Roberts Was Wrong to Apply the Canon of Constitutional Avoidance to the Mandate

• July 11, 2012 • 8:36 am

In his ACA decision, Chief Justice Roberts concludes that the individual mandate can be sustained as an exercise of the taxing power, even though he concedes that it is read “more naturally as a command to buy insurance than as a tax.”  His key move is an application of the canon of constitutional avoidance, which he explains as follows: “The text of a statute can sometimes have more than one possible meaning. To take a familiar example, a law that reads ‘no vehicles in the park’ might, or might not, ban bicycles in the park. And it is well established that if a statute has two possible meanings, one of which violates the Constitution, courts should adopt the meaning that does not do so.”

The Chief Justice’s account of the canon is accurate, but his application of it is exceedingly odd. The canon applies when a statute is ambiguous—when it can plausibly be read to mean two different things.  In the Chief Justice’s example, “vehicle” could be read to mean something like “device for transport,” which would include bicycles, or it could be read to mean “motorized device for transport” which would exclude them.  If including bicycles would violate the Constitution, then the canon of constitutional avoidance would counsel adopting an interpretation of “vehicle” that excludes them.  The statute includes an ambiguity about quite what it forbids, and the canon helps to resolve the ambiguity.

But the individual mandate is not ambiguous in the same sense that “vehicle” is ambiguous.  It is crystal clear what the mandate requires: get insurance or pay a certain amount to the IRS.  There is no alternative interpretation.  There is no marginal case, like bicycles, that is up for debate.  Whether it is called a “penalty” or a “tax”, the individual mandate has exactly the same legal consequences. Compare Oliver Wendell Holmes, Jr. in “The Path of the Law”:

If you want to know the law and nothing else, you must look at it as a bad man, who cares only for the material consequences which such knowledge enables him to predict . . . . [And] from his point of view, what is the difference between being fined and taxed a certain sum for doing a certain thing? That his point of view is the test of legal principles is shown by the many discussions which have arisen in the courts on the very question whether a given statutory liability is a penalty or a tax.

Whether the individual mandate is called a “penalty” or a “tax,” it means what it means and requires what it requires.

So when the Chief Justice “interprets” the mandate to be a tax, this is not interpretation in the ordinary sense.  This is, in reality, an issue, not of statutory interpretation, but of constitutional characterization.  The Chief is not interpreting the meaning of a statute to avoid a constitutional problem.  Instead he is characterizing a statute—whose meaning is not in doubt—to be a tax for purposes of the Constitution.  See, e.g., NFIB v. Sebelius at *24 (framing the issue as whether the individual mandate “may be viewed as an exercise of Congress’s taxing power”) [emphasis added]; id. at *25 (“the shared responsibility payment may for constitutional purposes be considered a tax”) [emphasis added]; id. at *30 (individual mandate “may reasonably be characterized as a tax”) [emphasis added]. The canon of constitutional avoidance only applies when a statute is ambiguous—that is, when it is unclear what rights and duties the statute creates.  Since there is no such ambiguity in the individual mandate, the canon does not apply.

This is not a mere technical objection to the Chief Justice’s reasoning.  By his own account, the canon of constitutional avoidance plays a crucial—indeed an outcome-determinative—role in his analysis. The Chief Justice acknowledges that the individual mandate most naturally reads as a penalty, not a tax.  Absent the canon, that would be the end of the matter. Since the individual mandate could not be sustained as a penalty, it should have been struck down.


  • Orin Kerr says:

    Nick, interesting post. I’m curious, though: What is your response to Mark Tushnet’s view that there are actually two different versions of the avoidance canon? Tushnet writes:

    There is a “canon” of statutory construction known as the “constitutional avoidance” canon. It comes in two versions, now labeled the “classical” version and the “modern” one. On the modern version, a judge faced with a statute that, most naturally read, raises difficult constitutional questions, should adopt instead a construction — if one is fairly available — that does not raise such questions. On the modern version, then, the Chief Justice didn’t have to address the Commerce Clause question; all he needed to do was to note that the question was difficult and that construing the statute to impose a tax was an available reading.

    The classical version is different. On that version, the canon of constitutional avoidance comes into play only when the statute, given its most natural reading, would in fact be unconstitutional. Only if it would be unconstitutional can the judge search for an alternate, available construction that would make the statute constitutional.

    The Chief Justice appears to have used the classical version of the canon of constitutional avoidance in NFIB.

    Do you agree with Tushnet, and if so, does that version of the canon change your analysis?

    • Nicholas Quinn Rosenkranz says:

      Orin Kerr asked about Mark Tushnet’s account of the two different versions of the canon of constitutional avoidance. Orin: thanks for the question. I agree with Mark that there are two different versions. The first, weaker version provides that ambiguous statutes should be interpreted to avoid rendering them unconstitutional. The second, stronger version says that ambiguous statutes should be interpreted to avoid raising difficult constitutional questions. I also agree with Mark that the Chief purported to apply the first version. Finally, for the record, I agree with Judge Posner that the first version is legitimate and the second is not.

      But the point of the post is logically prior to all this. My point is that (either version of) the canon is triggered by statutory ambiguity—two possible meanings of a statute, which have different legal consequences. The individual mandate is not ambiguous in this sense. Whether it is characterized as a “penalty” or a “tax,” it has exactly the same legal effect. Since there is no ambiguity, no version of the canon applies.

  • David Welker says:

    Interesting post. I think asking the question about the rights and duties created by a statute is a very useful direction to go. However, I think the implications are the opposite of what you suggest.

    In general, the canon of constitutional avoidance should be employed unless by choosing the less “natural” meaning, you would so greatly distort the meaning of a statute that it becomes highly questionable whether Congress would have wanted to pass the statute with the meaning that is constitutionally permissible.

    In the case the definition of vehicle that includes or excludes a bicycle, you are right that the rights and duties of citizens differ, based on which interpretation is used. So, by choosing a saving construction, the Court is changing the rights and duties of citizens. This is somewhat problematic, since by applying a saving construction, the question arises regarding whether Congress would have wanted to pass a statute the creates those rights or duties at all. But in the case where the rights and duties are not even changed by choosing the saving construction, any doubt about whether Congress would have wanted to pass a statute creating the rights and duties in question vanishes.

    Think about it. One of the main objections to applying saving constructions is the concern that the Court is acting as some sort of legislature, creating rights and duties that Congress never intended. But if the right and duties of the statute are not even changed by choosing the alternative interpretation, that would address this concern, since we in fact know that Congress would have wanted to pass a statute creating the rights and duties in question.

    Since nearly everyone agrees that a statute with the word “tax” instead of the word “mandate” or “penalty” would be Constitutional, the challenge to the ACA was always something of a game of “gotcha” anyway. It seems like an awfully bad reason to throw away a statute merely because Congress used the wrong word when an identical statute creating the same rights and duties would be constitutional.

  • Mark N. says:

    The substance of this post seems fine, but what I’m missing is the argument in favor of the conclusion, that Roberts’s approach was illegitimate. It was indeed characterization for constitutional purposes, rather than statutory interpretation to resolve an ambiguity. But why is the former not permitted? Roberts makes a fairly persuasive argument that it is not only permitted, but required, citing to several cases, some over a century old, that make that argument.

    In short, he’s arguing that if Congress could’ve done something under Power A, but mistakenly believed that they could also do it under Power B, and labeled it as a Power-B exercise, judicial caution counsels that the statute should be read as an exercise of Power A, despite Congress failing to call it that— if indeed Power A is a power Congress legitimately possesses and could use in the manner at hand. That is, laws should be struck down if they were substantively outside Congress’s power to enact, i.e. not permitted under any power Congress possesses; but not because, formalistically, Congress has cited the wrong chapter and verse.

  • BDG says:

    Here’s what the “Brief of Constitutional and Tax Law Professors” — one of the three briefs to advance the avoidance/taxing power argument — says on this issue. (This is the one signed by Balkin, Kleinbard, Metzger, Morrison, and this commenter.) Amici agreed with you, Nick, and asserted that the MECP is straightforwardly a tax under the court’s functionalist approach established in earlier taxing-power cases. But they also noted that if, for the sake of argument, one were to accept the claims of plaintiffs that congress’s choice of how to characterize the provision actually mattered, then the Court must interpret congress’ characterization according to familiar principles, including the avoidance canon.

    So I take it that the Court implicitly is assuming that Congress’ description of a provision can affect its constitutional status. I think that’s wrong, but it’s hardly a settled question — we were not aware of any prior decision even addressing it.

  • Asher Steinberg says:

    “But the individual mandate is not ambiguous in the same sense that “vehicle” is ambiguous. It is crystal clear what the mandate requires: get insurance or pay a certain amount to the IRS… Whether it is called a “penalty” or a “tax”, the individual mandate has exactly the same legal consequences.”

    I think that’s half-right and half-wrong. It’s true that however you construe the mandate, it has the same legal consequences. But what is ambiguous is whether people are acting lawfully or unlawfully if they don’t obtain coverage and pay the payment. No legal consequence turns on that question, but not everyone is a Holmesian bad man; whether something’s legal or illegal matters to a lot of people independently of sanctions. So there’s an ambiguity of statutory interpretation, even before you get to the constitutional tax vs. penalty question: is this payment a penalty for an explicitly unlawful action, or just a conditional tax incentive? That was the hard tax question in the case; on its face the statute says one must obtain coverage, but it’s possible to read the statute otherwise. After Roberts decides, invoking avoidance, that failure to obtain coverage isn’t illegal, there’s a related but distinct constitutional question – whether the payment’s a revenue-raising tax or a de facto penalty. That is, even after Roberts decided the payment wasn’t a penalty for an illegal act – a statutory question – he still could have held that the payment was so high as to completely suppress the condition for its payment and raise no revenue, taking a statutory tax out of the scope of the tax power. Of course, that was a very easy question in this case; the payment was nothing like the tax in the Child Labor Cases. It was relatively low and is expected to raise a lot of revenue. Your analysis misses the first question, sees only the second – to which I agree constitutional avoidance can’t apply – and then conflates the second question with the first.

  • John David Ohlendorf says:

    I think Prof. Rosenkranz’s initial post is largely correct, and I’d like to offer two lines of defense.

    First, with respect Prof. Tushnet’s distinction between the classical and modern versions of the avoidance canon, which Prof. Kerr raises. I don’t think Tushnet’s distinction cuts against Rosenkranz’s point. As I see it, the distinction between two versions of the canon runs something like this: under the classical version, a court actually had to find the ultimately-rejected interpretation to be unconstitutional before considering the saving construction; under the modern version, a court needn’t take this preliminary step: instead, it can just adopt a saving construction from the outset, avoiding any constitutional “doubts”. In one version, the court is avoiding actual unconstitutionality; in the other, it’s avoiding constitutional doubt. But in both versions, the ultimately-adopted construction has to be a “fair” or “reasonable” interpretation of the statute, by ordinary interpretive benchmarks. The stringency of this requirement varies, compare Presser v. Illinois, 116 U.S. 252, 269 (1886), with Knights Templars’ & Masons’ Life Indem. Co. v. Jarman, 187 U.S. 197, 205 (1902), but I don’t think the modern and classical versions of the canon differ in the degree of interpretive freedom they allow the court.

    Second, with respect to several suggestions in the comments that Roberts’s saving construction of the ACA is reasonable, I think the important point is not that Roberts stretched the statute past the limits of plausibility in NFIB. I personally don’t find Roberts’s interpretation very plausible, but it’s not the first time courts have advanced a saving construction that doesn’t pass the laugh test, nor is it likely the most egregious. Instead, the oddity here comes from the fact that Roberts admits the interpretation is a stretch-and-a-half. In his attempt to transform his commerce clause discussion from dicta into holding, he makes a big deal out of the fact that the provision in question is most plausibly interpreted as a mandate, and it is only because of the unconstitutionality of this interpretation that resort to the “tax” interpretation is justified. This, I think, is truly odd; few cases acknowledge this forthrightly that they are adopting a saving construction that is so clearly an inferior interpretation of congressional intent. Moreover, the difference matters: the avoidance canon is easily justified when we’re talking about choosing between two equally plausible interpretations, one constitutional and one not. But the strong version of the canon applied by Roberts in NFIB raises in spades doubts about the canon’s normative justification, as well as fears that this type of avoidance is faux rather than genuine judicial modesty.

  • Christopher J. Schmidt says:

    To build on Mr. Ohlendorf’s post, the danger of finding any constitutional provision to save the constitutionality of a law is that the provision cited as authority for upholding the law can be interpreted beyond its meaning. Here, the power to tax is generally meant to tax an action, such as earning income, purchasing a good, or transferring property. Moreover, taxes are usually imposed and collected when the taxable action occurs, not as a penalty for inaction. For example, it would be odd to say, “If you do not stop your vehicle at the stop sign, you will pay a tax.” Once the provision cited as support for the saved law is expansively interpreted, subsequent legislation can be passed and upheld under the expansive meaning. So, “avoiding” a constitutional question by saving a law can cause as many constitutional problems as were originally attempted to be avoided.