The research focused on three case studies in Italy, the Czech Republic, and the Netherlands. Throughout these three case studies,  researchers examined the extent to which Facebook, Google and Twitter fulfilled their commitments outlined in the Code of Practice on Online Disinformation regarding enhanced transparency of digital political advertising, in the context of the 2019 European Parliament elections. The researchers interviewed key stakeholders from political parties, civil society, national regulators and digital platforms to assess the extent of meaningful transparency, and the interplay between the EU Code of Practice and national legislation. Following this, the researchers submitted a series of recommendations to the European Commission, the EU Member States, and held multiple multi-stakeholder policy dialogues in Brussels and Member State capitals

Our research identified the following issues with regards to increasing the transparency of political ads:

1. Defining political ads: While establishing a unanimous definition of what classifies as a political ad is difficult, the lack of a clear definition within the Code of Practice was detrimental to its application. Without a clear definition, tech companies were forced to come up with their own definitions of political ads, which resulted in inconsistencies through platforms on what was labelled as a political ad and was subject to scrutiny, and what bypassed the verification systems set in place by the tech platforms to verify content and transparency. Particularly affected were issue ads (e.g. campaigns on environmental issues, migration, encouraging voter turnout) which were either completely overlooked or intensely scrutinized.

2. Ads labelling and verification of advertisers: The ad labelling system provides insufficient information on the reasons for targeting users and the data used to make these decisions. Meanwhile, the verification procedure for political advertisers is overly burdensome and lengthy, and the platforms could not deliver in verifying all advertisers who had embarked on the verification process in time for the European Parliament elections. Particularly, the verification system failed to account for third-parties, proxy advertisers and social media influencers, thereby failing to provide transparency on such political advertising practices and the advertisers themselves. Through this, it failed to identify and avoid inauthentic behaviour through both verified accounts and fake accounts.

3. Ad Libraries & Application Programming Interfaces (APIs): Ad libraries were created by the three platforms as the main transparency tool that was newly rolled out for the EU Member States ahead of the European Parliament elections in 2019. These ad libraries were supposed to provide a user-friendly overview of all political and issue ads. However, the different ad libraries showed a number of shared as well as platform-specific shortcomings. A major shared shortcoming is the lack of meaningful, complete and accurate data provided in the libraries, including targeting criteria, data on the intended and reached audiences, exact spend, ad performance, and targeting mechanisms. These libraries were also found to be missing content depending on the time and location of the user, and as there is no comprehensive repository of all paid content, which makes it impossible to verify the accuracy, completeness and consistency of political ads libraries. Tech companies were found to be reticent to allocate enough resources to ad labelling. The platforms failed to put in place adequate control mechanisms, with insufficient manpower allocated to issues such as identifying political (or issue) ads that were not labelled as such and political advertisers who had not been verified.

4. Policy Context at National Level: Our research revealed that National Regulators tend to have little to no guidelines for online political advertising. Because of this, the Code of Practice taken for the European Parliament elections in 2019 was significant in the sense that it raised awareness on the lack of regulation regarding party finance and political advertising transparency. Researchers also found that national regulatory authorities generally lacked the sanctioning mechanisms and mandates to hold political parties and platforms accountable to the transparency measures outlined in the Code of Practice.

5. Policy changes by the platforms since the European Parliament elections: Following the EU Parliament election, tech companies, under public pressure, took some initiatives to improve transparency. Facebook adapted some transparency tools, expanded the ad library to new countries, and is creating an oversight board. Meanwhile, Google limited targeting of political ads and Twitter decided to apply a ban on all political ads.

Recommendations

Considering the above findings and recognizing the importance of transparency within the digital sphere, in order to protect democracy, we recommend the following:

1. Co-regulatory approach: The EU and EU Member States need to develop a uniform, clear co-regulatory framework – beyond self-regulation – for digital political ads. The framework shall apply equally to all dominant platforms and set the rules for meaningful transparency and provide accountability and oversight mechanisms;

2. Transparency of all ads: The EU and the EU Member States need to make complete ad repositories, including political and commercial ads, a requirement for dominant platforms (such as Facebook, Google and Twitter). By expanding transparency requirements to all ads, such an approach would overcome the problem of defining ‘political ads’ and ‘political issues’. These libraries must meet a set standard such as the same level of granularity in information on targeting criteria and practices as advertisers see, including the data source, inferred profile, lookalike audiences, custom audiences, and A/B testing practices;

3. An additional layer of transparency for political ads: All political ads should have a  clear ad labelling and a verification procedure for advertisers. For this, the network of European electoral management bodies needs to jointly agree on a basic definition of political advertisements that the dominant platforms can use. An actor-centric approach would be sufficient for this additional level of transparency. This additional labelling system would also allow the complete ad library to be queried by political ads only, so users have an overview of political ads in their country.

4. European platform regulator: Appoint an independent European platform regulator that would hold the platforms accountable to their requirements for enhanced transparency, in close cooperation with the network of national electoral regulators regarding issues of political advertising.

5.Cooperation between national regulators: National electoral regulators remain essential for enforcing political parties’ transparency requirements and campaigning regulations, both online and offline. National electoral regulators need to cooperate with the dominant platforms for ensuring online enforcement of national electoral and campaigning legislation, with close support from the European network of national electoral regulators as well as the European platform regulator. The European Commission needs to continue to invest in the cooperation and support of the network of national electoral regulators, including with capacity building on digital advertising, mediation between national regulators and the platforms, and support in adapting national legislation to the digital era.

EPD believes that these recommendations, while not exhaustive, will bring us closer to the goal of increasing transparency in online political campaigning. Through close cooperation between the EU Commission, National Electoral Regulators and tech platforms we can create a system that ensures transparent political advertising and levels the playing field for all political actors.