Tyler B. Lindley1Tyler B. Lindley is a J.D. Candidate at The University of Chicago Law School, Class of 2021, and graduated from Brigham Young University with a B.S. in 2018. For very helpful feedback and discussion on previous drafts, he thanks Jonathan Mitchell, William Baude, Aziz Huq, Tom Ginsburg, Thomas Miles, Ernest Young, Jared Mayer, Micah Quigley, Eric Wessan, and the participants of the Canonical Ideas in American Legal Thought Seminar. He also thanks the editors of The University of Chicago Law Review for thoughtful comments and edits. Lastly, he thanks his wife, Katrina Lindley, for her indispensable discussion and support.


The Supreme Court’s changing composition and, relatedly, its increasing skepticism for the current structure and pervasiveness of the administrative state have given rise to increased constitutional challenges to agency actions that seem increasingly likely to be successful. However, the Court’s jurisprudence (and consequently the lower courts’ application of it) on the proper remedy when a constitutional violation is found is fractured and underdeveloped. Particularly, two strands of jurisprudence—justiciability and remedies—are in tension with one another. This conflict also puts pressure on the Court’s commitment to the doctrine of constitutional avoidance.

The root of the tension is the “writ-of-erasure fallacy.” The judicial power granted to federal courts in Article III does not allow judges to repeal laws or remove statutes from the U.S. Code; such a system was considered and rejected in 1787. Instead, judicial review allows judges to refuse to enforce certain laws if they deem these laws “repugnant to the constitution” and to enjoin executive officials from particular methods of enforcing the law.

Judges commit the writ-of-erasure fallacy when they purport to “strike down” or “nullify” laws they have determined to be unconstitutional and treat them as though such laws are vetoed, repealed, or erased. Although the linguistic part of the fallacy is important, it is the effect on the law that is key. That is, while judges often err when they purport to strike down or nullify a law, it is a no-harm–no-foul situation until they—or others—subsequently treat the law as vetoed, repealed, or erased. (And judges who do not use faulty language can still commit the fallacy by treating laws as such.) Judges who reject the writ-of-erasure fallacy accept a limit on the remedies that judges can order. To avoid the awkward double negative in the phrase “rejecting the writ-of-erasure fallacy,” I use the term “remedial limit” or the phrase “accepting the remedial limit.”

This remedial limit was recognized at the inception of judicial review, even though the rhetoric surrounding judicial review did not always reflect it. Over time, courts increasingly used lackadaisical and careless language, eventually coming to believe that Supreme Court decisions could actually nullify a law, as if the legislature itself had repealed it or the executive vetoed it.

Since Jonathan Mitchell presented the framework for the remedial limit in The Writ-of-Erasure Fallacy in 2018, two Supreme Court justices,2See Seila Law LLC v. Consumer Financial Protection Bureau, 140 S. Ct. 2183, 2219–20 (2020) (Thomas, J., joined by Gorsuch, J., concurring); United States v. Sineneng-Smith, 140 S. Ct. 1575, 1585–86 (2020) (Thomas, J. concurring) (arguing that the overbreadth doctrine of the First Amendment extends beyond facial challenges and unjustifiably relaxes standing requirements); Murphy v. National Collegiate Athletic Ass’n, 138 S. Ct. 1461, 1485–87 (2018) (Thomas, J., concurring) (arguing that current severability doctrines violate standing requirements and traditional statutory interpretation principles). sixteen circuit judges,3See Pool v. City of Houston, 978 F.3d 307, 309 (5th Cir. 2020) (Costa, J., joined by Graves & Engelhardt, JJ.); Jacobson v. Florida Secretary of State, 974 F.3d 1236, 1255 (11th Cir. 2020) (Pryor, William, J., joined by Luck, J.), superseding 957 F.3d 1193; United States v. Swenson, 971 F.3d 977, 989 (9th Cir. 2020) (Smith, N.R., J., concurring in part and dissenting in part); Texas v. United States, 945 F.3d 355, 396 (5th Cir. 2019) (Elrod, J., joined by Engelhardt, J.); National Collegiate Athletic Ass’n v. Governor of New Jersey, 939 F.3d 597, 612 (3rd Cir. 2019) (Porter, J., dissenting); Island Creek Coal Co. v. Bryan, 937 F.3d 738 (6th Cir. 2019) (Murphy, J., joined by Thapar & McKeague, JJ.); Collins v. Mnuchin, 938 F.3d 553, 610–11 (5th Cir. 2019) (Oldham & Ho, JJ., concurring in part and dissenting in part); Lester v. United States, 921 F.3d 1306, 1317 (11th Cir. 2019) (Pryor, William, J., statement respecting the denial of rehearing en banc); Close v. Slotheby’s, Inc., 909 F.3d 1204, 1210, 1210 n.1 (9th Cir. 2018) (order) (Watford, Bybee, & Boggs, JJ.). and a magistrate judge have cited the remedial limit approvingly. However, none of these opinions has squarely focused on its implications for current justiciability doctrines and remedies, particularly in administrative law. Generally speaking, when a court purports to nullify a statute, and doing so does not redress the plaintiff’s alleged harm, the court is issuing a mere advisory opinion. Thus, in administrative law cases where courts purport to nullify a constitutionally impermissible provision, the ultimate judgment must address the plaintiff’s harm. Because courts cannot strike down laws, an opinion untied to the remedy is merely an advisory opinion on how it would rule in a future case and not connected with a live case or controversy.

This Essay seeks to bring the doctrines of remedies and justiciability in administrative law cases into cohesion. Part I identifies and explains the error courts are making—deciding constitutional questions without first deciding whether that decision can remedy the plaintiff’s alleged harm or whether that decision is necessary to resolve the dispute. Part II explores changes needed to make the doctrines consistent by examining how they should apply to a recent case and a pending case. In particular, courts should ensure that determining that a provision is unconstitutional will give relief to the plaintiff or be necessary to the resolution of the case. And Part III briefly analyzes the systemic and practical consequences of such changes.

I.  The Writ-of-Erasure Fallacy and Justiciability

A. Foundation and Error

The writ-of-erasure fallacy has been around as least as long as the federal judiciary itself. At the Constitutional Convention, the Virginia Plan included a “council of revision” that would include judges from the “National Judiciary.” The council of revision would review each federal and state law and could veto any law it deemed unconstitutional or “unwise or unjust.” This “dissent,” as it was called, was subject to override if the law were passed again, similar to the presidential veto. Influential founders such as James Wilson, George Mason, and James Madison supported the council, but they were repeatedly defeated. The Convention also rejected Madison’s separate proposal to “give the Supreme Court a . . . separate and independent” veto power and Edmund Randolph’s various proposals to give the judiciary additional veto-like power over state laws and federal laws that conflicted with state laws. However, the defeated founders continued to assert that the judiciary had the ability to render statutes void. One reading of the language in Marbury v. Madison further advances the idea that when a court declines to enforce “an act of the legislature” or executive that is “repugnant to the constitution,” that law “is void.” As shown here, that perception is at least as strong and believed today, and in any case is certainly more prevalent given the proliferation of federal regulation and litigation.

Mitchell asserts that the use of the word “void . . . suggests a permanent nullification.” This is certainly a reasonable (and perhaps the most reasonable) construction of the word and seems to be how Georgia state courts, whose constitution expressly allows its courts to declare acts void, use the term. But it is also possible that “void” merely means that the judiciary must “treat [the] acts as null and void” (emphasis added) in performing its judicial duties, as the Supreme Court has stated. Because it is possible for a court to declare a law void but not “fall victim” to the writ-of-erasure fallacy, the fallacy is the act of treating the statute as though it has been vetoed, repealed, or erased—not merely the word choice. Although, to be sure, the verbiage does contribute to the growth and acceptance of the writ-of-erasure fallacy.

The opposite of the writ-of-erasure fallacy is to accept a “remedial limit,” recognizing that the power of the federal courts to remedy harm does not extend to vetoing, repealing, or erasing the offending statute. That limit means that even after a court declares a law unconstitutional, the statute still exists and can be amended, repealed, or enforced (especially if the court overrules its previous decision).

Since Mitchell’s article was published three years ago, a cross-ideological, but heavily originalist, array of judges have positively cited it—including Justices Clarence Thomas and Neil Gorsuch, sixteen federal appellate court judges, and a magistrate judge. Despite the considerable amount of traction it has gained, its application to current doctrines has not been fully fleshed out. Particularly, relatively little attention has been devoted to its application to justiciability doctrines, and that relationship remains generally undertheorized.

Justice Thomas first cited Mitchell’s article in a concurring opinion in Murphy v. National Collegiate Athletic Association. Justice Thomas called for a reconsideration of the Court’s severability jurisprudence. In doing so, he discussed why refusing to enforce a part of a statute—while indiscriminately continuing to enforce others—is not “lawmaking,” but rather the proper result of judicial review. Justice Thomas also questioned whether plaintiffs would have standing to challenge other parts of the statute when the Court considers whether those other provisions should remain enforceable during its severability analysis. Because the remedial limit’s application to current severability doctrine is distinct and complex, I do not address that relationship here; however, some of the same analysis presented below could be applied to severability.

Several judges on the Fifth Circuit have also explicitly accepted the remedial limit in opinions. In one case, Collins v. Mnuchin, the en banc Fifth Circuit addressed the constitutionality of statutory provisions that restricted the circumstances under which the president could remove the director of the Federal Housing Finance Agency (FHFA). On the constitutional question, a majority of the court held that the removal restrictions unconstitutionally infringed on the executive power of the president. Then, a separate majority in a “separate opinion” held that the proper remedy for the constitutional violation was to excise the provision restricting removal but to allow the challenged action to remain in force. Concurring, Judges Andrew Oldham and James Ho, who were in the majority on the first question but dissented on the second, argued that the remedy adopted—or more precisely, the refusal to grant a remedy—by the separate opinion caused the court to extend beyond its Article III powers. I discuss Judges Oldham and Ho’s opinion more in Part I.B, and the case is discussed further in Part II.A.

Other judges have adopted the remedial limit in various contexts—including First Amendment overbreadth challenges, preemption, habeas corpus, wrongful injunctions, agency adjudication, pre-enforcement challenges, and the effect of injunctions and orders. While justiciability is only one area in which adopting the remedial limit could alter current judicial doctrine, judges who adopt the limit should ensure both that the doctrine itself is consistent and that they apply it consistently. This Essay seeks to further the former aim by presenting a coherent theory of justiciability consistent with accepting the remedial limit.

B. Justiciability After Accepting the Remedial Limit

Article III of the U.S. Constitution gives the federal judiciary “the judicial power,” but limits its jurisdiction to certain “‘Cases’ and ‘Controversies.’” The standing doctrine is one of multiple justiciability doctrines used to determine when the “case-or-controversy requirement” is satisfied by any particular dispute. Over time, courts have set out three constitutionally irreducible minima for standing: First, the plaintiff must have suffered a concrete, particularized injury that is actual or imminent; second, the injury must be fairly traceable to the alleged illegal conduct by the defendant; and third, the requested relief must be reasonably probable to redress the injury alleged. There is perhaps an independent fourth prong—a prohibition against “generalized grievances”—which is constitutionally required and not only a prudential requirement. The generalized grievance rule is best understood as constitutional, but merely as a corollary or special application of the particularized-injury prong rather than a conceptionally independent bar.

The remedial limit has little to say directly about the first two prongs, injury in fact and traceability, but accepting that limit requires judges and scholars to rethink the redressability prong of standing. First, courts are powerless to redress harm that stems from the mere presence of a statute. (I do not focus on this implication here but have done so elsewhere.)

Second, and central here, relief that does not remedy that plaintiff’s harm faces two related justiciability problems: (1) it requires the court to go beyond its Article III power to resolve cases and controversies, and (2) the remedy purports to go beyond refusing to enforce the law because courts are either not being asked to enforce the law in that case or have already rejected the claim that would entitle a party to relief.

The problems identified in this Essay arise when courts order non-party-centric relief—that is, when courts use their judicial authority to order something that may or may not be inherently outside of their judicial power in a typical case, but where the order itself is not necessary to resolve the case or controversy before them. Courts who order non-party-centric relief commit the writ-of-erasure fallacy when the relief ordered is described and regarded as “striking down” or “excising” a statute. I do not explore here when orders that are at best tenuously related to the parties are problematic. I merely attempt to show that such orders are particularly problematic for a different reason—they purport to rely on a judicial power that courts do not have: the power to strike down statutes.

When courts purport to strike down a statute, and that order would not relieve the parties’ harm in the particular action even if they could strike it down, the court’s action is functionally nothing more than an advisory opinion. If a court did have the power to alter statutes, then an order purporting to strike down a statute could in theory have binding authority on non-parties. This holds at least insofar as it satisfies (and insofar as one accepts) Judge Daniel Manion’s formulation in City of Chicago v. Sessions: “[B]road relief, even relief that benefits non-parties, is sometimes necessary to provide complete relief to the actual plaintiffs. . . . In those cases, the relief to non-parties could be called a side-effect of the relief given to the plaintiffs.” If the order is necessary to provide complete relief to the plaintiff, then an order purporting to strike down a statute suffers only from the defect that it falls victim to the writ-of-erasure fallacy (rather than a specific justiciability problem).

In Collins v. Mnuchin, Judges Andrew Oldham and James Ho concluded that there was no basis for which courts “could purport to delete a statutory provision when there is no active case or controversy within the meaning of Article III.” While related, my argument is slightly different: regardless of whether it could be generally acceptable, it is uniquely problematic precisely because courts are not actually erasing the statute.

Because courts do not have such a power, at least so far as one accepts Mitchell’s account, judges are, at bottom, offering nothing more than pontifications as to how they would rule on a future case properly presenting the issue. They are no longer justified in ruling on the constitutionality of the statute before them because doing so would not affect the legal “rights of litigants in the case before them” and does not actually erase the statute. As the Supreme Court stated in 1969 in Golden v. Zwickler, quoting an 1885 case: “No federal court, whether this Court or a district court, has ‘jurisdiction to pronounce any statute, either of a State or of the United States, void, because irreconcilable with the Constitution, except as it is called upon to adjudge the legal rights of litigants in actual controversies.’” Our system does not allow courts to issue such advisory opinions.

To be sure, in many administrative law cases, such an order might affect one party to the suit: the executive official. But the relief must affect the rights of both parties—as shown by the combination of the injury and traceability prongs in the standing analysis, a case must have the right plaintiff and the right defendant. There are forceful practical arguments for allowing private citizens to challenge government conduct so as to avoid intra-executive branch disputes, but even in those cases, the challenger must still meet the Article III requirements. As Judge Gregg Costa explained in Collins, “affected individuals can bring [separation-of-powers] claims. . . . But that does not mean they don’t have to be affected by the allegedly unconstitutional law.”

Thus, as demonstrated above, those who reject the writ-of-erasure fallacy must also refuse to opine on the constitutionality of a statute when so doing would not give relief to or affect the legal rights of the parties before the court.

II.  Case Studies

Administrative law cases can raise these issues in which judges, by committing the writ-of-erasure fallacy, might violate the mandate of Article III or might prematurely decide a constitutional question. This Part now examines two such contemporary cases. In the first, a case currently pending before the Supreme Court, Collins v. Mnuchin, a fractured Fifth Circuit court sitting en banc held unconstitutional and purported to “strike down” a removal restriction and then refused to grant any relief to the plaintiffs. In the second case, Seila Law, LLC v. CFPB, decided by the Supreme Court last term, the Court purported to excise a removal provision despite leaving unresolved whether it was necessary to resolve the dispute.

A. Collins v. Mnuchin: Redressability and Unrelated Relief

In Collins v. Mnuchin, the plaintiffs, shareholders of the Federal National Mortgage Association (“Fannie Mae”) and the Federal Home Loan Mortgage Corporation (“Freddie Mac”), are challenging an action by the Federal Housing Finance Agency (FHFA), an independent agency. The statute creating the FHFA protects the sole director of the agency—appointed by the president and confirmed by the Senate for a term of five years—from removal by the president, except for cause. After taking over Fannie Mae and Freddie Mac, the FHFA signed a financing deal with the Treasury in 2008. The deal gave Fannie Mae and Freddie Mac much-needed financing and a continuing line of credit, and gave the Treasury senior debt, quarterly dividends, stock warrants, and other fees. The FHFA and Treasury also signed various amendments to the original financing agreement, including the “Third Amendment” in 2012. The Third Amendment required Fannie Mae and Freddie Mac to “give [the] Treasury nearly all their net worth each quarter as a dividend” to “expedite the wind down” of Fannie Mae and Freddie Mac.

The case does not present a simple question of the constitutionality of the removal restriction, but instead involves several preliminary statutory questions necessary to reach the removal question. For example, the Fifth Circuit concluded that acting directors—including the Acting Director who signed the Third Amendment—are also protected by the removal restriction. But the claim most germane to this Essay, and one the Fifth Circuit reached below, is that the “FHFA lacked authority to adopt the Third Amendment because its Director was not removable by the President.” A majority of the Fifth Circuit sitting en banc held that the removal restrictions were unconstitutional; a separate majority concluded that the proper remedy for that constitutional violation was to strike—or more precisely, attempt to strike—the restriction in the statute but to leave the Third Amendment in place; and two judges believed that the plaintiffs lacked standing to challenge the removal restriction altogether. The following sections address each argument.

1. Unrelated Relief

The court ultimately concluded that the removal restriction was unconstitutional and then purported to strike out the restriction but keep the Third Amendment in effect. Of the three approaches put forward by various judges and examined here, this one seems the most conceptually puzzling. A complete overview of agency structure and what should happen to regulations promulgated while the agency is unconstitutionally structured is beyond the scope of this Essay. But simply put, the majority concluded that a regulation promulgated by an executive official should only be invalidated if (1) the actor was granted “power inconsistent with their role in the constitutional program;” (2) the actor was “not properly appointed under the Constitution;” or, “perhaps,” (3) if plaintiffs can show that the removal restriction altered the regulation’s content or approval.

Although one could forcefully argue that appointing someone to a position with an unconstitutional removal restriction is also an unconstitutional appointment (or, at least that both officers equally lack constitutional authority), for purposes of this Section, I accept that unconstitutional appointments or delegations should be treated differently. Here, only the third circumstance is potentially applicable. The separate opinion in Collins held that even if vacating an action were generally permissible in the third circumstance, the removal restriction did not change the outcome because the president had ultimate authority of the Secretary of the Treasury and the action also required the Secretary’s approval. Even this conclusion is problematic because the two sides were not working together, and contractual negotiations require substantial give-and-take.

But taking that conclusion at face value, this “remedy” without relief still raises three related concerns: First, the court, including members of the separate majority that decided the remedy (Judges Stuart Kyle Duncan and Priscilla Owen), just weighed in on a constitutional question that was not before the court as a case or controversy.

Second, this approach clashes with the generally accepted principle that, in separation of powers cases, plaintiffs have standing even if they cannot show that they were treated differently than they would have been if the Constitution had not been violated. Seila Law explicitly rejected that argument with respect to the injury prong of standing, holding that a challenger need not show that “the Government’s course of conduct would have been different in a ‘counterfactual world.’” Despite the separate opinion’s assertion to the contrary, Seila Law’s reasoning required the plaintiffs to show that the president’s lack of control over the FHFA’s action caused their harm. It makes little sense to require a challenger to show more to get relief for an injury than to show a constitutionally sufficient injury.

Third, a court does not have the power, even when presented with a case or controversy that would be resolved by such a determination, to “excise” the removal restriction. A court can only refuse to enforce a provision or issue an injunction against its enforcement. Perhaps perversely, it may have been the separate majority’s concern for their “co-equal branches of government” that caused it to go beyond its own constitutional bounds. But, ultimately, the court’s ruling on the constitutional issue did not affect the legal rights of the plaintiffs or remedy their alleged harm because in either case, the Third Amendment would have remained in effect. Nor could it have because the court held that no matter the constitutional violation, the shareholders were not entitled to relief. The constitutional conclusion was thus untied from the case or controversy before it—if there ever was one.

The separate opinion’s criticism that invalidating the Third Amendment while leaving other FHFA actions in place was picking and choosing also reveals some confusion. It is unclear how the remedy majority thinks the plaintiffs would have gained standing to challenge every act by the FHFA, especially considering two judges in the majority did not think the plaintiffs had standing to challenge even the Third Amendment. Of course, picking and choosing the provisions of the Third Amendment to invalidate, while leaving the Treasury obligated to perform on the remaining provisions would be problematic. But the remedy dissent did not propose that result. Rather, it recommended that the Third Amendment be invalidated in its entirety and that, on remand, the district judge should determine how much benefit the shareholder derived from the Third Amendment—and, consequently, how much the shareholders should owe in restitution.

2. Lack of Standing

Judges Gregg Costa and Stephen Higginson argued that the plaintiffs had no standing. They agreed with the separate opinion’s conclusion that the removal restriction had no effect on the actions by the director of the FHFA. Thus, they argued, the plaintiffs did not have standing. This is an internally coherent theory insofar as one accepts that the injury and requested remedy require allegations and proof of different treatment caused by the removal restriction. Plaintiffs must show that they have been injured by unlawful conduct. If the agency head’s conduct was not different than if she had been removable, then any harm is not traceable to that conduct.

However, this position—at least as it is presented by Judge Costa in Collins—has certain drawbacks as well. First, the two judges also joined the separate opinion setting forth the remedy. There can be no remedy if there is no standing. Of course, had they not joined the remedy majority, the court would have split 7-7-2. Perhaps Judges Costa and Higginson wanted to avoid such a split, or perhaps their votes are best explained by majority-building rather than an effort to present a coherent legal principle.

Second, it presents the same problem as the unrelated relief ordered by the separate opinion: Plaintiffs generally do not need to show that they received worse treatment than they would have had if the Constitution had not been violated. There is independent harm in being subjected to action by a government actor that is not authorized by the Constitution to act. And it makes little sense to require plaintiffs to satisfy a higher burden to have their alleged harm remedied than to receive the judiciary’s opinion on the constitutionality of a statute. In fact, in Raines v. Byrd, the Supreme Court said that “our standing inquiry has been especially rigorous when reaching the merits of the dispute would force us to decide whether an action taken by one of the other two branches of the Federal Government was unconstitutional.” Requiring proof of different action would turn this principle on its head.

To be sure, the separate opinion claimed to have remedied the “ongoing injury” (emphasis added), but the plaintiffs never alleged any ongoing injury or requested ongoing relief with respect to the removal restriction. Nor is it clear that the plaintiffs could have plausibly alleged that there would be an ongoing injury even if they had alleged such harm. Thus, the court either refused to remedy the past injury or implicitly adopted the view of some of the judges that there was no past injury.

3. Gradual Excision

The final argument was that because the removal restrictions were unconstitutional, the Third Amendment should be subject to rescission. This is also consistent with the remedial limit. If the restriction is unconstitutional, then courts are completely within their power to refuse to enforce the regulation (or here, the contract) and to order action consistent with the absence of the regulation. In this case, the plaintiffs sought an order that the Net Worth Sweep was without force and that Fannie Mae and Freddie Mac therefore had different legal obligations with respect to the shareholders. The separate opinion aptly explains why principles of contract law would require the court to subject the entire Third Amendment to rescission—and not just clauses that are adverse to the plaintiffs—but that is also within the judicial power.

Of course, as soon as a court holds that a removal restriction is unconstitutional, many others negatively affected by that agency’s actions will soon seek redress for their particular injuries. At the same time, some actions may never be invalidated, perhaps because no one is harmed, no one has the proper incentives to sue, or the actions are repealed or the equivalent.

To the extent that the consequences of my argument for the remedial limit inform what burden of proof courts should use in deciding constitutional questions, there are at least two competing considerations. First, although courts will be constrained with respect to constitutional questions upon which they may pass, they may be more willing to lower the burden of proof, reasoning that the effect of that particular decision is smaller because only one government act is being invalidated as a result. Whereas when a court purports to invalidate a statutory provision, it believes to be altering a congressional enactment, as opposed to refusing to enforce a discrete agency action. Second, and conversely, in many instances the remedial limit would prohibit courts choosing exactly whether agency actions are to be invalidated because courts cannot chose to either erase the whole statute (if it is not severable) and invalidate all actions at once or alter the statute (if it is severable) and keep all actions in place. This may make wary courts more likely to defer to the legislature out of concern for perverse consequences that none of the executive, legislature, or judiciary intended or even foresaw.

B. Seila Law: Constitutional Avoidance

In Seila Law LLC v. Consumer Financial Protection Bureau, a private law firm sought judicial review when the CFPB issued and sought to enforce a civil investigative demand (“essentially a subpoena”) for certain documents. Seila Law defended itself on the grounds that the demand “was invalid and must be set aside.” Because, the argument went, the structure of the CFPB—its head being insulated from presidential removal by way of a statutory “for cause” termination provision—violated the separation of powers, the demand was fundamentally defective and could not be enforced.

The district court and Ninth Circuit Court of Appeals rejected Seila Law’s constitutional argument and concluded that the subpoena was enforceable. The Supreme Court disagreed, concluding that the CFPB unconstitutionally granted significant executive power to an unaccountable single agency head. The Court also concluded that the removal restriction was severable from the Dodd-Frank Act, which created the CFPB, effectively keeping the CFPB in operation.

But the Court never conclusively addressed whether the demand was actually enforceable—whether Seila Law was entitled to relief. The Court noted that the parties disputed whether the demand had been ratified during a time in which an Acting Director (terminable by the president at will) led the CFPB, purported to ratify the initial demand, and continued the enforcement proceedings. Thus, the Court engaged in the severability analysis to determine whether to reverse the Court of Appeals and reject the demand or to remand to sort out the ratification arguments. Ultimately, the Court concluded that the removal restriction was severable and remanded the case to the Ninth Circuit to resolve the ratification dispute.

Justices Thomas and Gorsuch—concurring in the constitutional conclusion, but dissenting as to the severability and remedial analysis—vigorously disagreed that the Acting Director’s limited role had any relevance to the enforceability of the subpoena. Instead, they argued, even if the Acting Director “properly ratified the issuance of the investigative demand and the initiation of the enforcement proceedings,” the then-director of the CFPB, Kathy Kraninger, was continuing the enforcement action, giving rise to an independent injury that could only be remedied by refusing to enforce the demand.

One key aspect of the Court’s decision identified here is that the Court seems to have purported to strike down a statutory provision without first discerning whether it was necessary to resolve the dispute. Unlike in Collins, Seila Law had standing in this particular case; it was the defendant in the enforcement action and consequently had standing to appeal the adverse decision from the lower court. Nevertheless, the Court seems to have embraced the constitutional question rather than avoiding it.

The Court sought to determine whether the plaintiff was in fact subjected to an unlawful action by first addressing whether the administrator was unconstitutionally insulated from presidential removal. But the Court never actually decided whether Seila Law itself had its legal rights altered by such an administrator. To be sure, the Court’s conclusion about the constitutionality of the CFPB’s structure made it possible that, in this particular case, Seila Law was subject to a constitutionally deficient demand. But even if the CFPB’s structure was unconstitutional, that conclusion was necessary only if the CFPB’s ratification argument failed. If its ratification argument prevailed, then Seila Law had no valid defense to the enforcement action. Even though the Court’s order of operations in this case did not offend Article III’s standing requirement, it seems to have violated the doctrine of constitutional avoidance.

Constitutional avoidance dictates in part that the Court not address a constitutional question unless its resolution is necessary to the proper disposition of the case. Just eleven years ago in Pearson v. Callahan, the Court unanimously reasserted the correctness of the “older, wiser judicial counsel ‘not to pass on questions of constitutionality . . . unless such adjudication is unavoidable’” (emphasis added). The Court’s decision in Seila Law seems to have run counter to that wise counsel.

But where does the writ-of-erasure fallacy come into play? The Court never actually did anything to the statute itself, and because it never did, it opined on a constitutional issue that (1) at least potentially did not change the parties legal standing, depending on the resolution of the ratification dispute, and (2) did not actually cure the constitutional violation because the Court has no power to do anything to the statute. Essentially, the Court merely signaled to future parties that it would conclude that the CFPB’s structure was unconstitutional in future cases. It did not decline to enforce or choose to enforce a statue—as the judiciary is empowered to do—and it did not alter or erase a statutory provision because it had no power to do so.

This confusion is evident in the aftermath of the decision. Director Kraninger purported to ratify the agency’s previous decisions to issue the demand and then initiate and continue enforcement proceedings. On remand, the Ninth Circuit held that analysis of the Acting Director’s potential ratification (the purpose of the remand) was ultimately unnecessary. After the Supreme Court’s decision, Director Kraninger “knew” she could be removed by the president at will and “nevertheless ratified” the present actions. Thus, Seila Law’s injury and concerns “ha[d] now been resolved.”

That conclusion, though, raises the question of how Director Kraninger “knows” that she is removable at will by the president. It cannot be that the statute no longer protects her, because it does—the director is still statutorily insulated from presidential removal. If she knows it because five members of the Court have expressed that opinion, it is unclear why that is legally sufficient. The composition of the Court can change, and the new Court could be more receptive to the CFPB’s argument. To be sure, with the addition of Justice Amy Coney Barrett, the director may well be safe in assuming as a factual matter that the Court would uphold Seila Law’s fundamental conclusion, but it is unclear why that has the legal effect sufficient to allow for ratification. Perhaps the distinction is that, as a party to the suit, the CFPB is bound by the opinion. However, even that argument needs more; Director Kraninger herself was not a party to the suit, and the extent to which an opinion binds the parties as opposed to a judgment is unclear.

In sum, the Court seems to have embraced, rather than avoided, the constitutional question. As a result, there remains the possibility that Seila Law was never subject to an unconstitutional demand and thus that the Court’s conclusion that the CFPB Director was unconstitutionally insulated was wholly unnecessary. Further, the Ninth Circuit’s reasoning ensures that as long as the constitutional question is decided before the proper remedy is determined, no relief will ever be granted. If the Court had resolved the ratification dispute before answering the constitutional question, then Seila Law might have been entitled to relief. And because it did not, Seila Law’s fate was effectively sealed.

III. Systemic Effects

Reembracing the judiciary’s remedial limit in administrative law challenges will both (1) make such challenges less likely to be properly presented before a court, and (2) ensure that when such challenges do arise, the challenger will obtain relief upon a favorable resolution of the constitutional question. Further, this combination might also incentivize judges to hesitate to enforce constitutional limitations on agency actions; to relax prudential, non-constitutional constraints (such as constitutional avoidance) to avoid upsetting the current state of government; or both.

One of the ways in which enforcing the judiciary’s limited remedial powers reduces the number of properly presented challenges is through redressability. For example, if the Collins majority is correct that the removal restriction was unconstitutional, but Judge Costa is also correct that the shareholders did not have standing—or, at least, that the shareholders were not ultimately entitled to their requested relief—there are very few parties that could successfully challenge the structure of the FHFA. And if the proper remedy is never vacatur or rescission of the agency action, as the separate opinion suggested, there are even fewer potential litigants. In that case, only the president or the FHFA Director would seemingly be able to successfully challenge or enforce the removal restriction, respectively. Given the exceptionally high costs of intrabranch litigation, the challenge may never arise because it is probable that, even if the president disagreed with the executive official, she would not attempt to remove the relevant official. And, conversely, if an official were wrongly removed, she might well decide that the costs of litigating against the president outweigh the benefits from a successful lawsuit.

Fortunately, Supreme Court precedent almost entirely forecloses that line of analysis. Seila Law explicitly rejected it with respect to standing, citing the reasoning of Bowsher v. Synar to support its conclusion. In Bowsher v. Synar, an executive official was subject to removal by the legislature. The Court concluded that ripeness does not require waiting until “that provision is actually used” because the threat of removal changes behavior. The Court’s conclusion that the threat of removal changes behavior is almost tautological, but that does not justify expanding the Court’s remedial power beyond that allowed by Article III. If anything, it shows that acts by unconstitutionally insolated executive officials are always different than they would be if the officials were subject to complete control of the president. And thus, contrary to the Collins separate opinion, challengers to have been subjected to such unconstitutional acts are always entitled to a remedy.

Even assuming challengers have standing and would otherwise be entitled to a remedy, adherence to the doctrine of constitutional avoidance will further narrow the cases in which the constitutional questions are decided. As described above, constitutional avoidance—combined with the recognition that holding a statutory provision unconstitutional does not change the statute going forward—requires that before answering the constitutional question courts determine both whether plaintiffs would be entitled to relief and whether there are other reasons to grant relief. Of course, in many cases, there will be a reason, such as ratification, that the challenger will not be entitled to relief even if correct on the constitutional question; or there may be another reason to grant relief, such as the agency exceeding its statutory authority.

To be sure, constitutional issues will be slower to reach courts, particularly the Supreme Court. On the other hand, once the issue is squarely presented, a court’s decision that the agency is unconstitutionally structured might effectively undo many past actions the agency has taken. Perhaps this weighty result is why the Seila Law plurality chose to decide the constitutional issue and leave the ratification issue to be decided on remand. In that way, the Court could vindicate constitutional separation of powers principles while avoiding the risk of collapsing the entire agency and crippling the agency’s ability to act.

However prudential or constitutionally permissible it may have been to reach the constitutional question in Seila Law, the Court’s decision still did not and cannot change the language of the Dodd-Frank Act. It still purports to protect Director Kraninger from removal. Left open is whether congressional intervention (in the form of an amendment removing the statutory protection or removing some of the director’s power) is required to enable the CFPB to constitutionally perform any act. There is nothing that independently requires such intervention, but, at minimum, more justification is needed to defend why Seila Law renders the CFPB’s past and future actions constitutional.


Accepting the remedial limit also requires accepting that courts have no power to veto, repeal, or erase an unconstitutional law that has otherwise been duly enacted. If this is correct, then federal courts are hearing and deciding constitutional issues that they should not be hearing—or at least that are not necessary to resolve the present dispute. An opinion holding that an agency is unconstitutionally structured does not change the statute. Thus, to issue such an opinion, courts must ensure that it resolves a case or controversy, and they should ensure that reaching the constitutional question is necessary to resolve it.

 * * *

Tyler B. Lindley is a J.D. Candidate at The University of Chicago Law School, Class of 2021, and graduated from Brigham Young University with a B.S. in 2018. For very helpful feedback and discussion on previous drafts, he thanks Jonathan Mitchell, William Baude, Aziz Huq, Tom Ginsburg, Thomas Miles, Ernest Young, Jared Mayer, Micah Quigley, Eric Wessan, and the participants of the Canonical Ideas in American Legal Thought Seminar. He also thanks the editors of The University of Chicago Law Review for thoughtful comments and edits. Lastly, he thanks his wife, Katrina Lindley, for her indispensable discussion and support.

* * *

Photo credit to Joe Gratz.

Leave a Reply