Defining 'political activity' for the internet: Learning from the IRS

Defining 'political activity' for the internet: Learning from the IRS
© Hill illustration

Social media giants Facebook, Twitter and Google have been struggling to develop policies regarding political ads. Facebook has taken the position that it will not fact check political ads. However, it recently announced that it is considering increasing the minimum number of people who can be targeted by specific political ads from a hundred to a few thousand. Twitter has decided not only to end political ads, but also to impose restrictions on cause-related advertising. Google has determined to stop allowing highly targeted political advertising, but to continue to fact check such ads, as part of a general policy forbidding misleading advertising.

All of these media giants have encountered strong objections to their announced policies. Furthermore, the inconsistency of the different policies itself creates challenges. In our politically charged atmosphere, both the attempts to adopt policies and unhappiness with them are understandable. The media platforms, however, irresponsibly underestimated the difficulties of formulating workable rules regarding political activity.

The experience of the IRS demonstrates the problems such attempts face.


The Internal Revenue Code prohibits tax-exempt section 501(c)(3) charities from supporting or opposing candidates for elected office. Public charities can engage in a limited amount of lobbying — seeking to influence legislation or legislators. Private foundations cannot lobby at all. Section 501(c)(4) social welfare organizations can engage in lobbying without limit and engage in campaign intervention so long as such is not their primary activity.

Tax regulations applicable to private foundations provide exceptions to forbidden lobbying. These include nonpartisan analysis, study, and research; technical advice or assistance, and examinations or discussions of broad social, economic and similar problems. Social media would benefit from adopting these exceptions and look to IRS interpretation of them.

The IRS, however, has struggled to give meaningful guidance about how to define campaign intervention for both these categories of exempt organizations. One important piece of official guidance from 2004 offers these criteria as key factors in distinguishing issue advocacy from campaign intervention:

  • a) The communication identifies a candidate for public office;
  • b) The timing of the communication coincides with an electoral campaign;
  • c) The communication targets voters in a particular election;
  • d) The communication identifies that candidate’s position on the public policy issue that is the subject of the communication;
  • e) The position of the candidate on the public policy issue has been raised as distinguishing the candidate from others in the campaign, either in the communication itself or in other public communications; and
  • f) The communication is not part of an ongoing series of substantially similar advocacy communications by the organization on the same issue.

Thus, the IRS has treated targeting voters as an important factor for many years. Social media platforms, however, can now target specific groups of voters — by gender, by race, by location — in ways not imagined in 2004. In the internet age, the beginning of an electoral campaign is difficult to identify. Deciding whether the position of the candidate distinguishes the candidate from others grants enormous discretion to government employees. More generally, “facts and circumstances” rules are inherently uncertain in application.

Dissatisfaction with such a facts and circumstances test, as well as the Tea Party controversy, led the IRS to try to define campaign intervention more precisely. It issued proposed regulations in 2013 to govern section 501(c)(4) social welfare organizations in this regard. The proposed section 501(c)(4) regulations introduced a new term, “candidate-related political activity,” and defined it broadly. For example, candidate-related activity under the proposed regulations included voter registration and get-out-the vote drives, even if conducted in a non-partisan and unbiased manner, which would have been a significant change in position. The new category also encompassed preparation or distribution of a voter guide that that referred to a candidate or a political party. 


The IRS received over 150,000 comments on these proposed regulations, most opposing them. The IRS then withdrew them. Congress also took action. Since 2015, it has barred the IRS from using any appropriated funds to define political activity under section 501(c)(4). Thus, the IRS has been unable to give clearer guidance about the meaning of political activity.

The IRS experience teaches that social media giants waited far too long to come to grips with this issue. A problem this fraught cannot be resolved so close to a presidential election. By not confronting the issue sooner, these tech companies have once again failed the public. The consequences not only for political activity, but also for public discourse generally could prove great enough to prompt Congress, after the 2020 election, to act — despite its enormous reluctance to regulate either political activity or social media.

Ellen P. Aprill is the John E. Anderson Chair in Tax Law at LMU Loyola Law School in Los Angeles, where she founded the Western Conference on Tax Exempt Organizations. LMU will serve as the venue for the Dec. 19 Democratic Presidential Debate.