Free Speech for Robocallers? A Preview of the Severability Issue in Barr v. American Ass'n of Political Consultants

by Michael C. Dorf

Listeners tuning in live to hear Supreme Court telephonic oral arguments today will likely be most keenly interested in the cases involving claims at the intersection of religious freedom and what's left of the Affordable Care Act's employer-mandated contraception coverage. But there's another case to be argued today, Barr v. American Ass'n of Political Consultants (AAPC), and it involves the free speech rights of robots! As longtime readers know, I favor rights for sentient robots on the same grounds that I favor animal rights, so you can imagine how excited I am to see the possibility of robots getting their due in the highest court in the land.

Okay, fine. AAPC has nothing to do with the rights of robots. It's a case about the rights of robocallers--that is to say, the human beings who want to send pre-recorded or auto-dialed messages to mobile phone users. That's right, this is a case brought by perhaps the least popular demographic in America--robocallers. After doing my best to gloss over without oversimplifying the substantive issues, I'll discuss what I regard as the most interesting aspect of the case: severability.

The plaintiffs (respondents in the Supreme Court) challenge the Telephone Consumer Protection Act ("TCPA"). The TCPA generally bans the use of automated dialing or the sending of pre-recorded messages to mobile phones. As originally enacted, the TCPA made exceptions for emergencies and for calls to people who had consented. In 2015, Congress added another exception for calls "made solely to collect a debt owed to or guaranteed by the United States." In the current case, the plaintiffs say that in light of the exception, the TCPA is impermissibly content-based.

The Solicitor General's opening brief argues that the exception is not content-based because "the applicability of the government-debt exception turns on the economic activity in which the caller is engaged." The reply brief reiterates this position, which is at least a bit counterintuitive.

Suppose you borrow $30 from a friend. Two weeks later, your friend calls you. Is this a call "made solely to collect a debt?" The answer depends on what your friend says. If she says "hey buddy, I need the $30," then yes. If she says "happy birthday" or "I'm just calling to check on how you're feeling," then no. It's frankly difficult to make sense of the government's position that the "economic activity" is independent of content when the context is a telephone call.

Not surprisingly, both the district court and the appeals court found the exception content-based. That's especially unsurprising given the breadth of the current definition of content-based, which can be found in the SCOTUS opinion in Reed v. Town of Gilbert. There, Justice Thomas wrote for the Court that a law is content-based if it "applies to particular speech because of the topic discussed or the idea or message expressed." Meanwhile, there's a further reason why the exception could trigger strict scrutiny under First Amendment doctrine: It's speaker-based. One can disagree with the holding of Citizens United v. FEC, but the Court was correct in stating there that speaker-based distinctions should generally be treated as content-based distinctions. 

Accordingly, there is a very good chance that the SCOTUS will reject the government's argument that the government-debt-collection exception to the TCPA is content-neutral. Perhaps it will even be unanimous on this point.

I also would not be at all surprised if the Court goes on to find that the exception fails strict scrutiny. The government brief argues that the exception serves the "important interest in protecting the federal fisc." However, note that in describing the interest as "important" rather than "compelling," the government implicitly concedes that the exception cannot satisfy strict scrutiny; it can at best satisfy intermediate scrutiny. Thus, if the government loses on the content-based question, by its own admission, the exception will be deemed invalid. Meanwhile, even if the government somehow persuades five justices that the exception is content-neutral, it's by no means clear that it satisfies even intermediate scrutiny. Of course the public fisc is important, but does the ability to make robocalls to debtors really do very much to protect it? And is the government situated any differently from private creditors, for whom debt collection is also important?

The foregoing analysis leads me to conclude that the government should and very likely will lose on the substantive validity of the government-debt-collection exception to the TCPA. The question then is whether that benefits the plaintiffs. That's the severability issue, which has two dimensions: (1) statutory construction; and (2) constitutional and related limits.

(1) The statutory construction piece seems like a slam-dunk against the plaintiffs. As a matter of imputed congressional intent, it seems very clear that given the choice between no TCPA at all and a TCPA without the government-debt-collection exception, Congress would prefer the latter. That, after all, was the legal regime that operated between 1991 and 2015.

The plaintiffs do not argue that Congress would prefer the elimination of the TCPA. Rather, they contend that they are not challenging the exception at all. They say they're challenging the TCPA's restriction on their own speech as content-based. Of course, what makes it content-based is that it contains exceptions, so whether one deems the challenge as aimed at the regulation of their speech or the exceptions for other speech seems a matter of semantics rather than substance. Thus, all of the force in the plaintiffs' argument comes from (2).

(2) The plaintiffs offer three main arguments, which I'll address out of order. 

(a) The plaintiffs say that eliminating the government-debt-collection exception would not cure the TCPA's unconstitutionality because its other exceptions are also content-based. I very much doubt that the Supreme Court will base its ruling on this ground, although it's possible that if the Court rejects plaintiffs' other arguments, this point would be open on remand.

(b) The plaintiffs say that whenever the Court has confronted content-based regulations of speech, it has treated the problem as the regulation, not the exception. I suppose that might be true, but it seems to me illogical. Certainly a legislature could cure an unconstitutionally under-inclusive statute by expanding its coverage. Suppose an ordinance providing "no gatherings of more than 1,000 people on the Great Lawn, except for assemblies expressing support for the mayor." That's clearly unconstitutional. It could be cured by deleting the exception or by expanding the exception to cover other kinds of expressive assemblies.

Now there are reasons why a court has less remedial freedom than does a legislature (or city council in the example above). Where the substitute provision could take a great many forms, a court will rightly hesitate to rewrite the statute in order to cure it. And sure enough, the plaintiffs make that argument in the Supreme Court. But it's a bad argument in this case, where--putting aside the objection to other exceptions noted in (a)--Congress rendered a previously permissible regulation impermissibly content-based by adding an exception. In such a case, it involves no substantial "rewriting" for the Court to say simply that the exception is invalid and to sever it from the prior law.

(c) The plaintiffs raise a further objection. If they lose on severability, they say, then the Court will have eliminated the incentive for similarly situated plaintiffs to challenge content-based laws in the future. Such "leveling down" leaves plaintiffs with no benefit. Knowing that, plaintiffs won't sue, and the government will get away with keeping unconstitutional provisions on the books.

This strikes me as a good objection, but it probably proves more than the plaintiffs intend. The leveling-down incentive problem exists in equal protection cases no less than in free speech cases. And yet the Court allows leveling down as a remedy in equal protection cases. The plaintiffs argue that free speech and equal protection cases are different, but it is hard to see why that is--especially when what makes a law impermissibly content-based is the fact that it forbids proscribable speech but doesn't forbid comparable speech by others. Put simply, the content-based branch of free speech doctrine really is very much like equal protection.

To be sure, the Court's cases sometimes are more favorable to free speech plaintiffs than to other sorts of plaintiffs when it comes to severability. Because of the so-called chilling effect, free speech plaintiffs can bring overbreadth challenges where other plaintiffs cannot. I've expressed doubts (e.g., here and here) about whether treating speech plaintiffs as special with respect to overbreadth makes sense. But even if it does, the current case is not an overbreadth case. It's what we might call an underbreadth case. There is no special status of free speech plaintiffs with respect to underbreadth. Thus, if the leveling-down incentive problem suffices for invalidation in the current case, it also ought to suffice in equal protection cases.

And maybe it does. The problem for the plaintiffs is that the Court generally has not seen matters that way. Nor is it entirely clear that it ought to. So much of our major litigation challenging the constitutionality of legislation is driven by plaintiffs whose incentives are ideological rather than more direct. That's probably why Heckler v. Mathews makes sense. In that case, the Court allowed that a plaintiff has standing to bring an equal protection challenge even if it is certain (because written into the statute) that any victory will result in leveling down. Why would anyone bring a case in those circumstances? The Court didn't give an answer in Mathews but the combination of ideological motivation for institutional plaintiffs and schadenfreude for individuals will usually suffice.