW held on to 61%. He also very helpfully flags up the key rounds for those who do not want to look at all 37 rounds.
  • As Andy Lawton also points out, both Laura Pidcock and Luke Akehurst had around 13% of the surplus vote go to candidates not just in their slate but with different politics, perhaps illustrating that the politics of members are more complex than sometimes assumed. Tom Laing estimates an overall vote leakage of 27% for Laura Pidcock and 31% for Luke Akehurst. In the end, GV voters voted more “down the line” than LTW voters.
  • As people number-crunch the spreadsheets in the coming weeks, more insights will be found. No doubt the slates will be analysing them too as they plan ahead for 2022.

From all the analysis above, an interesting point to make about the results is that membership composition change since November 2019 may have had less political impact than some expected. In November 2019, paid-up membership was 430,000. It rose by 120,000 for the leadership election and 30,000 since then, though as the data above implies we have now had at least 20,000 leave and 65,000 in arrears or in the process of leaving, to take us down to 495,000.

From the polling for the leadership election and subsequent membership changes, the following seem to be significant points to consider:

• Taking account of the actual leadership election result, plus the majority of Lisa Nandy transfers, Keir Starmer won around 89% of new Labour members compared to Rebecca Long Bailey securing 11%.
• Among pre-November 2019 members (i.e. the same franchise as the NEC elections), Rebecca Long-Bailey polled at 38%, which was 8% lower than the left slate vote of 46%, thus some people had voted for the GV predecessor slates and Keir Starmer too.
• When adding in the new members and taking account of the 8% bonus that the left slate got over Rebecca Long-Bailey amongst pre-November members, you might expect the left vote compared to then to now be 37% (i.e. down 9% from the 46% in the November NEC by-election franchise) as a result of membership composition changes.
• Excluding the Independents for a moment, if you add the GV and LLA vote together, you get a 42% vote share. This is 5 percentage points above what current membership composition might suggest.
• However, we also know from the transfer data above that there was a strong pro-left transfer not just from LLA – to be expected – but also from Independents, which helped GV.
• Whilst there are dangers of comparing a 27% turnout NEC election with a 73% turnout leadership election, I think it is a large enough sample of the membership to make two important points for all sides in the party to consider. These may in some ways look mutually exclusive, but reflect a complexity in the party that people need to better understand:
  • The membership, having voted for Keir Starmer, may have voted for “electability” but it is still quite left-leaning (probably over policy) and thus is prepared to transfer a significant number of votes to GV in NEC elections between leadership elections.
  • The membership having the chance to vote for the left in the same way as the strong social norm of the 56% #JC9 in 2018 instead opted to spread their first preference vote much more. When given the chance, a very large 32% chose to vote outside the two main slates. That included some to the left of the main left slate, but also many to a much wider space broadly between the two main slates. This led to both OL’s increased vote share but also a wide range of leftish independents not in a slate. Something for the three main and most active slates to note is that STV does show members have a wider range of views than the two main slates traditionally encompass.

When it comes to how well the Labour election geeks did on predictions, I have to say that I, James McAsh and even GV itself did not expect them to get five seats. 5-3-1 was my third most likely scenario, after 4-3-2 and 4-4-1. I did also try to estimate vote shares, and I had GV at 36% after nominations and Independents at 13%, so I was very pleased with being within 1% of them. I was further out, by around 3%, on LTW and OL and LLA vote shares. Since much of our analysis was done on the basis of nominations and some STV AMM vote samples, I think we can say we were out, but not far out. In the end, it was always going to be hard to determine the outcome of the ninth seat, even though most people studying the data were essentially right on the other eight.

Other NEC seats

Below I will explore the other NEC results, which are set out in full here.

• Councillors. Incumbents were easily re-elected, as expected. But the Socialist Campaign Group of Councillors candidates, with 28%, did better than the 21% independents had polled against the incumbents in 2016. This is an interesting indicator of the views of Labour councillors at present and may represent a bit of change compared to a few years ago.
• Disabled. This was so close as predicted, but in the end Ellen Morrison’s (GV) member vote overcame George Lindars-Hammond’s (OL) trade union lead. It was so close that if Community Union had nominated George Lindars-Hammond rather than Emily Brothers, he would have won.
• Treasurer. Diana Holland was easily re-elected, with 74% to her nearest rival’s 8%. With close to 50% of the vote already in the bag from affiliates, she polled close to half the member vote too.
• Youth. No great surprise that incumbent Lara McNeill won here with her already-in-the-bag trade union support.
• Wales. This was always going to be close as I expected, and I did say Carwyn Jones had the profile to win despite his poor nomination numbers. Many missed that this was a vote of members and registered affiliate supporters, not an electoral college, which would have misled them over nominations. What this contest and the earlier leadership election result may tell us is that registered affiliate voters are not voting quite as left as they did in 2015 and 2016. Affiliated trade unions will no doubt note this development.
• Young Labour. A very strong showing for the GV/Momentum-backed For a Socialist Future slate. Results are available here. However, OL did win two seats and may have established themselves as an alternative option to GV in Young Labour. This also might be a significant indicator in the future politics of a re-established Labour Students, which I assume will finally move to one-member-one-vote balloting, perhaps with an STV element it always had in the past.

NEC composition

This is always difficult to get right, as voting patterns are less consistent than people think and the unions naturally strike deals, but the NEC has moved from leadership/right trade unions/LTW 17, OL/GMB three and GV/left trade unions 17 to a new NEC of leadership/right trade unions/LTW 19 (+2), OL/GMB four (+1) and GV/left trade unions 16 (-1). The letter this week to the general secretary regarding recent events over Jeremy Corbyn had 14 left NEC member signatories, and this might represent a more firm indicator of left support on the NEC in many votes.

It is important to note that the NEC gained a new disabled place and a suspended NEC member with no voting rights has now been replaced by a new full NEC member. Overall the leadership is in a stronger position on the NEC, however so is OL/GMB, which do hold an influential balancing role. Left NEC members will nevertheless be pleased that they out-performed early expectations and won a CLP seat and a Disabled Place that both could have gone elsewhere.

Ann Henderson’s loss means a vacancy for chair of equalities, and no doubt there will be the other usual changes to NEC officer posts. Luke Akehurst, Laura Pidcock and Carwyn Jones are clearly big political names on the new NEC, whilst Ann Black brings back a lot of experience. At the same time, GV has seen a big change in its representation with Yasmine Dar the only one of the “JC9” of 2018 elected and a new generation of mainly women now elected as GV activists. This may have an impact on the culture of the NEC.

Labour’s 2022 NEC elections

Having analysed the 2020 elections, what might it tell us about the options for slates in the 2022 NEC CLP STV elections? Here are a few thoughts that may stimulate debate.

Next time, Grassroots Voice will face the issue of where it goes from a better-than-expected result. Its chances of getting six seats are low, so does it focus on an incumbency strategy of holding its current five or does it run six in the knowledge that if it is a new person they may struggle against incumbent comrades? GV also needs to resolve its own decision-making on this after accusations of rush and stitch-up in recent candidate selections.

Momentum’s leadership is committed to a democratic candidate selection process, and they have already held primaries for youth candidates, joining OL who were already undertaking this process. As some have pointed out, it’s a bit hard for left NEC members to argue for open selection for MPs if left NEC candidate selection processes are opaque to its own supporters. This demand may see the end of the old CLGA “groups around the table” brokerage process, however that element might still play a role in negotiating various quotas for any supporter primary.

Labour to Win may reduce down to four candidates and operate a regional vote strategy, which may give them a better chance of a fourth seat. They also may continue to support various individual OL and Tribune candidates.

Open Labour is likely to run two again as they came pretty close to a second. They are already clear that they did not have the data of the other two slates, and I imagine rectifying that will be their focus in the coming period as they build on their support of close to one in ten activists in this election.

Tribune may reduce down to one candidate. They may strike deals with LTW and even OL.

Labour Left Alliance may think of reducing down to two, bearing in mind how well their top two candidates did, though it is also possible that their vote does represent the “most likely to leave” segment of membership in the current political circumstances.

There may be less Independent candidates after their generally poor showings. Thus we might see just 14 slate candidates and perhaps between eight and ten Independents, producing a ballot paper of 22-24 rather than the 42 we saw this time.

An issue that may impact on candidates is the issue of reform of the nomination process to bring in STV nomination meetings for NEC nominations. This would reduce allegations of “gaming” of the nomination process. This happened because the nomination requirement was for nine nominations, yet the main slates only nominated six candidates, which meant that especially in nominations where FPTP was used it was in a majority’s interest to not only nominate their six candidates but also three others who might be sympathetic to stop three candidates of another slate being nominated. We therefore saw CLPs nominate the six GV candidates and three LLA candidates, or alternatively six LTW candidates along with two Tribune and one OL.

This also likely contributed to a larger-than-average number of Independent candidates, as people voted for some of them to stop candidates of another slate. Some have argued that this goes against the whole spirit of the STV system, where these local results were not going to be reflected in the outcome. It was therefore interesting to see a significant minority of CLPs, whether general committees (GC, using a delegate system) or all-member meetings (AMM) adopt STV for the nominations too.

All slates now on the NEC may now have an incentive to reduce vote leakage by changing the nomination system to STV (perhaps leaving it optional whether GC or AMM) to possibly reduce the number of Independents without being seen to “undemocratically” increase the five CLP’s nomination threshold. It would also give us election geeks a lot more direct data early on to make more accurate predictions of the final outcome, not that I and some of the people I name above have an interest in this of course!

Hopefully all the above shows that Labour NEC results tell us a lot about the state of the party. After recent events and with EHRC implementation and the Forde report still to come, the next NEC elections in 2022 will again allow us to take the temperature of the party membership.