A review of political advertising online during the 2019 European Elections and establishing future regulatory requirements in Ireland

ABSTRACT

Ahead of the 2019 European Parliament elections, the lack of regulation addressing online political advertising with the potential to undermine the integrity of the electoral process was a significant concern. Within Ireland, the Standards in Public Office Commission regulates ‘ethics, electoral, state finance and lobbying legislation’, however, there are few requirements made of political parties to disclose their funding of online political advertising. As a wider EU issue, the European Commission sought to address this lack of oversight by engaging with social media platforms to agree a voluntary Code of Practice on Disinformation. Signatories to this, including Facebook, Google and Twitter, committed to ‘ensuring transparent, fair and trustworthy online campaign activities ahead of the European elections in spring 2019’ (Kirk, N., Culloty, E., Casey, E., Teeling, L., Park, K., Kearns, C. Suiter, J. (2019). ElectCheck2019: A report on political advertising online during the 2019 European elections. Institute for Future Media and Journalism. https://doi.org/10.13140/RG.2.2.18135.01449). The report here summarises the data and key findings from monitoring the levels of transparency in the political advert libraries relating to Ireland over the European Election campaign from April 18th to May 24th, 2019. The report finds that, although platforms proactively engaged with their commitments under the EU Code, inconsistencies in the companies’ approach to this result in a lack of transparency and comprehensive understanding of political and issue-based advertising online in such election campaigns.

KEYWORDS:

• Europe
• elections
• disinformation
• political advertising
• digital media

The impact of online political advertising on the electoral process

Advertising online has become an integral component to any modern political campaign (Bimber, 2014).¹ It offers an effective means of reaching audiences on a large scale whilst also facilitating the microtargeting of campaign adverts at audiences thought to be most susceptible to their messages. The potentially harmful impact of this type of communication on democratic debate has become a focus of concern in recent years, particularly in relation to elections and referenda (Matz et al., 2017; Marwick & Lewis, 2017; ADRF, 2019). Specifically, research into the 2016 US presidential campaign and the Brexit referendum of the same year revealed the online environment for political advertising as one highly susceptible to manipulation by a range of unregulated actors. Numerous examples of concerning advertising practices were revealed, questionable both in terms of volume and transparency regarding sponsorship (Borgesius et al., 2018; Risso, 2018).

An absence of regulation at the Irish state level

The election campaign environment has altered significantly over the past 20 years, but domestic regulatory practices have failed to keep pace. Currently, political advertisements on Irish television and radio are prohibited, with the exception of a system of party-political broadcasts permitted ahead of national elections (BAI, 2017). There are no regulations regarding political advertising via other forms of media in Ireland such as billboards and, more critically, online media. Ireland’s Standards in Public Office Commission (SIPO) has responsibility for regulating financial matters regarding political parties and campaigns, including political donations. It imposes spending limits and requires disclosure of some information about spending. However, as noted by the Transparency Referendum Initiative (TRef, 2018) it does not require a level of transparency that would involve a publicly available record of online adverts purchased. Additionally, SIPO regulates lobbying and maintains a register of third parties, non-political individuals or organisations that make donations for political purposes. However, it does not have the remit to address individuals or organisations who buy advertising online on behalf of political parties or deal with electoral campaign issues or matters of political policy development such as environmentalism or immigration.

There is a gulf between what activities SIPO can regulate and the emerging activities of online political activism. According to the Central Statistics Office, 79% of Irish people access the internet every day (CSO, 2019). Additionally, the Digital News Report (2019) finds that the top five social media platforms used in Ireland are owned by Google (YouTube) or Facebook (Facebook, Messenger, WhatsApp and Instagram), and 84% of the Irish electorate now use digital and social media as a source of news (Kirk et al., 2019b). The Online Advertising and Social Media (Transparency) Bill 2020, proposed legislation that could potentially address the lack of transparency in online political advertising, was not implemented, and the time for its progression lapsed (Merrion Street, 2018a, 2018b). Similarly, the establishment of a standing electoral commission, first advocated in 2015, has yet to come to fruition (Department of Housing, Local Government and Heritage, online). In September 2020, the Irish government established the Future of Media Commission to chart the course for media regulation and funding in the digital era, and a new audio-visual regulation system is to be established with a remit for digital video services. However, it will no doubt take some time to develop recommendations and appropriate infrastructure, lacunas in relation to online advertising remain, and there are still no regulatory limitations in Ireland governing political advertising online. Regulatory practices developed in the analogue era continue to be critically ineffective in the digital age.

Until recently there has been little insight into the scale, spending, and distribution of political adverts online in Ireland. A discussion paper published by the Taoiseach’s office as part of the Report of the Interdepartmental Group on the Security of Ireland’s Electoral Process and Disinformation (2019) suggested that the risk in broadcast and print media was low, as robust systems to mitigate against abuses were in place. However, it saw the risk to the integrity of the electoral process from online advertising as as being high. The paper cited risks including the lack of transparency regarding the sources of funding, microtargeting, the absence of any limits on the volume of adverts, the speed at which the online information travels, the potential for artificial manipulation via internet bots, and the absence of editorial control over content. It noted a critical gap that online political adverts are ‘not subject to direct regulation and an absence of transparency of sourcing and targeting of advertising’ (Department of AnTaoiseach, 2018). This lack of transparency presents significant risks and challenges, and could potentially undermine the integrity of the electoral process in Ireland. Projects such as the Transparent Referendum Initiative (TRef, 2018) have served to highlight the lack of regulation of political advertising and sought to address the lack of transparency during Ireland’s recent referendum on the 8th Amendment. During this campaign, Twitter declined to take any advertising regarding the Repeal Referendum, while Google and Facebook refused to take adverts from any non-Irish advertisers. Nonetheless, journalists and media monitors from the Transparent Referendum Initiative found adverts both for and against the referendum present on Facebook.

A response to these challenges, the EU Code of Practice on disinformation

The absence of any regulation of online political advertising, not just within the Irish state, but throughout the European Union, was a serious concern ahead of the European Elections in 2019, with fears that the election could be manipulated by unknown digital political advertisers.² To address the potential of online advertising to undermine the integrity of the elections and propagate the spread of disinformation, the European Commission engaged with social media and digital advertisers to agree a voluntary, self-regulatory Code of Practice on Disinformation, published in September 2018 (EU Commission, 2018, 2019).³ Signatories to this, including Facebook, Google and Twitter, committed to ‘ensuring transparent, fair and trustworthy online campaign activities ahead of the European elections in spring 2019’ (Kirk et. al., 2019a). The signatories to the Code are required to report on their progress in addressing five key issues: (1) scrutiny of advertisement placements (aimed at demonetising online purveyors of disinformation); (2) political advertising and issue-based advertising (aimed at making sure that political adverts are clearly identified by the users); (3) integrity of services (aimed at identifying and closing fake accounts and using appropriate mechanisms to signal bot-driven interactions); (4) empowering consumers (aimed at diluting the visibility of disinformation by improving users’ ability to find trustworthy content and by make it easier for them to discover and access different news sources representing alternative viewpoints); and (5) empowering the research community (aimed at granting researchers access to platforms’ data such as are necessary to continuously monitor online disinformation). To address each of these issues, social media companies agreed to develop a range of initiatives in each member state.

In response to commitments on the second issue, political advertising, Facebook, Twitter and Google provided online libraries or archives of political adverts placed on each of their platforms. Twitter provided a Transparency Center, a webpage that is accessible using a web browser (Twitter, 2019a, 2019b, 2019c). Google provided a Transparency Report, a reasonably accessible online archive that could be filtered, with additional information also provided in Comma-Separated Values (CSV) files that could be directly downloaded (Google, 2019a, 2019b). Facebook provided an Ad Library, essentially an interactive database, accessible via an Application Programming Interface (API) (Facebook, 2019a, 2019c, 2019d). Later, Facebook provided an Ad Library Report (Facebook, 2019b), but it was not accessible via an API, and displayed inconsistent information when compared to the initial Ad Library. Following the elections, Facebook provided access to a CSV report with limited fields or categories of information. The information provided by Facebook continued to be inconsistent across these archives. The archives are discussed in greater detail in the Elect Check 2019 report (Kirk et al., 2019a). The inadequacy of these archives has also been highlighted in reports by Mozilla (2019a, 2019b) and the Office of the French Ambassador for Digital Affairs (2019). Each of these reports identify substantial problems with data access and the consistency and presentation of information.

While Facebook developed definitions of issue-based advertising and implemented protocols to ensure transparency in their adverts, neither Google nor Twitter addressed issued-based advertising. As such, there is no way to establish the total volume, range or types of adverts that were in circulation among the Irish electorate during the European Election 2019 campaign. Despite the limitations to access of the necessary data to evaluate political advertising across these platforms, some information regarding the range of advertisers, advert funders and beneficiaries during the 2019 Election campaigns could be identified and analysed.

Monitoring the implementation of the EU of Practice on Disinformation in Ireland

The European Regulators Group for Audio-visual Media Services (ERGA) is the body tasked with supporting the European Commission in monitoring the implementation of its Code of Practice on Disinformation. To do so, 13 national media regulators agreed to participate in an ERGA-led project to monitor the implementation of the Code on a national basis (ERGA, 2019). Ireland’s national media regulator, the Broadcasting Authority of Ireland (BAI), commissioned the Institute for Future Media and Journalism at Dublin City University to implement monitoring of the political advertising archives provided by Facebook, Google, and Twitter, which culminated in the Elect Check (2019) report on political advertising online during the 2019 European Elections (Kirk et al., 2019a). This paper addresses some of the findings from the ElectCheck report and adds additional analysis.

Research design and methodology

Data collection

The collected data consisted of 1,554 political advertisements broken down as follows: Facebook’s Ad Library (280); Facebook’s Ad Library Report (1,091); Twitter’s Ad Transparency Centre (88); and Google’s Advertising Transparency Report (95) (Table 1).

Table 1. Overview of findings from each archive.

Source	Extraction Method	Total Adverts	Adverts Coded
Facebook Ad Library	Application Programming Interfaces	399	280
Facebook Ad Library Report	Scrape (Python script)	1,091	1,091
Twitter Transparency Centre	Scrape (Python script)	88	88
Google Transparency Report	CSV file download	95	95

Data coding

The data was manually coded by three coders. One coder was assigned to Twitter, one to Google, and all three coded adverts from the Facebook Ad Library (280) and Facebook Ad Report (1091). The three coders collectively tested 20 adverts from Facebook’s Ad Library, achieving an intercoder reliability result of 100% agreement on all categories on analysis. Each platform was manually assigned categories related to the presence or absence of given details on the adverts, these being extracted from databases via Application Programming Interfaces (APIs), Python scripts and CSV downloads.

ERGA identified a common set of nine questions to assess the extent to which each platform disclosed information to identify those funding political advertising; the volume of political advertising; the targets of advertising; and the amount spent. To address these questions, researchers had to rely on the information provided by the platforms through their advertising archives. Only six of ERGA’s questions could be answered by monitoring the four archives in question.

Specifically, these six questions were as follows:

1. What is the degree of transparency of political and issue-based advertising?

2. Is the platform adopting appropriate and efficient measures to enable users to understand why they have been targeted by a given advertisement?

3. Is the platform adopting appropriate and efficient measures to ensure that political adverts are clearly distinguishable and are readily recognisable as a paid-for communication or labelled as such?

4. Is the sponsor’s identity publicly disclosed?

5. Are the amounts spent on the political adverts publicly disclosed (at least in terms of price ranges)?

6. What progress has been made on the commitment to publicly disclose ‘issue-based advertising’? Have platforms provided a definition of issue-based adverts and complied with it?

To answer these questions the adverts were assessed for the presence of the following information: (i) whether the advert was paid for; (ii) who paid for it; (iii) if it carried a disclaimer stating that it was a political or issue-based advert; (iv) information on micro-targeting options; and (v) spending information. From this the following categories were developed for coders to assess by examining the spreadsheets extracted from online databases via the APIs, Python scripts and CSV downloads and comparing this with individual adverts (Kirk et al., 2019a).

• Disclaimers – a label on an advert that stated that it was political or issue-based.

• Payer – provision of the name of the individual or organisation that paid for an advert.

• Microtargeting – provision of details on the age, gender, and/or location of those to whom an advert was distributed.

• Sponsors – a label on an advert that clearly signified that the content promoted was paid for.

• Spend – provision of details of how much a payer spent (either in brackets or in aggregate).

• Currency – provision of details of the currency in which an advert was paid.

• Issue – a label on an advert that signifies that it might be issue-based.

The categories were labelled with a ‘1’ if the data point was present and a ‘0’ if omitted, and the results quantified to measure the level of transparency.

Findings – political advert transparency

RQ1. What is the degree of transparency regarding political and issue-based advertising?

This question was addressed by identifying whether an advert carried a disclaimer stating that it was a political advert and whether it provided information about the entity that paid for the advert. Overall, we found that Facebook and Twitter provided disclaimers in the plurality of cases, but Google did not. Facebook also made the payer clear in the plurality of cases, as did Google.

We found several discrepancies between the information extracted from Facebook’s Ad Library (API access) and the Ad Library Report (no API access). At times there were more adverts listed for an advertiser than appeared in the Ad Library. For example, more adverts were listed for advertiser the Andrew Doyle than appeared on his page. In one case, an advert from April 22nd to April 26th 2019 ran without a sponsor name and a disclaimer was not collected by either method. However, it appeared in the Ad Library archive (Table 2).

RQ2. Is the platform adopting appropriate and efficient measures to enable users to understand why they have been targeted by a given advertisement?

This question was addressed by identifying the number of adverts that carried labels on micro-targeting information. Three targeting categories were coded: geography, age, and gender. Overall, we found that micro-targeting information was available in all cases, but that it was limited to geography, gender and age. We were unable to assess whether other targeting options were available to the advertisers (such as special interests or perceived political persuasion) (Table 3)).

Table 2. All platform payer and disclaimer labels.

Source	Disclaimer Present		Payer Name Present		Total Adverts
	Yes	No	Yes	No
Facebook Ad Library	237	43	239	41	280
Facebook Ad Library Report	1,039	52	1,037	54	1,091
Twitter Ad Transparency Centre	88	0	88	0	88
Google Advertising Transparency Report	0	95	95	0	95

Table 3. Overview of labels regarding micro-targeting details (Geographic Location, Gender and Age).

Source	Geographic Details	Gender	Age	Total
Facebook Ad Library	280	280	280	280
Facebook Ad Library Report	1,091	1,091	1,091	1,091
Twitter Ad Transparency Centre	88	88	88	88
Google Ad Transparency Report	95	95	95	95

RQ3. Is the platform adopting appropriate and efficient measures to ensure that political adverts are clearly distinguishable and are readily recognisable as a paid-for communication or labelled as such?

This question was assessed by establishing the number of adverts that carried: (i) a label to say that it is an advert or sponsored (as opposed to a post, tweet or vlog, for example); (ii) the name of the sponsor; and (iii) a disclaimer stating that advert is political or issue-based. We found that all platforms were satisfactory in labelling adverts as being sponsored and the sponsor name was present in the plurality of adverts, as was a disclaimer although this was not wholly consistent (Table 4).

RQ4. Is the sponsor’s identity publicly disclosed?

This question was answered by establishing the number of adverts that carried the name of the payer. This happened in most cases, but Facebook’s record was more mixed than the other platforms. All adverts in Twitter’s Ads Transparency Centre and Google’s Transparency Report provided a named sponsor. However, Facebook’s databases were not as consistent. In the data extracted from both its Ad Library and the Ad Library Report, the results were mixed (Table 5).

RQ5. Are the amounts spent for the political adverts publicly disclosed (at least in terms of price ranges)?

This question was answered by establishing the number of adverts that carried spending information. We found that all adverts carried some spending information (Tables 6 and 7).

RQ6. What progress has been made on the commitment to publicly disclose “issue-based advertising”? Have platforms provided a definition of issue-based adverts and complied with it?

These questions were answered by establishing the number of issue-based adverts that could be identified, and whether there was a label for a specific issue. We found that only Facebook labelled any of its adverts as issue-based, and many examples of issue-based adverts that were not clearly labelled as such (Table 8).

Platform transparency of political adverts spending

In its 2019 report, SIPO highlighted the need for enhanced regulation of online advertising, suggesting that this should be the remit of an electoral commission (SIPO, 2019). Similarly, the Code of Practice on Disinformation only addresses the need for transparency regarding how much is being spent by advertisers through online platforms; it does not seek to limit spending by political actors. Again, in the absence of national regulation to compel detailed disclosure of information relating to spending by Irish political actors, the online platforms’ own advert libraries offer some insight, but not a complete picture.

Table 4. Transparency in terms of advert labelling across the archives.

Source	Labelled as Being Sponsored	Sponsor Name Present	Disclaimer Present	Total Adverts
Facebook Ad Library	280	239	237	280
Facebook Ad Library Report	1,091	1,037	1,039	1,091
Twitter Ads Transparency Centre	88	88	88	88
Google Transparency Report	95	95	0	95

Table 5. Display of payers’ identity across platforms.

	Yes	No
Facebook Ad Library	239	41
Facebook Ad Library Report	1,037	54
Google Transparency Report	95	0
Twitter Ads Transparency Centre	88	0

Source: Display of Payers’ identity.

Table 6. Provision of spending information.

	Yes	No
Facebook Ad Library	280	0
Facebook Ad Library Report	1,091	0
Google Transparency Report	95	0
Twitter Ads Transparency Centre	88	0

Source: Spending Information Present.

Table 7. Provision of details of the currency in which an advert was paid for.

Currency	Yes	No
Facebook Ad Library	280	0
Facebook Ad Library Report	1,091	0
Google Transparency Report	0	0
Twitter Ads Transparency Centre	88	0

Table 8. Issue-based advertising across platforms.

	Labelled as Issue-Based	Disclaimer Present	Disclaimer Absent	No. of Adverts labelled	Specific Issue Label
Facebook Ad Library	Either/or	237	43	Unclear	0
Facebook Ad Library Report	Either/or	1,039	52	Unclear	0
Twitter Ads Transparency Centre	No	0	NA	0	0
Google Transparency Report	No	0	NA	0	0

Google

Google had a relatively robust procedure for registering adverts commissioned by political parties and the volume of adverts targeting Ireland during the 2019 election cycle was comparatively smaller than on other platforms. Only two Irish parties,, Fianna Fail and the Green Party, registered with Google to boost their message to the public, spending a combined total of €4,650. Fianna Fáil commissioned 46 adverts, and the Green Party 22 adverts. The European Parliament also registered to campaign and spent €57,500 to encourage participation in the election. However, as also noted, Google did not label or disclose issue-based advertising and, in this regard, there is no way to establish if there was such advertising, for example on environmental issues or immigration, from political parties, actors, campaigners or supporters (Table 9).

Table 9. Spending by political actors for adverts as reported by Google.

Political Actor	Regions where Advertised	Total Adverts	Total spend (€)
Fianna Fail	EU, GB, IE	46	1,850
Green Party of Ireland	EU, IE	22	2,800
European Parliament	International	338	57,500

Twitter

In most cases, Twitter disclosed information regarding the page that an advert was run from and gave the payer’s name as well as the total spent over time. However, Twitter only publicly displayed such information for seven days and did not easily facilitate national-level analysis by filtering by region. Similarly, Twitter did not label issue-based advertising, instead noting only political actors that were registered and verified. As such, it is not possible to establish if there was advertising based on environmental or immigration issues, for example, and, if there were, who was paying for them. Over the course of the campaign we found only seven Irish political parties or candidates registered to advertise on Twitter, spending a combined total of €4,370 with the platform. However, recently Twitter announced that it would not be taking political or issue-based advertising from November 2019, citing an aversion to facilitating false political messaging and a preference for politicians to achieve organic growth in messaging through users sharing content via Tweets, Retweets and Mentions (Table 10).

Table 10. Spending of selected political actors for adverts as reported by Twitter.

Twitter User/Handle	Total Spent
Andrews4Europe	€551.30
DeirdreCluneMEP	€417.30
Fianna Fail Party	€2400
Greenparty.ie	€235.20
Hermann Kelly	Not Disclosed
SeanKellyMEP	€262.70
Sinn Fein Ireland	€766.50

Facebook

Facebook was the platform that attracted the highest level of advertising from Irish parties and candidates, the majority of which were running in the 2019 European Election. Table 11 reports the top 30 political actors that spent the most on Facebook at this time. However, in total there is estimated to be more than 150 individual political advertisers on Facebook over the course of the 2019 campaign, in both local and European elections. Facebook did not label adverts for the specific election an advertiser was addressing, and in many cases this was not clear from the advert, such as when parties advertised their political stance more generally. Among the top 30 political actors advertising on Facebook, more than €283,000 was spent, the majority coming from the Irish Referendum Commission or the European Parliament, both providing information on how to vote. Among these top 30 advertisers, more than €126,000 was spent by Irish political campaigners. Compared with Google and Twitter, the way this spending on Facebook is presented does not facilitate easy identification of total spend by sponsors, nor who of who benefited from their funding.

Table 11. Spending by political actors for adverts as reported by Facebook (Top 30 Advertisers).

	Political Actor	Handle	No. of Adverts	Total Spend (€)	Sponsor
1	Referendum Commission	@referendumcommission	29	87,563	Referendum Commission
2	European Parliament	@europeanparliament	58	56,399	European Parliament
3	European Greens	@EuropeanGreens	20	18,835	European Greens
4	Dominic Hannigan	@DominicHannigan	29	17,446	Dominic Hannigan
5	Facebook Ireland	@facebookdublin	4	13,069	Facebook Ireland
6	Luke ‘Ming’ Flanagan MEP	@Lukemingflanagan	7	10,430	Luke ‘Ming’ Flanagan MEP
7	Maria Walsh	@MariaWalshEU	16	9764	Fine Gael
8	Matt Carthy	@MattCarthySF	30	6713	Matt Carthy
9	Fianna Fáil	@fiannafail	54	6087	Fianna Fáil
10	Fine Gael	@FineGael	69	4378	Fine Gael
11	Mark Durkan	@MarkDurkanEU	12	4359	Fine Gael
12	Pro Life Campaign Ireland	@prolifecampaignireland	5	4315	Pro Life Campaign Ireland
13	Leo Varadkar TD	@leovaradkar	12	4207	Fine Gael
14	Malcolm Byrne	@malcolmbyrneMEP	30	3840	Malcolm Byrne Page
15	Deirdre Clune MEP	@deirdreclunemep	24	3784	Deirdre Clune MEP
16	Women for Election	@womenforelection	10	3327	Women for Election
17	Sean Kelly MEP	@sean.kellymep	30	3193	Sean Kelly MEP
18	Saoirse McHugh for MEP	@Saoirse2019	26	2951	Green Party Ireland
19	Frances Fitzgerald	@FitzgeraldFrances	4	2918	Fine Gael
20	European People’s Party	@epp.eu	30	2552	European People’s Party
21	Andrew Doyle	@AndrewDoyleTD	19	2313	Andrew Doyle
22	Anne Rabbitte TD	@annerabbitteTD	16	2308	Rabbitte Anne Marie
23	Labour Party	@thelabourparty	15	2290	Labour Party
24	Alex White	@AlexWhite4EU	10	2219	Anthony Eunan McKinney
25	Sheila Nunan for Europe	@SheilaforEurope	9	1724	The Labour Party
26	Cllr. Eilis Ryan for Europe	@cllreilisryan	13	1518	The Workers’ Party of Ireland
27	Barry Andrews for Europe	@Andrews4Europe	30	1210	Barry Andrews
28	Michael O’Dowd for Europe	@1michaelodowd	14	1192	Michael Joseph O’Dowd
29	Social Democrats	@socialdemocratsireland	30	1185	Social Democrats
30	Sinn Fein Ireland	@sinnfein	17	1183	Sinn Fein

Incomplete insight into buying attention during the European Elections

While the voluntary self-regulation system implemented by the EU Commission addressed the lack of national regulation to some extent, the picture of online political advertising is far from complete. In the absence of national regulation in Ireland requiring public disclosure of funding political advertising online, as well as a lack of regulation at the EU level compelling complete transparency by social media companies, the advertising libraries offer only a limited insight into political advertising online in the Irish context.

Although Facebook, Twitter and Google did engage with their commitments under the EU Code of Practice on Disinformation by providing some transparency regarding political advertising, this was incomplete, and the overall objectives underpinning these commitments were not achieved in Ireland. The level of transparency exhibited across the three platforms was mixed, and each platform adopted a different approach to classifying and labelling relevant adverts. Furthermore, of the three platforms, only Facebook labelled any adverts as issue-based, and further investigations found examples of issue-based adverts on Facebook that were not labelled. Neither Twitter nor Google made efforts to identify issue-based adverts, and there is little to no insight into the extent to which adverts were bought to address political issues relating to certain parties or candidates such such as climate change or housing. Election campaigns provide a critical juncture in the political calendar where political stances on issues like housing and health can evolve into policies and, with elections of MEPs, can contribute to the development and implementation or rejection of certain Europe-wide measures. In this regard, the absence of insights into issue-based advertising is particularly concerning. It is not clear to what extent the political issues that certain candidates are platforming are being bolstered or undermined by a range of non-political actors. While Twitter has announced a ban on all political advertising, it remains unclear how issue-based advertising and the broader political discourse on contested issues such as global warming and Covid vaccine development, would be affected and thus how these matters might be monitored in the future.

However, it must be recognised that by signing up to the Code of Practice on Disinformation and engaging with the European Commission and European Regulators Group for Audiovisual Media Services in relation to monitoring its implementation these digital platforms are volunteering to support democratic participation, electoral transparency and external oversight of their activities in this regard. The advertising archives provided by Facebook, Twitter and Google examined in this study do provide useful information for individual users regarding the particular demographic groups that are seeing certain adverts on their social media feeds (as distinct from any consistent explanation as to why adverts appeared on their own timeline), and most of the time users can see who paid for an advert. However, these archives showed significant limitations with regard to transparency for the purposes of systematic monitoring of election spending and the accuracy of advert content. This raises questions regarding their ability to fully support comprehensive monitoring initiatives.

The findings of this research indicate that the digital platforms have much room for improvement if they are to satisfactorily comply with their commitments under the EU Code of Practice on Disinformation and facilitate effective monitoring of electoral advertising. As a basic requirement, monitors should be able to quantify the scale and pervasiveness of advertising within a region, identify all of the adverts that are political or issue-based, identify the payees and beneficiaries of political adverts, and what microtargeting options were selected by political advertisers. Critically, media monitors need to be confident in the data they are analysing. In utilising the advertising archives provided by Facebook, Twitter and Google, it was not possible for the authors of this study to be wholly confident that the libraries capture and report all political and issue-based adverts. Our findings highlight the continued lack of clarity regarding the scale, nature and financing of online political advertising in Irish elections.

Future regulation at a national level

Eight months after the European Elections, Ireland called a General Election at short notice. Reviews of the political advertising for that showed that in the four weeks leading up to the poll there were more than 6,000 political adverts on Facebook and Google placed by competing candidates and parties. This involved a spend of about €32,000 was on Google and €350,000 on Facebook, with most individual adverts falling far below the €126 SIPO threshold for disclosure (Kirk, 2020). We know very little about the nature of candidate advertising for this election and digital political adverts remain a blind spot in terms of assuring electoral integrity. Social media giants can help address transparency of political advertising but so too can national governments, through electoral regulations that help us to understand the advantage bought rather than just the advantage algorithmically gained through engagement, via likes and shares, among our online communities. The limitations in online platforms’ disclosure indicates that the Irish national regulator should demand more comprehensive disclosure from Irish political organisations. While SIPO requires some disclosure regarding payments made during election campaigns, this is limited to only those above €126, and most of the adverts posted were substantially below this. Where possible, SPIO should require more detailed disclosure of a political party’s or candidate’s online spending and related activities, including a list of all digital platforms used; precise details on the spending for (including for content creation) and duration of each advert;; information on microtargeting options used to either target or exclude certain platform users; and specific information on activities at different political levels (central political party, local branch, individual candidate, etc.) across different social media platforms. By requiring more detailed information from political parties and candidates, regulation will also facilitate better identification of and the ability to address malicious actors, nationally and internationally. The digital landscape is a critical arena for political debate, and it is essential that it is a fair playing field for all stakeholders, political candidates, and citizens alike. As it stands, those with the larger budgets and more substantial organisational resources are at an advantage. This inevitably means large political parties, such as Fianna Fail and Fine Gael, that can sponsor multiple candidates across multiple, long-established social media networks, have a competitive advantage over newer parties and independent candidates who would not have established a similar level of audience reach across social and digital media.

The mediation of political campaigns in the digital era requires a substantial overhaul of national regulation and Ireland is lagging far behind comparable nations. Digital media is one of the key battlegrounds for political campaigners and political advertising and marketing is becoming increasingly sophisticated in using social media to identify and target certain electorates. During April and May 2019 in Ireland, this was apparent at a local level with the Local and Mayoral elections, the national level with the Divorce Referendum, and at an international level with the European Elections. On one hand, digital political advertising is a critical mode of communication between parties, candidates, and constituents, and provides an important platform to enable citizens and civil society to articulate their own concerns. On the other, it is an unregulated environment ripe for political exploitation, and which currently offers greater advantage to those with the deepest pockets.

Quality of information in the public sphere and the fairness of electoral practices are central to democracies and, now as ever, regulatory infrastructure is required to maintain both. Ireland’s regulator, SIPO, is in need of a substantial extension of its remit to survey the digital landscape. Also urgently required is the establishment of an Electoral Commission as outlined in the Programme for Partnership Government (2016) and recommended by the Interdepartmental Group on the Security of Ireland’s Electoral Process and Disinformation (2019). Furthermore, in Ireland, there is a pressing need to develop a media monitoring body that can combine the skills and expertise needed to address issues such as fact-checking and verification, source tracking, image and video manipulation, and emerging technologies such as blockchain. Effective monitoring in conjunction with appropriate regulation and transparency in terms of online advertising are essential in combating the spread of disinformation and promoting an informed Irish electorate.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Niamh Kirk

Niamh Kirk is a digital media and communications researcher. She received her PhD from the School of Communications, DCU and is currently the Newman Fellow in Digital Policy at the School of Information and Communication Studies, UCD.

Lauren Teeling

Lauren Teeling is a Media and Communications scholar from the School of Communications, DCU.

Notes

1 Special thanks to Edward Casey, Eileen Culloty, Kirsty Park and Jane Suiter who contributed to the ElectCheck 2019 project.

2 A new Bill, the Online Advertising and Social Media (Transparency) Bill, has not been formally debated in the Dáil since 2017.

3 This agreement is between the EU and media platforms, and does not impose any obligations on European member states.