Pre-Existing Conditions, Severability, and the "When" Question in Statutory Construction

by Michael C. Dorf

In my latest Verdict column, I take aim at one of the many lies in the op-ed that appeared last week in USA Today under Donald Trump's name--the claim that the president has kept his promise to protect health insurance for people with pre-existing conditions. I explain that the claim does not pass the laugh test. Among the reasons I give is the administration's support for a pending lawsuit by Texas and nineteen other states that, if successful, would invalidate the Affordable Care Act's prohibition of screening out or charging extra for persons with pre-existing conditions.

The column describes the lawsuit's key argument in greater detail, but the very short version goes like this: (1) the ACA's individual mandate was upheld by the SCOTUS as a tax; (2) Congress eliminated the tax late last year as part of its tax cut law; (3) therefore the mandate no longer has a constitutional basis; (4) the mandate was designed by the Congress that enacted the ACA to work hand-in-glove with the protections for people with pre-existing conditions; (5) thus, a court cannot sever the mandate from those protections; and (6) accordingly, the protections are invalid.

To decide a severability question, a court must determine whether Congress would have wanted the portions of the statute that do not by themselves violate the Constitution (here the pre-existing conditions protection) to remain operative without the  invalid portion (here the mandate without the tax). I say in the column that the non-severability claim made by Texas is extremely weak, because we don't need to guess what Congress would prefer: Congress, just last year, told us what it prefers. It left the rest of the ACA intact when it reduced the tax to $0.

Nonetheless, at a hearing last month, Federal District Judge Reed O'Connor seemed sympathetic to the non-severability argument. My column notes that while that is alarming, it should not be surprising. Over the last eight years, Republican-appointed federal judges have been remarkably receptive to what we might politely call "creative" arguments that aim to destroy the ACA. But there is--or at least should be--a difference this time. While in prior challenges to the ACA, conservative legal scholars crafted some of the arguments that conservative judges accepted, to their credit, conservative legal scholars have joined with more liberal ones in arguing that Texas's non-severability argument is unpersuasive.

I am thus cautiously optimistic that the courts will ultimately reject the argument for non-severability. The issue deserves some attention, however, because there is no guarantee that the case will come out right, and it raises a question of wider importance.

Subject to some constitutional limits not relevant here, severability is a matter of statutory construction. Thus, one way to understand the dispute in Texas v. United States is as an argument over which Congress's intent matters. The argument I laid out above straightforwardly points to the most recent Congress to have addressed the issue, in 2017. But Judge O'Connor expressed sympathy for the claim by Texas that the key Congress is the one that passed the original ACA in 2010. On its face, that is a highly peculiar view. After all, a fundamental principle of statutory construction holds that in cases of conflict, the later-in-time provision prevails. Why might Judge O'Connor think otherwise?

Perhaps he thinks that the 2017 action did not alter the original ACA in a relevant way. Certainly there are circumstances in which a later-in-time law alters an earlier one but not in a way that is relevant to some case. Suppose that in 2017 the legislature in some state amends the traffic code to redefine "vehicle" to include "battery-assisted bicycles." There then arises a dispute over the meaning of a provision of the traffic code enacted in 2010. If that dispute involves a car rather than a bicycle, then insofar as legislative intent matters, it would be appropriate to look to the legislature's intent in 2010 rather than in 2017, because the 2017 amendment is simply not relevant to the case.

But of course that hypothetical example has very little to do with the actual dispute in Texas v. United States. Even assuming that Congress in 2010 would have preferred no protection for people with pre-existing conditions to protections that were not backstopped by a mandate--itself a contestable assumption--the 2017 amendment spoke much more clearly to the severability question than the original 2010 enactment did. Congress in 2017 could have chosen to eliminate both the protections for pre-existing conditions and the tax penalty for failing to obtain insurance, but it eliminated only the latter. Texas v. United States is like a case about battery-assisted bicycles, not cars.

Readers might find amusing the fact that in arguing for the ACA's invalidity, Texas relies on the supposed intentions of the Democratic Congress that passed it in 2010 rather than the intentions of the Republican Congress that effectively repealed the mandate in 2017. That's not just ironic; it's telling. Although nearly everyone in the 2017 Congress who voted for the legislation that reset the tax penalty to $0 would have wanted (and likely still wants) the rest of the ACA to be repealed along with it, Republicans could not quite garner a simple majority in the Senate to repeal all the portions of the ACA that could be repealed under the reconciliation procedure, much less the 60 votes that would be needed to repeal those portions that could not be eliminated using reconciliation. Texas and perhaps Judge O'Connor must look to the imagined intentions of the 2010 Democratic Congress because, strange as it may sound, the actual intentions of the 2017 Congress would appear to preserve the pre-existing conditions protections of the ACA.

Meanwhile, we might problematize the whole notion of asking what Congress would have intended. After all, there is a serious tension between the argument Texas makes for non-severability and the approach that judges--especially conservative judges--tend to follow.

Textualists have long argued against being guided by the subjective intentions of lawmakers as opposed to the texts they enacted. Textualism is often defended in opposition to intentionalism, but I have argued (as have others) that for some time now the chief rival of textualism is purposivism, not intentionalism. Purposivist judges share textualist judges' skepticism of subjective legislative intentions, but when text is not clear purposivists look for guidance in the broad purposes that can reasonably be ascribed to the legislature. We need not rehash the debate between textualism and purposivism here. The key point is that under either of these two leading approaches, a judge would reject as hopelessly indeterminate the effort to discover what the legislature would have intended if it had been told that part of a statute it enacted was invalid.

Finally, I would note that the problem presented in Texas v. United States is extraordinarily common. Severability issues arise or can arise in just about every constitutional case. Given the frequency with which legislatures amend statutes, there will then be some need to harmonize or prioritize among statutory provisions enacted at different times. That task, in turn, is simply a particular application of a maneuver required in just about all cases of statutory interpretation.

Put differently, a judge called upon to decide a severability question arising out of an amended statute is a bit like Molière's bourgeois gentilhomme, who is surprised (and delighted) to learn that he can speak prose. To decide a severability question arising out of an amended statute is to engage in a task in which judges have a great deal of experience. In a reasonably mature legal system, statutory interpretation frequently is inter-temporal synthesis or prioritization.