As readers are no doubt aware, last week the Supreme Court invalidated the Trump administration’s rescission of the DACA program. In this post, I want to examine an argument that seemed not to matter to the decision, though arguably the Chief Justice implicitly rejected it. The argument holds that an agency can rescind (or decline to enact) a discretionary program on the sole basis that the agency believes the program to be unconstitutional, even if the reviewing court would find that the agency’s belief is incorrect. (Sometimes it is said that the agency’s belief must at least be reasonable.) In the wake of the Supreme Court’s decision some have championed this argument in explaining why the case was wrongly decided. I think the argument is wrong.
Let’s assume the following: DACA is not required by any statute and could be rescinded provided adequate justification is given. The President and all relevant agencies have a good faith belief that DACA is unconstitutional. The only basis given by the executive for rescinding DACA is that unconstitutionality. The reviewing court would, however, conclude that the agency was wrong and that DACA is constitutional. I’m also assuming the rescission is reviewable; that, to me, is a separate question. What result?
The answer, I would think, is that the court should set aside the agency action, correct its misunderstanding, and send the issue back to the agency. That approach is the normal one when it comes to legal errors of the kind at issue—where the agency thinks it is commanded to do one thing or another, but the agency is wrong, according to the court. Take Chenery I. There, the agency thought it had to apply certain common-law principles to the case at hand and dutifully applied those principles. The Supreme Court held that the agency was wrong that those common-law principles applied and remanded the case. As the Supreme Court wrote, an agency’s “order may not stand if the agency has misconceived the law.” The agency was then free to exercise its independent discretion and, in round two, prevailed. (Hence Chenery II!)
Or take what Daniel Hemel and Aaron Nielson call “Chevron Step One-and-a-Half.” In cases applying that doctrine, the agency thinks Congress has unambiguously commanded it to do something. The court disagrees. In fact, the agency has discretion. As in Chenery I, the solution is to correct the agency’s “mistake” and allow the agency to re-decide the issue but now with knowledge that it is operating within a zone of ambiguity. The Supreme Court arguably applied Chevron Step-One-and-a-Half in Negusie v. Holder. It is regularly applied in the D.C. Circuit.
These doctrines make good sense. When Congress grants an agency discretion, that discretion is undermined to the extent the agency misperceives the limits of its own freedom. They would also point toward vacating and remanding in the hypothetical scenario described above regarding DACA.
What are some possible responses? One is that the Trump administration’s determination regarding DACA rests, at least partly, on the Constitution, unlike the cases cited above. As an initial matter, the alleged constitutional infirmities with DACA are seemingly all derivative of the statutory ones, a byproduct of the unfortunate recent tendency to turn every statutory case into one about “The Separation of Powers.” But even granting that DACA presents constitutional issues, I don’t get the distinction. Bracketing certain caveats regarding Chevron not relevant here, agencies are equally bound by statutes as by the Constitution. And the practical effect of a mistake is the same. In each case, the agency perceives some external force as cabining it in, and, according to a court, it is wrong.
To shore up this point, proponents of the contrary view may repair to the concept of departmentalism. Roughly, departmentalism holds that each branch has independent authority to interpret the Constitution in its sphere of authority. So, for example, President Jackson was free to veto a bill re-chartering the Second Bank of the United States because he thought the bank was unconstitutional, even though the Supreme Court had decided otherwise. But the DACA case isn’t like vetoing a bill. It’s a properly presented case-or-controversy, which falls within the courts’ sphere of authority. If the governing law requires the courts to set aside an agency action based on a perception of the Constitution with which a court disagrees—which is what we’re arguing about—the courts should follow that law in deciding cases before it.
To drive the point home, imagine that the APA were amended to be super clear on this point. It now says that an agency action shall be set aside “if the agency action is based on a misperception of any law, including the Constitution. In making this determination, the court shall give no deference to the agency’s views.” Is this statute unconstitutional? It’s hard for me to believe that it is. And if that’s right, we’re really just talking about what the content of administrative law is, and the cases above are relevant once more.
Another argument in favor of the wrong-but-reasonable position is based on the APA that we do have. The challenges to DACA have been largely litigated on “arbitrary and capricious” grounds. The courts often say that arbitrary and capricious review is pretty deferential. When it comes to the ultimate policy judgment being made, an agency can be a little bit wrong (or even more so) and still get away with it. And so too here, an agency should be allowed to be a little bit wrong about the Constitution.
The problem is that arbitrary-and-capricious review is chameleon-like, and how deferential courts are depends on what form it properly takes in the given setting. For example, one component of arbitrary and capricious review says that an agency must consider the factors made relevant by statute and may not consider factors Congress has barred the agency from considering. An agency that considers most of the relevant factors and only one prohibited factor will not win because it got things more right than wrong. That’s just not a situation in which courts apply the “little bit wrong” standard to cut agencies slack.
So then the question is whether the “little bit wrong” standard applies to agency decisions based on constitutional positions with which the court disagrees. I think the answer is no. First, the case law cited above does not support it, and there is no reason to think the APA speaks differently to constitutional mistakes than statutory ones. Second, the reasons for deference do not apply here. Courts do not flyspeck agency decisions based on the agency’s policymaking discretion because courts are not competent to do so. Courts defer in Chevron cases for similar functional reasons and because under the prevailing conception of Chevron that’s what the law requires when Congress has vested an agency with authority to administer the statute in question. But agencies do not administer the Constitution, and there is no reason to think their constitutional judgments are more likely to be correct than the courts’ own. (I’m open to the idea that in some constitutional cases, such as those involving procedural due process, some degree of deference to agency judgments might be appropriate, as Adrian Vermeule has argued. But the considerations supporting deference there do not translate to the DACA context.)
Finally, allowing agencies to be “a little bit wrong” about the Constitution would have strange consequences. Take an agency operating in good faith that really wants to enact a certain program, but the smartest minds in its General Counsel’s office conclude that the agency cannot do so consistent with the Constitution. In theory, the agency would be thrilled to “lose” the resulting litigation. It is now freed! But the “little bit wrong” position would saddle the agency with its incorrect view.
Now take an agency operating in bad faith. Say the EPA has discretion to apply a statutory program to certain power plants, but the statute doesn’t compel it to either. The policy basis for applying the program is very strong, but the EPA likes power plant owners and doesn’t want to apply it. Under the “little bit wrong” view, the agency can simply allege a constitutional problem with the program and hide its weak policy case. Now perhaps this type of behavior is what the “reasonability” caveat is supposed to address, or perhaps it could be addressed through engrafting a “good faith” requirement. But my head hurts just thinking about the litigation. Here’s a simpler idea: Let the courts decide what the Constitution allows, and let agencies exercise their delegated policymaking discretion within those limits.
Daniel T. Deacon is a Lecturer at Michigan Law School. Thank you to Dan Hemel for helpfully commenting on a draft of this post.