268.It is the responsibility of all of us to ensure that representative democracy is supported, rather than undermined in the digital era, and that, above all, elections remain free, fair and trusted. The inability of our electoral law to reflect the rise in digital campaigning places that trust at risk.
269.Digital campaign tools can make it easier and cheaper for campaigners to communicate with voters. This in itself is no bad thing. Electoral Commission statistics show that the proportion of money spent on digital political advertising has steadily increased. It is also possible, given the currently unknown impact of COVID-19 that digital campaigning will become ever more important.
270.However, there is growing uncertainty about the way in which digital campaigning is regulated and how much online electoral processes can be trusted. The Nolan Principles of honesty, leadership and openness are of particular relevance here. Holders of public office, and those who seek it, should act in an open and transparent manner; they should be truthful, and they should actively promote these principles and be willing to challenge poor behaviour wherever and whenever it occurs. We cannot be certain that our leaders are adopting these behaviours when it comes to online electioneering, which creates fertile ground for trust in democracy to be undermined. The Committee on Standards in Public Life launched a review of electoral regulation in June. We look forward to reading their conclusions and hope that this Chapter will inform that review.
271.The LSE Truth, Trust and Technology (T3) Commission expressed the urgency of the situation neatly:
“If the policy framework is not updated, the ability of ‘rules of the game’ to ensure that elections are free, fair and legitimate will increasingly be called into question. The UK should not find itself having to go to the polls again before the legislative framework is modernised.”
272.There is no clear definition of a ‘campaigner’ in law. We formulate a definition in the box below.
There is no single definition of campaigning in law that can be readily used to define what a campaigner is. The Political Parties, Elections and Referendums Act 2000 (PPERA) has separate definitions of campaign spending for registered parties, third parties and referendum campaigners. The Representation of the People Act 1983 has another definition for election spending by candidates.
Advancements in digital technologies also bring the question of what a campaigner is into focus. The PPERA, drafted 20 years ago, does not account for the present situation.
In their 2018 ‘Democracy Disrupted’ report, the Electoral Commission used the term ‘campaigner’ loosely as an umbrella term for political parties, third parties, permitted participants, unregistered referendum campaigners and candidates.
We define a ‘campaign’ as coordinated activity that promotes electoral success or promoting a referendum outcome. This applies to political parties, registered and unregistered third parties and candidates in local and national elections.
273.We add our voices to those calling for reform of UK electoral law. We detail what should be included in this piece of legislation and what the Government should urgently consider as part of this reform.
274.In the most fundamental sense, the voters have to trust the electoral process in order for them to accept its results. Elections are the mechanism through which we decide our representatives; that they are trusted, and their results accepted, are the keystones upon which democracy rests.
275.Electoral law has not kept pace with changes to digital campaigning and as such it fails to regulate modern electoral campaigns effectively. In recent years a succession of public bodies have called for the significant reform of electoral law. In March 2020, the Law Commission published the final report of its nine-year review of electoral law. It concluded that our electoral laws are outdated, confusing and no longer fit for purpose. Its main recommendation was that reforms were needed to bring electoral law into a single, consistent legislative framework.
276.As has been put to us many times, developments in digital technology far outpace the response rate of legislation. Spending on digital campaigning has increased at each General Election in recent years, as Figure 3 below illustrates. The graph below shows spend by political parties and permitted participants and only looks at spending on a small selection of digital platforms which were reported as ‘spending on advertising’. Given that this chart does not reflect third-party spending and the full range of digital platforms, it is likely that the true picture of digital spend is much greater than is shown.
277.And yet, the main Acts governing our electoral law date from 1983 (Representation of the People Act) and 2000 (PPERA) respectively. The Law Commission pointed out that the 1983 Act has its origins in the Corrupt and Illegal Practices Act 1883 and that much of it had not been updated since. The Figure below illustrates the absurdity of this.
278.The Law Commission suggested that modernising the framework of electoral law would mean that making changes to the law would be less complicated, allowing the law to respond faster to societal and technological developments. The Electoral Commission told us that wholesale reform of electoral law would be their main request to Government.
279.Furthermore, the Electoral Commission suggested that not only were these Acts outdated, there is a deficiency in how they interrelate. This means that the party and candidate regulatory frameworks are not always well-aligned. This lack of alignment has led to several deficiencies in compliance with, and enforcement of, the regime. Louise Edwards, Director of Regulation at the Electoral Commission, suggested that the public were probably confused about the current state of electoral law, explaining that the current legal framework does not work for some of the techniques that people use online.
280.Devolution will have a significant impact on reforming electoral law. Some electoral law is devolved to the Scottish and Welsh Parliaments. The Scottish and Welsh governments have the power to make rules for parliamentary and local government elections in those jurisdictions. However, as the Law Commission explained, there are some rules for those elections that remain the responsibility of the UK Parliament. These include the rules regulating the registration of parties and donations to registered political parties.
281.The Law Commission recommended what it calls ‘rationalisation’–a single, consistent framework that applies to all elections and referendums, with a consistency of approach across all the UK legislatures. The devolved administrations would maintain their current devolved legislative competence; rationalisation of electoral law here would not mean repatriation of devolved powers.
282.When we asked the Government about their response to the Law Commission’s report, Chloe Smith MP, Minister of State at the Cabinet Office, told us that consolidation was not her priority, and that her focus spanned “operations, resilience and security”. Whilst we would not dispute these goals are important, the two are not mutually exclusive: the rationalisation of electoral law to make it fit for the digital age should improve operations, security and most importantly, public trust in democracy. It is also long overdue.
283.Electoral law should be modernised, rationalised and brought into the 21st century. Whilst we focus in this Chapter on how this should be done in relation to digital technologies, we urge the Government to consider seriously the Law Commission’s proposals in their final report on reform of electoral law. In this report we focus on modernising electoral law to make it fit for the digital age, without altering the devolution settlement, in line with our remit; however, the Law Commission’s report goes much further than this and focusses on simplifying and clarifying the law.
284.The Government should bring forward a Bill based on the proposals set out by the Law Commission that comprehensively modernises electoral law. This should be completed in all its stages before the next General Election.
285.The issue of digital imprints is unavoidable when considering how to make electoral law fit for the modern era.
Election material is published material that can reasonably be regarded as intended to influence voters to vote for or against a political party or a category of candidates at certain elections.
Whenever (offline) election material is produced, it must contain certain details, which are referred to as an ‘imprint’, to show who is responsible for the production of the material. This helps to ensure there is transparency about who is campaigning.
286.Imprints are an important consideration because they mark a key difference, and a gap in transparency, between the regulation of offline and online election material. The PPERA requires printed campaign materials to carry an imprint, but, dating from 2000, it makes no equivalent requirement for digital campaign materials.
287.The Electoral Commission has been calling for digital imprints since 2003. They argue that the gap in transparency is affecting voter confidence and the Commission’s ability to enforce the rules. The Electoral Commission’s 2019 Election Report found that nearly one in five people were not confident that the election was well run. Many of these people selected reasons related to concerns about campaigning or the media to explain their lack of confidence. The Electoral Commission stated that it is often too unclear who is behind digital election campaign material, and significant public concerns about the transparency of digital election campaigns risk overshadowing their benefits. Extending imprints to digital campaigning would go some way towards reducing this confusion. The Electoral Commission currently advises campaigners to include an imprint even though it is not required under law, however, it is unclear if there has been widespread, consistent application of this advice.
288.In our evidence, there was overwhelming support for introducing digital imprints. The Network for Media and Persuasive Communication at Bangor University, for instance, told us that imprints were important for transparency, and that it was important to make clear who was behind an advert. The recommendation to introduce digital imprints was endorsed by academics and many different organisations. The Royal Statistical Society noted that extending imprints to include online political campaigning had been recommended no less than eight times in the last two years by other committees and inquiries. MySociety suggested that introducing transparency through measures such as imprints would not automatically resolve actual or perceived issues in campaign finance, but it would enable greater potential for examining and discovering wrongdoing sooner. The Conservative Party told us that they support and advocate the extension of imprints to electronic campaigning material, and that the imprint would provide a check and balance against inappropriate content, as the publisher or promoter could be held to account publicly for the content. Many other pieces of evidence advocated the introduction of imprints on digital campaign material.
289.In calling for digital imprints, we add our voice to a strong chorus of bodies and organisations who have consistently recommended this measure. The Law Commission recommended extending imprints to online material, to reflect the growth of online campaigning and the use of social media in recent years. The House of Commons DCMS Select Committee’s inquiry into fake news suggested a similar reform to the law. The Committee suggested that there was a need for additional clarity and new definitions of what constitutes online political advertising and that there should be “absolute transparency of online political campaigning, including clear, persistent banners on all paid-for political adverts and videos, indicating the source and the advertiser.” The Committee on Standards in Public Life has endorsed the Electoral Commission’s recommendation. The Oxford Internet Institute’s Technology and Elections Commission final report recommended that imprints should be provided on all digital ads and sponsored content as one of their immediate proposals for action.
290.The Government consulted on introducing digital imprints in late 2018, and in May 2019 committed to bringing forward the technical proposal for imprints by the end of 2019. No such proposal has emerged, and an election has taken place since that commitment. In their evidence to us dated April 2020, the Government provided no further indication of when imprint legislation, or even legislative proposals, will be introduced, saying that further details would be announced “in due course.” It is extraordinary that with such overwhelming support for this proposal, to which the Government has repeatedly stated its commitment, no progress has been made. This delay is unacceptable; “in due course” provides no clarity and we question how many more election campaigns will be run, and how much further public trust will be eroded, before this proposal is put into law. We draw the Government’s attention to the substantial number elections that were due to take place in May 2020 which have been postponed until May 2021 due to the COVID-19 pandemic, including English local councils, 13 directly elected mayors in England and 40 police and crime commissioners in England and Wales. These elections will take place on the same day as the Welsh and Scottish Parliament elections which are due to take place in 2021.
291.We note that there is a digital imprint requirement to show the name and address of the person responsible for the online campaign material contained in the Referendums (Scotland) Act 2020, passed by the Scottish Parliament on 19 December 2019. Digital imprints were also in place for the Scottish Independence Referendum in 2014. This is notable because it provides an example of a poll conducted in the UK that took place with digital imprint legislation enacted and therefore simply proves it can be done. Each of the devolved administrations must decide for themselves, but we advocate consistency across the UK in line with the rationalisation aim of the Law Commission’s proposals.
292.Whilst we add our voices to those who have been calling for digital imprints to be enshrined in primary legislation, we also wish to draw the Government’s attention to the fact that the PPERA contains a specific power for the Secretary of State to use regulations to extend the imprint rules to non-printed material, which would include digital and online materials. As the Law Commission’s report makes clear, electoral law in the UK is fragmented and complex. There are various pieces of legislation that cover rules relating to parties or candidates and separate pieces of legislation covering UK-wide elections, and elections in England and the devolved administrations. There are imprint rules in all those pieces of legislation, which could be extended via secondary legislation. This secondary legislation power means that the Government has more flexibility to change this part of electoral law compared to other aspects of law which can only be changed via primary legislation, and therefore progress can be made whenever the Government decides it would like to prioritise this.
293.Whilst we urge the Government to undertake a holistic reform of electoral law, we are mindful that the Government has committed to action on the issue of imprints and that developments in technology move fast. Clarity about the Government’s intention is desperately needed, and the use of secondary legislation could provide a remedy in the short-term. We note that recent advert labelling initiatives by the technology companies, albeit in a patchy and inconsistent fashion, show that digital imprints would be workable. We encourage the Government to introduce this as an urgent first step in reforming electoral law to make it fit for the digital age.
295.Campaign spending is regulated by the Electoral Commission, which derives its powers from the PPERA, and reports to Parliament. On numerous occasions we heard that the Electoral Commission’s powers were insufficient to properly regulate electoral conduct in a digital age. Their budget is determined by the Speaker’s Committee on the Electoral Commission. Baroness O’Neill of Bengarve told us that the one thing she would impress upon the Government to enact in the short term would be, “sorting out the powers of the Electoral Commission.”
296.The Electoral Commission has extensive powers once it has declared a formal investigation into an alleged breach or breaches of electoral law, but limited powers beyond this. One of the rules of the PPERA is that political parties, campaigners, members associations and elected officials must report spending, donations and loans above £500 to the Electoral Commission. If these rules are broken, the Commission will investigate.
297.Louise Edwards from the Electoral Commission told us that the resourcing of the Commission was not the problem, but that the present legal framework does not allow them to regulate elections in a digital age. For example, they do not currently have the power to go to a social media company outside of an investigation to find out who is behind a particular online campaign. The Commission requested further powers to obtain information from others, beyond those they regulate, where it is in the public interest to do so. They argued that this would allow them to deal with compliance issues in real time, and compel organisations, for example social media companies, to give them information about the source of an online campaign. In their 2018 report, they asked for the power to request relevant information and explanation from outside parties who may hold relevant material outside of formal investigations. They argue that this would bring the Electoral Commission’s powers in line with those of the ICO. Craig Westwood told us that the most important thing was to “move quickly and reassure the public in what is a live campaigning period, when people are trying to influence people every day and people are already starting to put in postal votes.”
298.Our concern here is trust. Voters must be able to trust what they believe they are being told and who is trying to influence them. In the context of increasing voter concerns about the conduct of online campaigning, people need to be assured that the regulatory system is working in an effective way, and that those who breach it will be dealt with adequately and in a timely manner. To do this, the Electoral Commission must be able to acquire information from third parties such as social networks. We note that the ICO also have the power to issue assessment and enforcement notices in addition to information notices. We limit our recommendation to information notices for the Electoral Commission, on account of a lack of evidence that assessment and enforcement notice powers are desirable for the Electoral Commission.
299.The reform of electoral law should grant the Electoral Commission the power to acquire information from external parties such as social networks about campaigners’ activities outside of a formal investigation.
300.The rise of digital campaigning has arguably confused the public. The Electoral Commission’s 2019 General Election Public Opinion Survey found that public perceptions of political information online were that it is unclear and untrustworthy. 46 per cent of respondents were concerned about why and how political ads were targeted at them; only 29 per cent agreed that they could find out who had produced the political information they saw online, such as who had paid for it and who produced it; more than half (52 per cent) of respondents felt that ‘inadequate control of political activity on social media is a problem’ and almost two-thirds (60 per cent) of respondents thought that information available online about politics is not trustworthy.
301.Members of the public expressed concern to us about the lack of clarity as to how much was being spent online during campaigns. Christina Eager suggested that there must be rigorous checks and the ability to impose punitive sanctions on organisations and individuals who break the rules. John Brace told us that he routinely inspects candidate election expense returns in the local government area and has observed that candidate expenses do not always reflect the true cost of digital campaigning. He stated that some costs that should be declared are sometimes left out or other basic errors are made on returns.
302.The Electoral Commission recommended in their 2018 report that campaigners should be required to submit detailed and meaningful receipts from their digital suppliers and that parties should have to break down their spending returns into additional sub-categories. As it currently stands, national regulations in each of the UK’s legal jurisdictions do not require sufficient detail in campaigners’ submissions for observers to have a meaningful understanding of parties’ online activities. Spending on campaign activity is declared under broad categories such as ‘advertising’; there is no specific legal category for digital campaigning. This is unhelpful and provides no real insight into how campaigners are spending money on digital technologies and online. The information must be meaningful for everyone who uses it, so that the public can better understand how online campaigns are being conducted.
303.However, Dr Luke Temple from the University of Sheffield and Dr Ana Langer from the University of Glasgow raised the issue of additional rules on transparency possibly having a detrimental effect on small civil society organisations’ ability to take part in democratic debate. They stated that these organisations do not have access to the lawyers and accountants needed to navigate these rules. They suggested that transparency rules should be accompanied by well-resourced providers of free advice on how these rules apply.We agree that the barriers to entry to democratic debate should not be raised unnecessarily, yet current spending categories are insufficient and do not provide any meaningful or practical insight into how campaign money is spent online. The Electoral Commission should therefore be encouraged to provide clear and accessible guidance about the level of detail that campaigners should include in their spending returns.
304.Electoral law needs to change to ensure parties must submit detailed returns and provide a meaningful account of their online spending. A possible further reform for the Electoral Commission to consider would be to link digital imprints to spending returns, so that the full information and context is available to the public.
305.The reform of electoral law should support the Electoral Commission in creating statutory guidance on the level of detail campaigners set out in receipts concerning digital spending and in their spending returns to the Commission, to provide the public with a greater understanding of the breadth and nature of online campaigns.
306.Since 2010, the Electoral Commission has had powers to investigate potential breaches of the rules set out in the PPERA, and to issue sanctions if breaches are found to have occurred, including variable fines up to a maximum of £20,000.
307.Looking at national spending, in the 2017 election the Conservative Party spent more than £18.5 million. The Labour Party spent £11 million and the Liberal Democrats spent £6.8 million. Taken together, there was a total spend at the 2017 election of £41.6 million by a combination of 75 parties and 17 non-party campaigners.In this context, the Electoral Commission’s maximum financial sanction of £20,000 is a grossly insufficient sum indicating that the regulator lacks the power to drive proportionate compliance. In the Electoral Commission’s 2018 report and evidence to this Committee, it suggested that the current £20,000 maximum fine could be considered the cost of doing business for some campaigns and does not act as an effective deterrent to stop campaigners from committing offences.
308.Dr Robert Dover from the University of Leicester also argued that fines must be levied at a significant rate to avoid the conclusion being drawn that investigations and fines associated with breaches of electoral rules are now transaction costs in the process. The principle of the application of bigger fines by the Electoral Commission was also endorsed by the LSE T3 Commission. The APPG on Electoral Campaigning Transparency in its report on Democracy in a Digital Age suggested removing the cap on the fines the Commission could levy.
309.By way of comparison with the Electoral Commission, the ICO can levy fines of up to €20 million or in the case of an undertaking, up to four per cent of the total worldwide annual turnover of the preceding financial year, whichever is higher. Even before the introduction of the large fines available under the General Data Protection Regulation (GDPR), the Information Commissioner was able to fine up to £500,000 under the Data Protection Act 1998. The Electoral Commission noted that this is an uneven benchmark with the ICO able to levy substantially higher fines than themselves. The Information Commissioner herself advocated an update in the Electoral Commission’s powers, resources and compliance tools. She stated that the maximum fine the ICO were able to levy was a significant power to drive compliance.
310.We note that many fines are caused by local volunteers not accurately reporting the donation received. It is important that the fines levied are proportionate to the offence; many of the fines currently given out by the Commission are in the £200-600 range. However, proportionality works both ways: the Electoral Commission must be given effective deterrent powers in an age where millions of pounds are spent at UK wide elections and referendums, especially on digital campaign activities. Campaign spend here uses the definition of campaign expenditure as set out in the PPERA, “expenses incurred on or on behalf of the party… for election purposes.”
312.A concern repeatedly raised with us was the ability of digital campaigning techniques to blur the boundaries between local candidate and national spending. The emergence of micro-targeting techniques, particularly targeting based on location, means there is greater uncertainty about whether an advert falls under local or national-level spending limits. The Open Rights Group suggested that the targeting abilities of social media platforms does indeed blur the line between differentiated spending limits traditionally placed on actors by candidate, party or national designations to the point where these are essentially cosmetic. Dr Martin Moore also suggested that parties were using digital technology to circumvent limits on local spending. Bethany Shiner told us that the regulation of the different caps in spending for local and national elections must be strengthened to avoid the funnelling of money into swing seats which should fall under local spending, but might be paid for out of the national budget. Dr Luke Temple and Dr Ana Langer suggested that increased transparency would mean rethinking the distinction between constituency and national spending, as digital technologies have made them increasingly hard to distinguish.
313.The LSE T3 Commission pointed out the discrepancy whereby targeting voters online in particular constituencies is defined as national spending, for which limits are set far higher than for constituency spending. They argued that this undermines the principle of a level playing field at a local level. The Law Commission in their electoral law report also noted concerns about blurring the line between national and constituency expenditure.
314.The distinction between national and local spending is somewhat murky: the criteria is whether a particular advert is trying to promote the election of a candidate or encourage people to vote for that party. Louise Edwards from the Electoral Commission told us that, “In practice, even if an advert is targeted at a single constituency, it comes down to whether the content of that advert is about the candidate or the party’s national policies. If it is about the party’s national policies, it is still party spending.”
315.However, Ms Edwards told us that there was no evidence that the distinction between national and local spending was not working. After the 2015 Election, the Electoral Commission fined a party a significant amount of money for reporting some candidate spending as party spending, but Ms Edwards told us this was an anomaly.
316.Nevertheless, the APPG on Electoral Campaigning Transparency recommended that the Electoral Commission should oversee local candidate spending in addition to national spending, which they currently regulate. The Electoral Commission told us that they had wished to oversee local candidate spending for a long time and that there were obvious benefits to the same regulator overseeing national and local spending declarations. Currently, the Electoral Commission cannot take further action themselves if they find inaccuracies in candidate spending; it is the responsibility of the police to deal with any allegations that a return is inaccurate.
317.The Law Commission stated that it excluded national spending regulations from its review as the issue was too political for a neutral body to consider but did state that any holistic reform of electoral law should include both areas. The Law Commission’s report suggested merging as many of the different rules governing elections into simple single codes that are easier to understand. On this proposal specifically, the Law Commission concluded that there are a number of practical barriers to overcome because the Electoral Commission does not have a register of candidates and so would not be able to determine if a candidate had failed to submit a spending return. In its interim report, it recommended that should spending returns go digital in the future they should be overseen by the Electoral Commission. The Law Commission also concluded that it could be an area of future work for itself.
319.Under the PPERA, any contribution of goods, property or services given to a political party with a value of less than £500 is not considered a donation and therefore does not have to be registered. Actors who spend below the £500 limit therefore avoid scrutiny.
320.In its report, Defending Democracy in the Digital Age, the APPG on Electoral Campaigning Transparency’s first recommendation was to regulate all donations by reducing permissibility check requirements from £500 to 1p for all non-cash donations, and £500 to £20 for cash donations. The APPG made the point that the £500 limit on donations was conceived at the time when tracking such sums of money was relatively simple; today £500 can be transferred from anywhere in the world instantaneously. Dr Bethany Shiner told us that foreign and automated donations made through PayPal make it possible to escape the prohibition on the £500 limit.
321.The Open Rights Group suggested that digital technology had allowed for a proliferation of ‘astroturf’ campaigns, that claim grassroot status but are coordinated and funded by larger established commercial lobbying organisations or pre-existing political campaigns. Most organisations do not have to register as a non-party campaigner unless they are spending £20,000 in England or £10,000 in Scotland. The Open Rights Group stated that it is relatively easy under present circumstances for political campaigners to obscure the sources of their funding.
322.The Electoral Commission told us that lowering the current spending limits would impose a regulatory burden on small campaign groups and would require a large amount of resources from the Electoral Commission. They contrasted the US system where campaigners are required to register with the Federal Election Commission (FEC) with the UK’s system which has no such comprehensive register. The FEC register allows online imprints through Facebook to connect directly to the organisation’s registration with the FEC. MySociety told us that stronger requirements for parties to know the identity of donors would help address concerns that funding limits can be circumvented through multiple donations below the £500 limit, before stronger disclosure standards come into effect. It cited the example of the US FEC site as a good model.
323.We propose that the Electoral Commission should look into the feasibility of a secondary registration scheme for campaigners that fall below the £20,000 threshold at which they have to register as non-party campaigners with the Electoral Commission. This is with the aim of transparency, which here refers to an understanding of linked business interests; it is not our intention to make onerous demands on small campaigners, and hence we recommend that only the identity of a small campaign’s trustees if they are incorporated or legally responsible persons if they are not, and the identity of their five largest funders. This should go some way in restoring the public’s trust that democracy is protected from automated donations.
324.The Electoral Commission should explore whether it would be feasible to create a secondary registration scheme for campaigners who would otherwise fall below current spending limits. These campaigners would only be required to register the identity of their trustees or legally responsible persons and the identity of their five largest funders. They would not be required to disclose spending. This information could then be used to improve the transparency of online imprints.
326.We received copious evidence calling for the creation of a database for political adverts. The Electoral Commission’s 2018 report called for social media platforms to keep online databases of adverts that followed the UK’s rules for elections and referendums.
327.Facebook and Google both highlighted their current attempts at advertising transparency. Facebook told us that they launched a public advert library that contains a comprehensive, searchable collection of all adverts that are active and running on Facebook and Instagram. Google highlighted their transparency report that includes a searchable election advert library. Louise Edwards from the Electoral Commission acknowledged that social media companies have stepped in and taken voluntary action in this area, which she welcomed.
328.However, both Google and Facebook’s advert libraries have been criticised as incomplete and Google’s has been criticised for containing inaccurate information. Ms Edwards told us that the current advert libraries do not capture everything that falls within the legal definition of election material. For example, she explained that the Electoral Commission wanted to see better and more reliable data on spending; some of the advert libraries have large spend brackets, which makes it difficult to understand exactly how much a particular campaign has spent. During the 2019 General Election, Google published a weekly transparency report including updates on how much money had been spent by political parties on adverts on their platform. Major inaccuracies were recorded weekly. For example, in the week the election was called Labour had a recorded spend of just £50 on adverts and nothing at all in the following week. However, Labour was advertising heavily in that period, spending tens of thousands of pounds on adverts on Google Search. Google told us that occurred because of human error, but they have not yet stated what types of quality control they conduct on their spending transparency report or why they were initially reluctant to correct the record about a party’s spending during a contested election campaign. Furthermore, the reliability of these advert libraries has been called into question. Less than 48 hours before the 2019 General Election, tens of thousands of political ads went missing from Facebook’s archive.
329.Chloe Smith MP, Minister of State at the Cabinet Office, told us that the efforts of tech companies have shown that the sector is capable of creating transparency to voters itself, and that she was not persuaded the Electoral Commission needed more powers in this area.
330.Whilst we welcome voluntary efforts made on the part of technology companies, unlike the Government we do not believe it appropriate that these companies should be setting the criteria for transparency. Alex Krasodomski-Jones from Demos told us, “we do not know what we do not know, and that is not acceptable. I cannot accept the idea that Facebook, Google and Twitter can tell us what an advert library or transparency looks like.” Leaving the companies to establish transparency for elections is at best naïve, and at worst complacent; the examples above establish why we cannot rely on technology companies to correctly design, update and police these advert libraries.
331.Mozilla have suggested that platforms should create a database of all adverts on their platforms. This would ensure that the public does not have to rely on the platform’s definition of what a political advert is and reduces the ways in which bad actors can exploit the system. Mozilla have also provided a list of criteria that these databases should adhere to. We endorse this list and encourage future regulation of online advertising databases to use these principles as a basis.
Mozilla have published five guidelines that advert archive APIs must meet in order to truly support election influence monitoring and independent research.
Comprehensive political advertising content
The APIs should include paid political adverts and issue-based adverts, without limiting access on the basis of pre-selected topics or keywords. Non-paid, public content that is generated by users who are known political content purveyors should also be available.
The content of the advertisement and information about targeting criteria, including:
Functionality to empower, not limit, research and analysis, including:
Up-to-date and historical data access, including:
The API itself and any data collected from the API should be accessible to and shareable with the general public.
332.According to these guidelines, John Lloyd from Mozilla told us that Google’s advert API met four of the five minimum standards, with important omissions. There was no information on its targeting criteria, nor does the advert API provide engagement data. Facebook fared even worse; their API met only two of the five minimum standards. He pointed out that a lot of the time, the advert archive did not work at all.
333.The LSE T3 Commission advocated a regulator to encourage the introduction of a UK political advertising directory and monitor outcomes of the initiatives of relevant institutions to ensure that databases such as Google and Facebook’s advert libraries are independently overseen.
334.The Electoral Commission in their report on the 2019 General Election recommended that in order to support trust and confidence in election campaigns, social media companies needed to provide more detailed and accurate data about campaign spending in advert libraries on their platforms, so that they and voters can see more information about who is campaigning. The Electoral Commission suggested to us that the Online Harms regime should set common standards and obligations for what the social media companies should publish in these libraries. It should also define political campaigning.
335.This responsibility should lie with Ofcom, rather than the Electoral Commission because the former look at content, whereas the latter primarily regulate political donations and expenditure by parties, rather than social media platforms. The Centre for Data Ethics and Innovation (CDEI) also recommended in their review of online targeting that the Online Harms regulator should have powers to enforce online platforms’ compliance with hosting accessible advertising archives.
336.The Government has not announced any intention to regulate these advertising libraries. The Government’s response to the DCMS Committee’s inquiry into fake news highlighted that the Online Harms White Paper proposed that the code of practice for disinformation include expectations for companies to improve the overall transparency of political advertising on their platforms. This would suggest that Ofcom would regulate the databases of political advertising. Kevin Bakhurst from Ofcom told us that if this area was within scope of its regulatory powers then advert libraries could be a powerful tool for transparency.
337.Ofcom should issue a code of practice for online advertising setting out that in order for platforms to meet their obligations under the ‘duty of care’ they must provide a comprehensive, real time and publicly accessible database of all adverts on their platform. This code of practice should make use of existing work on best practice.
338.Elections in the modern age are increasingly data driven, both online and offline. Campaigns use digital technologies to collect and analyse large amounts of personal data about individuals in order to target their messaging both through targeted online advertising and through leaflets sent through the post. Although this data analysis is not new, there is a need for far greater transparency in how campaigners obtain and use data. Polling by the CDEI found that the level of understanding of the technology that drives online targeting was low and participants were shocked at the scale and sophistication with which it takes place.
339.Under the GDPR, there must be a valid lawful basis in order to process personal data. There are six available lawful bases under Article 6 GDPR for data falling short of “special category data”: consent, contract, legal obligation, vital interests, tasks in the ‘public interest’ and legitimate interests. Section 8 of the Data Protection Act 2018 (DPA) has augmented the definition of “public interest” in Article 6 in the GDPR to include “an activity that supports or promotes democratic engagement.”
340.The GDPR provides heightened protection for certain sensitive categories of information, as “special category data” in Article 9 GDPR. ‘Political opinions’ are “special category data”. Article 9(1) GDPR prohibits the “processing of data revealing” special category data, including political opinions, unless one of ten exemptions in Article 9(2) GDPR apply. These include Article 9(2)(g), which allows processing that ‘is necessary for reasons of substantial public interest’. The DPA extends the concept of “substantial public interest” with a particular exemption for political parties to process special category data.
341.The ICO’s draft guidance on political campaigning states that the vast majority of processing for political campaigning falls under the GDPR bases of consent or legitimate interests or the DPA extension of processing for democratic engagement. However, it warns that the democratic engagement basis is often misunderstood as an overarching exemption for political campaigning activities. This use only allows the processing of personal data that is proportionate and necessary for an activity that supports or promotes democratic engagement. The ICO states that this must be more than just useful or standard practice but must be targeted and proportionate in achieving the specific purpose.
342.Bethany Shiner from the University of Middlesex suggested that there was currently ambiguity about the democratic engagement provision in the GDPR. She told us that often data processing reveals political opinions, which is special category data attracting higher protections under the GDPR and DPA. Political opinions can be inferred through combining freely given information with other datasets such as the electoral register to make a prediction about that individual’s lifestyle, habits and political views. However, Ms Shiner told us it is unclear whether the parties consider inferred data is subject to the additional safeguards. She pointed out that political parties’ privacy notices cite this justification generically, as if by existing as a political party there is automatically a ‘public interest’ justification. However, the converse is true, that the parties are processing data likely to be revealing of political opinions which in turn needs to meet the safeguards within the data protection regime. In particular, that processing must be limited to what is necessary and proportionate. The ICO make clear in their 2018 Democracy Disrupted report that inferred data is personal information and the requirements of data protection law apply to it.
343.Pascal Crowe of the Open Rights Group told us that the public needs to understand that data-driven political campaigning is not limited to the use of political ads on social media. He explained that electoral roll data was used by the political parties, mixed with sources of commercially available data, such as Experian credit history information. That information is processed and turned into a profile. He also told us that responsibility lies at the political parties’ door just as much, if not more, as it does for social media platforms.
344.We invited the Conservative Party, Labour Party and Liberal Democrats to give us evidence, as the main nationwide parties. Only the Liberal Democrats were willing to send a representative; the Conservatives offered to send in written evidence in lieu of appearing before the Committee; the Labour Party were due to send a representative but pulled out the day before the session. We then asked for written responses to the questions we had wished to ask them in person. We found their responses wanting in ways specific to each party.
345.The political parties’ understanding of the provisions of the DPA appear to differ widely and seldom conforms to the best practice suggested by the ICO. The Conservative Party cited the extensions in Section 8 of the DPA for “processing of ‘democratic engagement’ as a task in the public interest” and argued that Section 8 was not an extension at all. They further cited provisions of the DPA “for political parties to process the ‘special category data’ of political opinions.” However, the Conservative Party did not explain their position on those extensions, such as how they approach the necessity and proportionality tests in those provisions or if they considered there to be any limit on their processing activities. The Labour Party maintained that specific statutory bases for processing ‘democratic engagement’ data was in the public interest and special category data was essential to the functioning of political parties. Like the Conservative Party, Labour did not explain how they interpreted these provisions, nor if they considered there to be any limit on what they considered permissible processing. The provisions within the DPA are not a door to immunity for data users, rather, they allow for necessary and proportionate data processing for certain ends. Both parties’ stances contradict the ICO’s guidance that in most circumstances campaigners should not use special category data for political messaging without the explicit consent of the individual. Only the Liberal Democrats stated that they only use special category data with the explicit consent of the individual, however this may not be true, as the party reportedly used special category data to profile voters without their explicit consent in the run up to the 2019 UK General Election.
346.The Labour Party stated that they do not purchase any information from outside companies and that they primarily use their own data. The Liberal Democrats also claimed that they do not purchase any information from data brokers or social listening platforms and that they primarily use their own data. This directly contradicts the evidence given to us by the Open Rights Group. Furthermore, we heard from the University of Bangor that the Labour Party spent more money on Experian data in the 2017 election than the 2015 election. After the 2017 election, the Information Commissioner sent all parties a warning about the usage of such data brokers. We asked the Labour and Liberal Democrat parties to confirm that they did not use any data brokers’ services in the 2019 election and to detail the reasons for this change in operating procedures. The Liberal Democrats confirmed this and stated that they decided that using data brokers would not be GDPR compliant and therefore unsuitable for online campaigning and that social listening tools would be of limited value to their campaign objectives. The Labour Party did not reply.
348.Ailidh Callander from Privacy International told us that her organisation had written to all political parties ahead of the 2019 General Election asking for transparency in how they use data. None were forthcoming.
349.Pascal Crowe told us that regulators need to keep a closer eye on the data gathering that parties are doing outside of the regulated period. He argued that there is a need for a mechanism to hold political parties to account for their use of personal data, both for the value that might bring to their political campaign and the ways it might help them reduce their spending in other areas of their campaign. The Open Rights Group suggested that the Electoral Commission and ICO jointly conduct data audits of parties prior to the regulated period. These audits would include a financial assessment of all data assets that are to be used in a campaign. What is difficult to ascertain is the value of the data that political parties hold. Louise Edwards told us this is difficult to calculate.
350.The ICO’s Democracy Disrupted report, published in July 2018, called for political parties to be required to do due diligence on whether third-party data brokers received appropriate consent for their use of individuals’ data, including suggested audit powers. This followed the finding of a significant shortfall in transparency and provision of fair processing information by political parties. It also called for a statutory code of practice to be introduced governing political campaigners’ use of personal data. In August 2019 the ICO published a draft framework code of practice for consultation. This followed a previous consultation launched in October 2018 on what should be included in this code. The draft framework includes guidance for parties on using third-party data brokers and provides clarity and advice to help those processing personal data in political campaigning to comply with the law. The ICO told us that this guidance is a candidate to become a statutory code under Section 128 of the DPA and that there is now time to turn this guidance into such a code. The responsibility for this lies with Government and Parliament.
351.It is important to note that political advertising is not limited to the regulated period before an election. Political advertising should be seen as a year-round issue, with greater stringency required in election periods. The political advertising infrastructure in election periods would be more sophisticated but having data collection between election periods is required to give us a full picture of how parties use personal data.
352.When we asked the Minister about the Government’s intention to put the ICO’s code of practice onto statutory footing, we were told that the Minister had “concerns” about the code and that she would be concerned that such a code should not restrict elected representatives in being able to report back to their constituents. This was an unconstructive response and the Government should work with the ICO to address their concerns. The framework provides guidance to parties processing personal data to comply with the law, which as we have seen in our evidence, is currently interpreted in various ways by the main political parties. The Government must take this seriously and should not be complacent with the use of people’s data.
405 Electoral Commission, Report: Digital campaigning – increasing transparency for voters (August 2019): [accessed 2 June 2020]
406 Committee on Standards in Public Life ‘Review of electoral regulation – public consultation’
(8 June 2020): [accessed 9 June 2020]
407 Written evidence from the LSE T3 Commission ()
408 The Law Commission, Electoral Law: A joint final report (March 2020) pp 17: [accessed 27 April 2020]
409 (Craig Westwood)
410 (Louise Edwards)
411 The Law Commission, , p 14
412 (Chloe Smith MP)
413 The Electoral Commission, ‘Response to UK Government policy consultation: Protecting the debate’ (August 2019): ( [accessed 27 April 2020]
414 The Electoral Commission, ‘UK Parliamentary General Election 2019’ (April 2020) p 10: [accessed 27 April 2020]
415 Written evidence from the Network for Media and Persuasive Communication at Bangor University ()
416 Written evidence from the Institute for Practitioners in Advertising (), Doteveryone (), Full Fact (), the Coalition for Reform in Political Advertising ()
417 Written evidence from the Royal Statistical Society ()
418 Written evidence from MySociety ()
419 Written evidence from the Conservative Party ()
420 Written evidence from Mr Roy Jakes (), Dr Ana Langer and Dr Luke Temple (), Avaaz ()
421 The Law Commission, , p 157
422 DCMS Committee, (Eighth report, Session 2017–19, HC 1791)
423 The Committee on Standards in Public Life, ‘Intimidation in Public Life: A Review by the Committee on Standards in Public Life’ (December 2017) p 61: [accessed 28 April]
424 Oxford Technology and Elections Commission, ‘Ready to vote: Elections, Technology & Political Campaigning in the United Kingdom’ (October 2019) p 9: [accessed 27 April 2020]
425 Supplementary written evidence from HM Government ()
426 The Electoral Commission, Overview of referendum campaigning - referendum on independence for Scotland (2014) p 8: [accessed 18 June 2020]
427 The Electoral Commission, ‘Response to UK Government policy consultation: Protecting the debate’ (August 2019): [accessed 27 April 2020]
428 (Baroness O’Neill of Bengarve)
429 (Louise Edwards)
430 Written evidence from the Electoral Commission ()
431 (Craig Westwood)
432 (Louise Edwards)
433 The Electoral Commission, , pp 12–3:
434 Written evidence from Christina Eager ()
435 Written evidence from John Brace ()
436 The Electoral Commission, Digital campaigning: Increasing transparency for voters (June 2018) p 12: [accessed 28 April 2020]
437 Written evidence from Dr Luke Temple and Dr Ana Langer ()
438 ‘Tories spent £18.5m on election that cost them majority’, The Guardian (19 March 2018): [accessed 13 May 2020]
439 The Electoral Commission, , p 22
440 Written evidence from Dr Robert Dover ()
441 Written evidence from the LSE T3 Commission ()
442 ICO, ‘Penalties’: [accessed 28 May 2020]
443 ICO, ‘Data Protection Act 1998’: [accessed 28 May 2020]
444 Written evidence from the Electoral Commission ()
445 (Elizabeth Denham)
446 Political Parties, Elections and Referendums Act 2000,
447 Written evidence from the Open Rights Group ()
448 (Dr Martin Moore)
449 Written evidence by Dr Bethany Shiner ()
450 Written evidence from Dr Luke Temple and Dr Ana Langer ()
451 Written evidence from the LSE T3 Commission ()
452 The Law Commission, , p 156
453 (Louise Edwards)
454 (Louise Edwards)
455 APPG on Electoral Campaigning Transparency, Defending our Democracy in the Digital Age (January 2020): [accessed 28 April 2020]
456 (Louise Edwards)
457 The Law Commission, , p 159
458 The Electoral Commission, ‘Overview of donations to political parties’, p 3: [accessed 28 April 2020]
459 APPG, Electoral Campaigning Transparency Democracy in the Digital Age (January 2020) p 11: [accessed 28 April 2020]
460 Written evidence by Dr Bethany Shiner ()
461 Written evidence from Open Rights Group ()
462 Written evidence from MySociety ()
463 The Electoral Commission, , p 13
464 Written evidence from Facebook ()
465 Written evidence from Google (), (Vint Cerf)
466 (Louise Edwards)
467 (Louise Edwards)
468 Supplementary written evidence from Google ()
469 ‘Google admits major underreporting of election and spend’ The Guardian (19 November 2019): [accessed 29 April 2020]
470 ‘Facebook under fire as political ads vanish from archive’ The Financial Times (10 December 2019): [accessed 29 April 2020]
471 (Chloe Smith MP)
472 (Alex Krasodomski-Jones)
473 (Jon Lloyd)
474 Written evidence from the LSE T3 Commission ()
475 The Electoral Commission, (April 2020) p 14: [accessed 27 April 2020]
476 Written evidence from the Electoral Commission ()
477 Centre for Data Ethics and Innovation,’ Review of Online Targeting: Final report and recommendations’ (February 2020) p 107: [accessed 28 April 2020]
478 DCMS Committee, (Seventh Special Report, Session 2017–19, HC 2184)
479 (Kevin Bakhurst)
480 International Institute for Democracy and Electoral Assistance, Digital Microtargeting (2018): [accessed 30 April 2020]
481 Centre for Data Ethics and Innovation Public Attitudes Towards Online Targeting (2019): [accessed 30 April 2020]
482 Oxford Technology & Elections Commission, ‘Political Campaigning: The law, the gaps and the way forward’ (October 2019) p 22–23: [accessed 26 May 2020]
483 ICO, ‘Guidance on political campaigning: Draft framework code for consultation’ (2019) p 38: [accessed 30 April 2020]
484 Written evidence from Bethany Shiner ()
485 ICO, Democracy disrupted? Personal information and political influence (July 2018) p 30: [accessed 30 April 2020]
486 (Pascal Crowe)
487 Written evidence from the Conservative Party ()
488 Written evidence from the Labour Party ()
489 ICO, ‘Guidance on political campaigning: Draft framework code for consultation’ (2019) [accessed 30 April 2020]
490 Written evidence from the Liberal Democrats ()
491 Sky News, ‘The Lib Dems are using data to profile every voter in the UK – and give you a score’ (14 October 2019): [accessed 26 May 2020]
492 (Pascal Crowe)
493 Supplementary written evidence from the Liberal Democrats ()
494 Supplementary written evidence from the Conservative Party ()
495 (Ailidh Callander)
496 (Pascal Crowe)
497 Written evidence from The Open Rights Group ()
498 The Information Commissioner’s Office, ‘Democracy disrupted? Personal information and political influence’ (July 2018) p 5: [accessed 30 April 2020]
499 (Elizabeth Denham)
500 (Chloe Smith MP)