Among the three major online platforms, none can agree on how to deal with political ads: Twitter bans them, Google limits microtargeting, and Facebook allows most paid political content. However different and flawed these approaches are, they show that U.S. tech companies have recognized that paid political communication on their platforms carries certain risks.
This recognition was driven by the experience of the 2016 U.S. presidential election, when negative, divisive ads were not only seen on television but also targeted to millions of social media users. More spectacularly, Russian operatives used platform ads to spread their message among American voters, as detailed in Special Counsel Robert Mueller’s report on the Kremlin’s meddling. U.S. policymakers reacted by proposing the Honest Ads Act, which would subject online ads to the same restrictions in place for radio and television advertisements. Around the world, legislators are either implementing or considering reforms to update digital political advertising.
Europe, in particular, seems to be keen to replace Silicon Valley’s voluntary measures with some form of legal guidelines. The European Commission has put transparency of online ads on the legislative agenda, Ireland is debating updates to its electoral law, and Germany is enacting new media regulation that for the first time seeks to define rules for big social media networks, video portals, and search engines. But the current regulations under consideration fail to address the unique threats posed by digital political advertising, and European regulators need to update their proposals to more squarely address them.
Paid political messaging on big ad platforms such as Facebook, Instagram, and YouTube already plays a major role in campaigns in Europe. In Germany, political parties alone spent an estimated 1.5 million euros on Facebook and Google ads during the 2019 European Parliament elections. Granted, in U.S. terms, this laughable amount represents maybe a week’s expenditures for a major presidential candidate, but the roughly 80,000 ads purchased with that money were still seen millions of times. In the United Kingdom, third-party groups outspent the country’s political parties on Facebook ads during the 2019 general election campaign, which was marred by negative and divisive advertising. In the Netherlands, one political party used a massive spending spree on Facebook ads to become the country’s largest political party by membership, which resulted in the party receiving considerable sums in state subsidies.
Why online political advertising carries unique risks not seen offline
Spending money to reach voters online is not a bad thing. In fact, in a country like Germany, online voter outreach can be seen as a welcome innovation over the past couple of years. Political parties and candidates have long relied solely on communicating with voters at campaign rallies, and via speeches and press interviews, and they can now address a larger and potentially more diverse set of people online. But the benefits of online political advertising can be eclipsed if lawmakers do not address the associated risks. These risks are not unique to the United States and its experience with foreign interference from Russia. Rather, the risks are unique to the way platform advertising functions and is different from offline advertising.
For one, online advertising allows rich campaigns to flood user feeds with paid messages, drowning out other, financially weaker political voices. In a commercial setting, this might be brushed off as the market at work. In the political arena, it represents discrimination in favor of moneyed advertisers who gain name recognition and ensure that their ideas and opinions are discussed online and in the media. Former New York City Mayor Mike Bloomberg spent more than a half a billion dollars on political advertising during his failed presidential bid, with a significant chunk of that money going toward digital ads. In the first six weeks of 2020, 1.6 billion of the 2.4 billion presidential campaign ads shown to American Facebook users were Bloomberg ads. Though his bid failed, Bloomberg used his wealth to ensure that he received huge public and media attention.
Digital advertising allows political candidates to skirt the boundaries of and regulations governing traditional advertising. Offline, there are physical limitations as to how much print space and airtime ads can take up, constraining the ability of candidates to flood the zone with their messaging. In Germany, media regulation limits the number of political ads on radio and television in an effort to create a level playing field for political campaigns.
Online advertising offers political campaigns far more tools to narrowly target their audiences, and this allows them to use their campaign funds to polarize the electorate more easily. Advertisers can target people based on vast troves of personal behavioral data, speaking to their hopes and fears, and reinforcing their identity. This can also happen offline, but the local paper and television station simply lack the personal behavioral data of millions of users and the algorithmic delivery systems to tailor messages exactly to those homogeneous groups most likely to engage with them. With a data-driven, algorithmically delivered ad campaign, there is no need to appeal to a pluralistic, heterogeneous populace (or even to make opposing promises to different groups). Advertisers can just hammer home the message that certain voter segments want to hear and bank on those folks turning out to vote. This is what happened during the 2016 election, when the Trump campaign invested heavily in Facebook ads to target narrow slices of the electorate, as Steven Levy reports in his book.
These risks of polarization and flooding digital platforms with ads are exacerbated by the fact that it is hard for citizens, researchers, and independent oversight bodies to even grasp the volume and nature of political advertising on big platforms. There is little opportunity for independent research, as platforms’ transparency tools, such as ad archives and ad disclaimers, remain flawed. The numbers on German ad expenditures mentioned earlier are just estimates and do not allow any deeper analysis of how ads were targeted and delivered, what groups saw (and did not see) paid messages, and how users engaged with those messages. While such statistics are hard to come by for offline ads as well, advertising on broadcasting airwaves and in public along the street or in the front yard are, as their names suggest, rather broad and public, allowing for easier public-interest scrutiny.
Enhancing platform and advertiser accountability to safeguard political debates online
With the EU and individual European countries now seeking to establish guidelines on political online advertising, it is important for lawmakers to keep in mind the unique features and risks of digital advertising and draft rules to address them. So far, the proposed rules fail to address the risks. For example, the reform of German media regulations set to enter into force in the fall requires enhanced disclosures for political ads compared to commercial ads. Yet there are no detailed guidelines on data disclosures regarding algorithmic ad targeting and delivery to reveal who is seeing a given ad. On the EU level, it is not yet clear what exactly the Commission has in mind when demanding “more transparency” for political online advertising.
As legislators move to regulate digital political advertising, these are some concrete measures that European legislators should consider:
- Limit behavioral microtargeting. Political advertisers should only be allowed to target their paid messages based on broad demographic data, such as age, location and gender, as the chair of the Federal Election Commission (FEC) in the U.S. has argued. Behavioral or inferred data should not come into play. Crucially, these restrictions should apply to the actual delivery of ads, done by platform algorithms, as well—and not just the targeting done by advertisers. The idea behind such a limitation is that ads will have to be used to address bigger, more heterogeneous voter groups, allowing public-interest scrutiny of advertisements and for advertisements to be rebutted.
- Mandate transparency tools and reporting. Some in the European Parliament are already proposing mandatory ad archives, and this should be a minimum requirement, as the voluntary measures that many big companies agreed to ahead of the 2019 elections have not been sufficient. In addition, corporate transparency reports already being published on a voluntary basis by some tech companies should be expanded and made mandatory. Meaningful mandatory reporting would allow citizens—but mainly researchers and journalists—to scrutinize and check corporate compliance with their own ad targeting and ad delivery policies.
- Establish an EU-wide “fair play pledge” for digital political campaigning. It is crucial to address the “supply side” of platform ads, meaning the algorithmic delivery on big ad platforms. But political advertisers, i.e. the “demand side”, have a responsibility, too. Unfortunately, legislators have few incentives to curtail their own options for voter outreach. In Germany, political parties even failed to agree on a voluntary, cross-party self-commitment. Only individual parties published pledges, promising to clearly label ads and to adhere to European data protection rules, among other things. A fair play pledge, as seen in Ireland after pressure from activists and academics, could include a commitment to refrain from behavioral microtargeting, to deliver some ads without any targeting metrics and to abstain from spreading disinformation. Over the long-term, such a fair play pledge could be worked into member states’ laws.
Europe, and the United States, for that matter, cannot rely on laws made for the offline media and ad world to keep working for data-driven online campaigns. Neither should they rely on voluntary commitments by big tech companies to set the rules for how paid political communication works online. Instead, it should be legislators gathering opinions from a variety of experts and taking the lead to ensure fair digital campaigns.
Julian Jaursch is a project director working on disinformation and platform regulation topics at Stiftung Neue Verantwortung, a Berlin-based not-for-profit, non-partisan tech-policy think tank.
Facebook, Google, and Twitter provide financial support to the Brookings Institution, a nonprofit organization devoted to rigorous, independent, in-depth public policy research.