Brexit Breach: Labour Data Was Shared With Leave EU And Cambridge Analytica

By James Patrick – 30/04/18

Sensitive personal data of Labour voters was processed by a third party and shared with Arron Banks’s Leave.EU, Cambridge Analytica, and others associated with unofficial groups campaigning to leave the European Union in February 2016.

Data based upon demographics, class, finances and ethnicity, was used to identify core groups of Labour voters to be targeted with UKIP-led messaging and was instrumental in deciding where Nigel Farage appeared to speak during the Brexit campaign.

Leave.EU, Cambridge Analytica, the RMT Union and Trade Unions Against EU, and Labour MP Kate Hoey – associated with Labour Leave – gained access to the information via Labour’s 2015 general election data guru before referendum campaigns were officially designated by the Electoral Commission.

Blue Collar workers, struggling families, students, and ethnic minorities were among those specifically designated valuable to tailored social media targeting and doorstep canvassing. The data provided specific postcodes to be targeted on and offline, in order to attract millions of votes across the country – enough to swing the divisive referendum result.

The postcode and demographic briefings are being released in full, in the public interest, to assist any concerned voters in establishing whether they were affected as Labour have remained largely silent on the issue of Cambridge Analytica and concerns over data profiling.

“It was Labour data that formed our strategy and therefore where we deployed Farage – Andy Wigmore, Leave.EU”

Sensitive personal data, which includes ethnicity, was allegedly compiled from Labour Party information by a third party consultant and shared with Arron Banks’s Leave.EU campaign group, Cambridge Analytica, Brian Denny of the RMT Union, and the MP Kate Hoey.

The huge dataset, based on the information of millions of Labour voters across the country, was allegedly built using Mosaic demographics and the results of party canvassing. It is believed to have been amassed during 2015 by political consultant Ian Warren, before he passed it on in a series of detailed briefings and a postcode targeting spreadsheet in early 2016.

He first met with Cambridge Analytica to discuss the use of the information as part of Leave.EU’s campaign at the end of 2015.

Warren was head-hunted by Labour for the 2015 election campaign after his successful work with UKIP and continued to be closely associated with the party, polling members and working with Owen Smith on his leadership challenge during the remainder of 2016.

Leave.EU’s Andy Wigmore said: “He ran Ed Milliband’s team and the general election campaign in 2015 for the Labour Party. He was and still is the Labour Party guru.” A self-taught statistician and political consultant, Warren trades under the name Election Data Limited, based in Bolton.

When asked whether he was still working for the Labour Party at the time of the leak, Warren said: “I’m sick of speaking to journalists about this. I’ve nothing to say.”

When asked whether he had the right to retain and use the data, Warren terminated the phone call.

labour leak infographic

In redacted emails released during a recent session of Parliament’s Fake News Inquiry, chaired by MP Damian Collins, former Cambridge Analytica employee Brittany Kaiser showed that Warren had been in contact with the controversial election firm and Leave.EU in early January 2016.

The briefing documents mentioned as attachments to the emails were not made available to the public. These have now been provided by Andy Wigmore of Leave.EU, along with further correspondence dated February 14 2016 which was not released by Parliament or referenced in Kaiser’s written evidence to the committee.

The later correspondence includes the postcode targeting data, meaning the voters’ demographic and ethnicity data could be matched with additional information to potentially identify them as individuals. This appears to contravene data protection rules published by the ICO.

The February email was sent to: “Arron Banks; White Knight; Brendan Chilton; Kate Hoey; B.Denny@rmt.org.uk; julian.wheatland@cambridgeanalytica.org; Peregrine Willoughby-Brown; Brittany Kaiser; Jordanna Zetter; Liz Bilney; Andy Wigmore.”

The subject of this previously unseen email was: “Target postcode sectors – Labour core vote.”

Screenshot (695)

The spreadsheet identified 107,406 target postcodes for “blue collar” Labour households and 131,691 target postcodes for “deprived and disaffected” Labour voters, with Warren’s email instructions saying the blue collar households should be targeted first.

He wrote: “Please find attached the complete list of target postcodes for Labour core voters susceptible to Leave messaging. I understand this will feed into the social media campaign in the first instance but this data should also determine where doorstep and street stall canvassing activities take place. If any of you are planning Labour Leave or Leave.EU canvassing activity I suggest you consult these sectors or contact me directly (details below). On average there are 15 households to each postcode sector in the UK, meaning there are a couple of million in the country. The target postcodes have both a high concentration and spatial clustering of blue collar voters, as measured using Herfindahl indices and LISA maps.”

“I suggest we start with the group I describe as Blue collar voters – you should all have received my report on these voters (I attach it regardless). They turn out to vote in higher numbers than the deprived and disaffected voters. That said, there will be scope to target deprived and disaffected voters too. I suggest we also begin with the major metropolitan local authorities which have high numbers of target postcode sectors and work our way down the list. You will see that there are tabs showing the number of target postcode sectors in each local authority and parliamentary constituency,” he continued, adding: “I can provide detailed maps to support the data enclosed, and am happy to provide canvassing sheets should that be required.”

Postcode_targets_allPostcode_targets_all – (Target Type Redacted)

The postcode data was complimented by a briefing report which repeated information already sent in January 2016, but which neither Hoey nor Denny were copied in on at the time. Election PR firm Gerrard Gunster was however an addressee on the January mail, which featured an additional two pages at the front of the document.

One of five, the report seen alongside the postcode targeting data focuses on the “blue collar” group of core Labour voters.

The other four reports have not been provided for examination, but they are listed and referred to as: Struggling, financially strapped households, Young, well-educated, early-career liberal metropolitan voters, Current students, aged 18 to 25, and Ethnic minorities.

In his accompanying January email, Warren was clear he had already been working with Cambridge Analytica on the Labour data, writing: “Each of these groups have distinct demographic characteristics, personality traits and political behaviours. They will require subtely different messaging as a result. Julian and the team at Cambridge and I had a productive meeting in London before Christmas and I hope this report will start to flesh this out.”

“Labour voters have relatively poor turnout levels as a rule. The traditional approach is to have a phased campaign, with ‘persuasion’ messaging earlier in the campaign giving way to ‘get out the vote’ messaging later on. I anticipate this will be the case with the Labour voters here. That said, there are some socially conservative Labour voters with fairly closed personality traits which have either already made up their minds or view any messaging with suspicion if it doesn’t chime with their world view. With those voters the aim should be simply to ‘keep the pot boiling’ with issues like immigration rather than continually hammering them,” he continued.

“For younger, liberal voters who are high-information, open and extroverted personalities there is a place for persuasion messaging to be effective if it is dispassionate and objective,” he added.

Labour vote briefing

The “blue collar” briefing document identifies six key sub-groups of Labour voters as: Middle-aged families in less fashionable suburbs, Low income communities reliant on low skilled jobs, Relatively comfortable homeowners in former industrial areas, Older families in traditional industrial areas, Middle-aged families in former right-to-buy households, and Older people in social housing.

It goes on to identify the main characteristics of this group of voters as being early school leavers in lower paid jobs, and likely to be struggling to make ends meet with little or no savings to fall back on.

According to Warren’s analysis: “They are resentful of New Labour for allowing large-scale immigration and for failing to respond to their anxieties and concerns. For these groups the Labour party is dominated by professional politicians or socially liberal London-centric elites who treat them with disdain or condescension. In local and national elections these voters have been attracted to UKIP because of the lack of mainstream voice for their combination of social conservatism, working-class identity and economic ambition.”

“They are resentful of politicians in general. They don’t trust them or even like them. They believe they are ‘in it for themselves’ and are unlikely to trust them when they make policy promises,” he added.

He identified these households as being closed minded and dogmatic, in particular in relation to welfare, the EU, and immigration, adding: “They are socially conservative in many respects and not trusting of external sources of information such as politicians and the media if it does not conform to their world view. They are also slightly neurotic, which leads them to emphasise the unfairness in issues such as welfare or immigration.”

Warren then began to outline the need for short, simple to understand, emotional messaging to be used when targeting these Labour voters.

He wrote: “It is crucial to translate a high-level policy message into language they understand and with reference to their own lives. For example, during the last election Labour had a policy on local bus routes, meaning local people would have a say in which routes were kept open and which were closed. This was a simple message on a doorstep to such voters: “Labour will make sure the Number 24 goes to Market Street every week”. Tangible, easy to understand, and quite easy to deliver; all of which meant that mistrusting voters couldn’t say it was unachievable or pie-in-the-sky.”

“Of course, as it happens the election was never going to be decided on local bus routes (!) but the point remains. Make the messages real and tangible for people. After all, if you can’t persuade voters how leaving the European Union will make a tangible difference then why leave the European Union?” he added.

Crucially, Warren then identified the manner in which Nigel Farage should be deployed in any efforts to harness the Labour vote for leave during Brexit.

He wrote: “Should Mr Farage be used, it should be done sparingly and only to ‘keep the pot boiling’ for such voters. Mr Farage could be used effectively at times of a specific crisis in migration, for example, to underline the negative effects of immigration on working households.”

The report went on to assess the sub-groups of voters on a number of different factors, including:

  • Demographics: Age; household composition; marital status; length of residency in neighbourhood; social grade; religion; ethnic origin.
  • Property: Property type; home ownership tenure; council tax band; property value; age of property
  • Employment Economic activity; occupation (if employed); occupation (NS-SEC); industry of employment
  • Income & expenditure Household income (net); household income (gross); likely disposable income; individual monthly income; net household wealth; weekly household expenditure; main items of expenditure; grocery shopping value.
  • Finances Investments; debt levels; perceived ability to cope financially; credit or debit cards; benefits claimed.
  • Home lives Holidays taken in the last year; cost of last holiday; car ownership; transport needs.
  • Perspectives Environmental awareness; motivations; social attitudes; health; community safety.
  • Education Qualifications; age completed education; higher education level.
  • Communications Communications channel used; purchasing channels; where do they learn about products; mobile phone usage; satellite TV; favourite TV programmes; internet use; newspapers read.
  • Political behaviour Where these voters live (mapped); Westminster constituencies with the highest proportions of these voters; UKIP and Labour performance in the highest-ranked Westminster constituencies in 2015; identification with main political parties; voting patterns since 2005; turnout performance.

Set against the postcode data, and further information to be gathered during tailored campaigning – on and offline – each household and even voter becomes inevitably identifiable.

The ICO guidance takes a stepped approach to identifying personal data, asking first “Can a living individual be identified from the data, or, from the data and other information in your possession, or likely to come into your possession?” then “Does the data ‘relate to’ the identifiable living individual, whether in personal or family life, business or profession?”. If the answer is yes to both, the information is personal data and subject to the Data Protection Act.

Importantly, the ICO adds: “Simply because you do not know the name of an individual does not mean you cannot identify that individual,” and “Data which identifies an individual, even without a name associated with it, may be personal data where it is processed to learn or record something about that individual, or where the processing of that information has an impact upon that individual.”

labour leak infographic 2

Andy Wigmore confirmed a number of details about the emails and attachments, confirming they didn’t simply buy the information but commissioned Warren to complete the briefings and spreadsheets.

According to Wigmore, Warren used “canvass Labour Party card data and mosaic to create the reports.”

Highlighting that Brian Denny – a senior figure in the RMT Union who was also involved in the Trade Unionists Against EU campaign – was copied into the February email alongside Hoey, he said of the MP: “she was still trying to exit from Vote Leave at the time, who of course did not believe the Labour vote was that important!”

Hoey is still shown in rally images on the Labour Leave website.

Electoral Commission records show that Banks donated £108,000 in cash to Trade Unionists Against EU through his other company, Better For The Country, which is currently under investigation over alleged mis-spending.

Wigmore explained why he had commissioned Warren to produce the Labour data, saying: “Basically, if we had won designation then this data would form the base of our campaign – worth noting no physical data would have changed hands but it does show you that Cambridge would not have used any Facebook data or anything like that – the outcomes of the report would have decided the strategy.”

“It’s interesting stuff and basically it was Labour data that formed our strategy and therefore where we deployed Farage etc…these reports were also one of the reasons we felt we did not need CA,” he added.

According to Facebook’s own pages, audiences can be targeted by postcode along with a broad range of additional demographics which have caused alarm as public inquiries have gathered pace across the world.

labour leak infographic 3

According to the ICO guidance, permission should have been obtained from Labour voter’s before any personal information – for example canvass data – was used for any other purpose or transferred for use by third parties.

The Labour Party’s own privacy policy also states: “We will never sell or share your personal information with other organisations for their direct marketing purposes without your explicit consent. We may share your data with third parties to perform services on your behalf and to help promote the Labour Party by serving you advertisements and content online about our politicians, campaigns and policies we think you might be interested in.”

An official Labour party spokesperson said: “We do not comment on staffing matters. The Party has strict confidentiality clauses and processes around the use of personal data and under no circumstances would share this data or permit this to be shared with external agencies.”

A well-placed source within the party said: “Fuck me, this is a scandal and a half.”

When challenged over the correspondence and the right to retain or use any Labour data, Warren terminated the call.

While no ICO registration for Warren has been identified yet, potentially due to fault on the website, a spokesperson for the ICO declined to confirm or deny whether Warren was under investigation as part of their extensive Brexit related inquiry.

The Information Commissioner said: “As part of my investigation into the use of personal data and analytics by political campaigns, parties, social media companies and other commercial actors, the ICO is investigating 30 organisations, including Facebook.”

Hoey responded to queries on the correspondence and her involvement in writing, saying: “Sorry. Haven’t got a clue what you are talking about. Labour Leave was a separate entity and I was not a member of any boards / committees etc of either Leave EU or Vote Leave. All sorts of emails floated around during the referendum campaign.”

Denny has not responded to telephone calls or an email request for comment.

The Electoral Commission were asked if collaborating between the unofficial campaign groups was against the law, but provided only generic working together rules in response. It is, however, against the regulations to consult other campaigners about what to say in a campaign, to whom, or even to discuss how such activity should be organised.

Labour are currently engaged in a significant social media and door-to-door push to gain votes in Thursday’s local elections.

Update: 00:20 01/05/18

Warren has subsequently posted a lengthy entry on his blog in which he denies any wrongdoing, claims the data is his own, and apologises to his “friends in Labour,” which can be read here.

*This article and the documents have been referred to both the ICO and the Electoral Commission in order to assist them in identifying any potential offences.

InfoGraphic Background: <a href=”http://krulua / 123RF Stock Photo“>Copyright

Red ‘Error’ Header Image:<a href=”http://iunewind / 123RF Stock Photo“>Copyright

Advertisements