Arizona v. Biden

RECOMMENDED FOR PUBLICATION
Pursuant to Sixth Circuit I.O.P. 32.1(b)
File Name: 22a0144p.06

UNITED STATES COURT OF APPEALS
FOR THE SIXTH CIRCUIT

┐
STATE OF ARIZONA; STATE OF MONTANA; STATE OF
│
OHIO,
│
Plaintiffs-Appellees, │
│
> No. 22-3272
v. │
│
JOSEPH R. BIDEN, in his official capacity as President │
of the United States; UNITED STATES DEPARTMENT OF │
HOMELAND SECURITY; UNITED STATES OF AMERICA; │
ALEJANDRO MAYORKAS, in his official capacity as │
Secretary of Department of Homeland Security; CHRIS │
MAGNUS, in his official capacity as Commissioner of │
United States Customs and Border Protection; TAE D. │
JOHNSON, in his official capacity as Acting Director of │
United States Immigration and Customs Enforcement; │
UR JADDOU, in her official capacity as Director of U.S. │
Citizenship and Immigration Services, │
Defendants-Appellants. │
│
┘

Appeal from the United States District Court for the Southern District of Ohio at Dayton.
No. 3:21-cv-00314—Michael J. Newman, District Judge.

Argued: June 10, 2022

Decided and Filed: July 5, 2022

Before: SUTTON, Chief Judge; MOORE and COLE, Circuit Judges.
_________________

COUNSEL

ARGUED: Daniel Tenny, UNITED STATES DEPARTMENT OF JUSTICE, Washington,
D.C., for Appellants. Benjamin M. Flowers, OFFICE OF THE OHIO ATTORNEY GENERAL,
Columbus, Ohio, for Appellees. ON BRIEF: Daniel Tenny, Michael Shih, Sean R. Janda,
UNITED STATES DEPARTMENT OF JUSTICE, Washington, D.C., for Appellants. Benjamin
M. Flowers, Sylvia May Mailman, OFFICE OF THE OHIO ATTORNEY GENERAL,
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 2

Columbus, Ohio, Drew C. Ensign, OFFICE OF THE ARIZONA ATTORNEY GENERAL,
Phoenix, Arizona, Christian B. Corrigan, OFFICE OF THE MONTANA ATTORNEY
GENERAL, Helena, Montana, for Appellees. Daniel R. Suvor, O’MELVENY & MYERS LLP,
Los Angeles, California, Gina M. D’Andrea, IMMIGRATION REFORM LAW INSTITUTE,
Washington, D.C., for Amici Curiae.

SUTTON, C.J., delivered the opinion of the court in which MOORE and COLE, JJ.,
joined. SUTTON, C.J. (pp. 23–28), also delivered a separate concurring opinion.
_________________

OPINION
_________________

SUTTON, Chief Judge. In September 2021, the Secretary of Homeland Security issued a
memorandum to his deputies outlining the Department’s immigration enforcement priorities and
policies. Arizona, Montana, and Ohio filed this lawsuit in the Southern District of Ohio to enjoin
its implementation. The district court issued a “nationwide preliminary injunction,” applicable to
all 50 States, blocking the Department from relying on the priorities and policies in the
memorandum in making certain arrest, detention, and removal decisions. Our court granted the
National Government’s request for a stay pending appeal and ordered expedited briefing and
argument. We now reverse the district court’s grant of preliminary injunctive relief.

I.

Federal law gives the National Government considerable authority over immigration
policy. Consistent with its powers under the U.S. Constitution, U.S. Const. art. I, § 8, cl. 4,
Congress has enacted several statutes with respect to detention and removal.

As to detention, the Department of Homeland Security “shall take into custody” those
“criminal aliens” who are inadmissible or deportable by reason of their having committed certain
crimes—including aggravated felonies, firearm offenses, drug crimes, and crimes of moral
turpitude—or their having been involved in terrorist activities. 8 U.S.C. § 1226(c)(1).
Removable individuals often are in state custody after a state-law conviction. In such cases, the
Department issues a “detainer,” a notice to the State that it intends to take custody of the
noncitizens upon their release from state custody. 8 C.F.R. § 287.7(a), (d); Immigration and
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 3

Customs Enforcement Policy No. 10074.2 ¶¶ 2.4–2.6. The State then informs the Department of
the noncitizens’ release dates and holds them for up to 48 hours to allow the Department to take
custody. 8 C.F.R. § 287.7(a), (d). In other cases, the Department has discretion to have “an alien
. . . arrested and detained pending a decision on whether the alien is to be removed from the
United States,” and even when it decides to do so, it retains discretion to release the individual
with certain conditions. 8 U.S.C. § 1226(a).

As to removal, Congress has provided that, “when an alien is ordered removed,” the
Department “shall remove the alien from the United States within a period of 90 days,” except in
specified circumstances. Id. § 1231(a)(1)(A). During that time, the Department “shall detain the
alien.” Id. § 1231(a)(2). If, however, removal cannot be accomplished within the removal
period, continued detention is not required, and the Department has discretion to release
noncitizens under supervision. Id. § 1231(a)(3).

Congress has tasked the Secretary of Homeland Security, currently Alejandro Mayorkas,
with establishing “national immigration enforcement policies and priorities.” 6 U.S.C. § 202(5).
On September 30, 2021, the Secretary exercised this power by issuing “Guidelines for the
Enforcement of Civil Immigration Law.” R.4-1 at 1. This Guidance prioritizes enforcement
with respect to noncitizens who pose a threat to national security, public safety, and border
security. On November 18, 2021, 11 days before the Guidance took effect, two States from the
Ninth Circuit (Arizona and Montana) and one State from the Sixth Circuit (Ohio) filed this
action against the United States, the President, the Secretary, the Department, and other
Homeland Security officials (collectively, the Department or the National Government). They
filed the complaint in the Southern District of Ohio. Soon after filing the complaint, they
requested a preliminary injunction to prevent the Department from implementing the Guidance.

From where the claimants stand, the Guidance violates the Administrative Procedure Act
on the grounds that it is contrary to law, is arbitrary or capricious, and should have been
subjected to notice and comment. The legal centerpiece of their claim is that the Guidance fails
to honor 8 U.S.C. § 1226(c)(1), which requires the Department to take custody of certain
criminal noncitizens—those convicted of terrorist activities, aggravated felonies, firearm
offenses, drug crimes, and crimes of moral turpitude—when they are released from state or
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 4

federal prison, and fails to honor 8 U.S.C. § 1231(a)(1)(A), which requires the Department to
remove noncitizens within 90 days of receiving final orders of removal. Failure to respect the
requirements of the two statutes, the three States claim, has led to fewer detainers and removals,
prompting the release of more individuals from state custody into their communities and
imposing more costs and burdens on them: additional costs to pay for medical and educational
services and additional law-enforcement burdens given the risks of recidivism.

After rejecting a host of justiciability challenges to the lawsuit and after concluding the
Guidance likely violated the Administrative Procedure Act, the district court issued a
“nationwide preliminary injunction.” R.44 at 75. The National Government sought emergency
relief in this court. We granted a stay pending appeal and ordered expedited briefing and
argument. Arizona v. Biden, 31 F.4th 469, 482–83 (6th Cir. 2022).

II.

In deciding whether to grant a preliminary injunction, the federal courts ask several
questions. Has the plaintiff established “that he is likely to succeed on the merits”? Would the
plaintiff likely suffer “irreparable harm in the absence of preliminary relief”? Does the “balance
of equities” tip in the plaintiff’s favor? And does “the public interest” favor an injunction?
Winter v. Nat. Res. Def. Council, Inc., 555 U.S. 7, 20 (2008). We review the district court’s
ultimate decision whether to grant a preliminary injunction for abuse of discretion, and we
evaluate its legal determinations, “including the likelihood of success on the merits,” with fresh
eyes. Union Home Mortg. Corp. v. Cromer, 31 F.4th 356, 366 (6th Cir. 2022). We start, and
largely stop, with the likelihood-of-success inquiry.

In answering the likelihood-of-success inquiry, we have a few preliminary questions of
our own: What exactly does the Guidance do? To what extent does it limit officers’ discretion
to enforce our Nation’s immigration laws? How in this last respect does the Guidance differ
from guidelines issued by past administrations?

The memorandum offers “guidance for the apprehension and removal of noncitizens.”
R.4-1 at 1. It emphasizes that “the majority of undocumented noncitizens who could be subject
to removal have been contributing members of our communities for years” and instructs that
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 5

“[t]he fact [that] an individual is a removable noncitizen . . . should not alone be the basis of an
enforcement action against them.” Id. at 2. Noting that the Department lacks the resources to
apprehend and remove every one of the more than 11 million removable noncitizens in the
country, the Guidance explains that the agency will “prioritize for apprehension and removal
noncitizens” who fit within three categories: threats to “national security, public safety, and
border security.” Id. at 2–3. “Whether a noncitizen poses a current threat to public safety,” the
Guidance adds, “is not to be determined according to bright lines or categories,” but “requires an
assessment of the individual and the totality of the facts and circumstances.” Id. at 3.
Department personnel “should not rely on the fact of conviction or the result of a database search
alone.” Id. at 4. The memo calls for a “training program . . . to ensure the successful application
of this guidance” and a “review process” to “achieve quality and consistency in decision-
making.” Id. at 6.

By its terms, the Guidance, though aiming to focus resources in certain directions, does
not tie the hands of immigration officers. The Guidance cautions that the memo does not
“compel an action to be taken or not taken,” “leaves the exercise of prosecutorial discretion to
the judgment of” Department personnel, and “is not intended to, does not, and may not be relied
upon to create any right or benefit.” Id. at 5, 7. Through it all, the Guidance assures that the
Department “do[es] not lessen [its] commitment to enforce immigration laws to the best of [its]
ability.” Id. at 2. Consistent with this language, the Department of Justice acknowledges that the
Guidance “preserves officers’ discretion to enforce the immigration laws on a case-by-case
basis.” Appellants’ Br. 1. It “does not forbid officials from taking any enforcement action that
the States claim the [relevant federal statute] requires.” Id. at 14. It “permits officers to pursue
enforcement action against any removable noncitizen in the exercise of officials’ individualized
discretion.” Id. at 28. It “does not preclude enforcement in any category of case.” Reply Br. 13.
And “[i]n any given case,” it permits an officer to “determine that a noncitizen warrants
enforcement action—regardless of whether all, some, or none of the aggravating or mitigating
factors are applicable.” Id. at 19.

The national immigration authorities have issued similar memoranda under many
administrations. Because prioritization has long been necessary, prioritization is not new.
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 6

A 2000 guidance recognized the reality of “finite resources” and explained that immigration
authorities have traditionally “responded to this limitation by setting priorities” to achieve goals
such as “protecting public safety, promoting the integrity of the legal immigration system, and
deterring violations” of immigration laws. R.27-4 at 4. A 2005 guidance to Department lawyers
explained that “the universe of opportunities to exercise prosecutorial discretion is large,”
emphasized that “we must prioritize our cases to allow us to place greatest emphasis on our
national security and criminal alien dockets,” and directed lawyers to consider that “[s]ome cases
involve sympathetic humanitarian circumstances that rise to such a level as to cry for an exercise
of prosecutorial discretion.” Memorandum from William J. Howard, Principal Legal Advisor for
U.S. ICE 1, 2, 6 (Oct. 24, 2005), https://www.aila.org/infonet/ice-prosecutorial-discretion-memo.
The 2011 guidance prioritized “national security, public safety, and border security” while not
prohibiting “the apprehension, detention, or removal of other aliens unlawfully in the United
States.” R.27-5 at 1, 3. The 2014 guidance prioritized “threats to national security, public
safety, and border security” but allowed officers to pursue removal of nonpriority individuals if a
field office director determined that doing so “would serve an important federal interest.” R.27-7
at 1, 5. And the 2017 guidance directed officers to “take enforcement actions in accordance with
applicable law” but allowed agency heads discretion to “issue further guidance to allocate
appropriate resources to prioritize enforcement activities within these categories” such as by
focusing resources on “convicted felons” or those “involved in gang activity or drug trafficking.”
R.27-8 at 2.

With these preliminaries accounted for, Arizona, Montana, and Ohio face two
justiciability hurdles, both relevant to likelihood of success, in bringing these claims. Do they
have standing under Article III to bring the claims? And does the Guidance amount to
reviewable agency action? Each threshold question deserves a turn.

Constitutional standing. Article III of the U.S. Constitution permits federal courts to
adjudicate “cases or controversies,” not any political dispute that happens to arise between the
state and federal executive branches. To have standing to bring this lawsuit, the States must
show that they have suffered an “injury in fact” “caused” by the Guidance that a favorable
decision would “redress.” Lujan v. Defs. of Wildlife, 504 U.S. 555, 560–62 (1992). When a
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 7

claimant challenges the defendant’s actions with respect to third parties (here, the regulation or
not of noncitizens), it is “substantially more difficult” to establish standing given the causation
and redressability problems that invariably arise. Id. at 562 (quotation omitted).

In trying to meet these requirements, the trio of States points to monetary harms allegedly
caused by the Department’s failure to enforce the immigration laws more vigorously. Their key
concern is that the Department’s prioritization of some risks—public safety, terrorism, and
border security—will come at the expense of other statutory priorities. They worry in particular
that the Guidance will decrease the number of noncitizens detained and removed and will
shortchange efforts to detain and remove those convicted of drug crimes and crimes of moral
turpitude, all with downstream costs to the States in the form of additional crime and public-
welfare costs. But considerable speculation undergirds the claim.

As for injury, start with the reality that the Guidance does not directly injure the States. It
does not regulate the States by telling them what they can or cannot do in their jurisdictions.
And it does not purport to preempt any state or local law, whether criminal or otherwise. State
criminal sentences, for example, may be as long as each State wishes. The Guidance merely tells
federal employees what to prioritize in enforcing a federal law over which the U.S. Supreme
Court has said that the National Government has considerable, indeed often exclusive, authority.
See Arizona v. United States, 567 U.S. 387, 394–97 (2012).

Speculation abounds over whether and how the Guidance’s prioritization of the
apprehension and removal of noncitizens in the three States will injure them. That the National
Government decides to remove or detain person A over person B does not establish that it will
pursue fewer people, particularly with respect to a Guidance that never requires agents to detain
some noncitizens over others. Because the Guidance prioritizes the noncitizens who pose the
greatest risks to public safety, it also is hard to know whether fewer detentions and removals
means more injuries to States even on their own terms. Once one accounts for the twin realities
that there are many noncitizens who are illegally in the three States and that the Guidance
permissibly does not regulate what to do with those individuals, is it not possible that the
Guidance’s focus on public safety, border security, and terrorism will decrease burdens on the
States? Contingent injuries, especially those arising from the impact of regulations on third
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 8

parties not before the Court, rarely create cognizable cases or controversies. Lujan, 504 U.S. at
562–63.

The States’ asserted injuries also “hinge on the response” of individual immigration
officers to the Guidance. Id. at 562. Even if the Guidance places some process limits on how the
officers exercise discretion to pursue action against certain people, the States do not dispute that
the officers retain control over the volume of removals and detentions they effect. The States
acknowledge that the Department “has limited resources” and cannot fulfill its duties as to
“every covered alien.” States’ Br. 43. They concede that “DHS officers must exercise discretion
in deciding whom to arrest.” Id. And the States “take no issue with DHS’s guiding that
discretion with prioritization schemes.” Id. If an injury turns on choices made by others and if
those choices permit considerable “discretion,” the States have a burden to show “those choices
have been or will be made.” Lujan, 504 U.S. at 562 (quotation omitted); see Trump v. New York,
141 S. Ct. 530, 535 (2020) (per curiam) (“Any prediction how the Executive Branch might
eventually implement this general statement of policy is no more than conjecture at this time.”
(quotation omitted)).

Even the premise that the Guidance has coincided with a fall in immigration enforcement
overall does not lead to the conclusion that the Guidance is the culprit, let alone the challenged
portion of the Guidance. Other explanations exist. The National Government’s main
enforcement authority affecting noncitizens within a State after all has little to do with detention
and removal decisions at the back end. It has to do with prosecutorial discretion at the front end
when immigration agents and law enforcement decide whom to arrest and whom not to. The
States do not challenge this classic form of prosecutorial discretion, and the consequential
exercise of discretion when it comes to noncitizen populations in Arizona, Montana, and Ohio.
See Texas v. United States, 14 F.4th 332, 337–38 (5th Cir.), vacated, 24 F.4th 407 (5th Cir. 2021)
(en banc). Even if the injunction remained in place—even if in other words the Guidance were
removed—that would not necessarily result in the Department arresting more people, detaining
more people, or removing more people.

The States attempt to prove a causal connection by pointing to data indicating that
interior arrests, interior removals, and removals of serious criminal noncitizens have fallen from
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 9

2019 to 2022. States’ Br. 15–17. But this theory starts with conjecture and returns to it over and
over. The data show a drop in these enforcement categories beginning long before the
challenged Guidance went into effect on November 29, 2021, and in some instances even before
2019. See FY 2016 ICE Immigration Removals, U.S. Immigr. & Customs Enf’t,
https://perma.cc/5EY7-BFCA; Bipartisan Pol’y Ctr., Interior Immigration Enforcement by the
Numbers 2 (2014), https://perma.cc/7P9Q-GV56 (showing a downward trend in interior
removals beginning in 2010). The States also acknowledge that the data, quite understandably,
are incomplete as to more recent months. The National Government, for its part, points to
competing data that the Department arrested more individuals per month this year compared to
the same period in 2020, particularly at the border. Much has changed since 2019 anyway. The
Covid-19 pandemic surely influenced, and continues to influence, migration patterns and
immigration enforcement. As one prominent example, a national public health order went into
effect on March 20, 2020, that led to hundreds of thousands of expulsions at the border without
arrests, scrambling the statistics in all directions. See 42 U.S.C. § 265; 85 Fed. Reg. 17,060
(Mar. 26, 2020); Nationwide Enforcement Encounters: Title 8 Enforcement Actions and Title 42
Expulsions FY2021, U.S. Customs & Border Patrol, https://perma.cc/N7Z5-8D4X. Other
policies of a new presidential administration and other expectations created by any new
administration, all disconnected from the Guidance, also would affect enforcement and
immigration statistics. Even if we accept the States’ position that certain enforcement numbers
are lower today than they were in 2019, that does not show that the Guidance caused that decline
or, more to the point, that the drop will cause the claimed injuries to the States.

The States protest that the district court found that they would sustain these costs,
emphasizing that we review factual findings for clear error. True, the district court found that a
downward trend in removals under the Guidance could increase expenditures. But true or not,
this conclusion still runs into these same causation and redressability problems. The district
court did not connect the dots between the decrease in removals and the Guidance’s challenged
prioritization. Given the Department’s unrebutted statements that the Guidance reflects
longstanding application of this 1996 statute, it is difficult to show that the Guidance has made
the difference. The court’s fact finding might survive review, but its materiality likely will not.
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 10

We are not persuaded by the decision of another district court that recently vacated the
Guidance. Texas v. United States, No. 6:21-CV-00016, 2022 WL 2109204 (S.D. Tex. June 10,
2022). No less true there than here, the causation and redressability problems identified above
deflate its findings that this Guidance has resulted in a drop in arrests of criminal aliens and in
interior removals. See id. at *10–13, 17–18. And because the Guidance prioritizes noncitizens
presenting the greatest risks to public safety and border security, it remains speculative whether
any drop in detentions and removals in nonpriority categories injures the States by making them
worse off than they were without it.

The States dispute the Department’s portrayal of the Guidance as mere prioritization. In
their view, the Guidance prohibits officials from arresting or removing certain people. But
nothing in the Guidance, to repeat, prohibits a single agent from detaining or removing a single
person or for that matter any category of noncitizens identified in the two statutes.

Even if the States cannot meet Article III’s “irreducible” standing requirements, Lujan,
504 U.S. at 560, they disclaim any need to do so. In their view, Massachusetts v. EPA, 549 U.S.
497 (2007), relaxed the Constitution’s standing requirements if the claimant is a sovereign.
There is something to the point but not as much as the States make of it. Start with what
Massachusetts v. EPA does not say. It does not remove Article III’s imperative of a cognizable
case or controversy or the requirements of injury, causation, and redressability. Think of it this
way. Had the States of Arizona, Montana, and Ohio challenged the Secretary of the Interior’s
reprioritization of activities to protect endangered species in Lujan itself, it is difficult to believe
that the Court would have found the injuries any less speculative or conjectural in terms of
causation and redressability. In that sense, Article III’s foundational standing requirements
remain for private and public litigants alike.

What Massachusetts v. EPA does show is that States sometimes are entitled to “special
solicitude” in this area because they may incur “quasi-sovereign” injuries that private parties
cannot. 549 U.S. at 517–20. But while the States may have more theories of injury available to
them, that does not allow them to bypass proof of injury in particular or Article III in general.
Saginaw County v. STAT Emergency Med. Servs., 946 F.3d 951, 957 (6th Cir. 2020).
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 11

The States’ alleged injuries in this case do not fall within Massachusetts v. EPA’s
compass. They do not protest regulation of them as States or preemption of local lawmaking
authority. They do not protest any threatened incursions on their property or territory. And they
do not involve the “classic” sovereign case, “public nuisances,” in which a State invokes a desire
“to safeguard its domain and its health, comfort and welfare.” Kentucky v. Biden, 23 F.4th 585,
596 (6th Cir. 2022) (quotation omitted); see Alfred L. Snapp & Son, Inc. v. Puerto Rico ex rel.
Barez, 458 U.S. 592, 602–03 (1982).

Their main objection is to indirect fiscal burdens allegedly flowing from the Guidance.
But why would that humdrum feature of a regulation count as a uniquely sovereign harm? Most
regulations have costs. A State has no more, and no less, reason to fear harms to its bottom line
from federal regulations than a person or a business does. Massachusetts v. EPA itself relied on
authority that distinguished quasi-sovereign interests from those “capable of estimate in money.”
549 U.S. at 518–19 (quotation omitted). Are we really going to say that any federal regulation of
individuals through a policy statement that imposes peripheral costs on a State creates a
cognizable Article III injury for the State to vindicate in federal court? If so, what limits on state
standing remain? Even though it “would make a mockery . . . of the constitutional requirement
of case or controversy,” the States’ boundless theory of standing—in which all peripheral costs
imposed on States by actions of the President create a cognizable Article III injury—would allow
them to challenge a “disagreeable war.” Alexander Bickel, The Voting Rights Cases, 1966 Sup.
Ct. Rev. 79, 89–90 (1966). That is a bridge much too far.

The States persist that the Supreme Court already accepted their theory of injury in
Department of Commerce v. New York, 139 S. Ct. 2551 (2019). Not so. In that case, States
alleged that the addition of a citizenship question on the census would cause noncitizen residents
to fail to respond to it, leading to a direct loss of federal funds that are distributed based on state
population. The Court ruled that such a loss of federal funds was “sufficiently concrete and
imminent” to satisfy Article III. Id. at 2565. It reached that conclusion in part because adding
the question would reduce response rates among noncitizens, rendering the causal link between
the addition of the question and the loss of federal funds sufficiently direct. Id. at 2565–66. By
contrast, this Guidance does not impose any direct costs on the States or threaten the loss of any
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 12

federal funding. Any downstream costs of the Guidance to the States come about via individual
officers’ discretionary enforcement choices, noncitizens’ actions in response to those choices, the
States’ own crime-and-punishment decisions, and the States’ other social-welfare policy choices.

The States also claim that harm to their interest in excluding people who have no right to
be in their territory is a quasi-sovereign injury that entitles them to special solicitude under
Massachusetts v. EPA. But even if a State can be distinguished from a private entity in this way,
that does not show that the States are entitled to a break. The key sovereign with authority and
“solicitude” with respect to immigration is the National Government, not the States. See
Arizona, 567 U.S. at 394–97. The States have distinctly less, not more, solicitude in this area.
Unlike a claim about potential losses of territory and property, as in Massachusetts v. EPA, a
State’s flawed assumption about the authority to exclude individuals offers no purchase for
relaxing the causation and redressability inquiries. Even if the States somehow had some
authority to prevent people from entering or leaving their territory, it would not free them from
establishing causation and redressability. The States express their injury in terms of rising costs
from crime and public services associated with playing host to more noncitizens. When defined
in terms of the costs of crime, the States’ injury fails to satisfy Lujan. How can we assume that
prioritizing apprehension of immigrants who pose a threat to public safety will drive up the
States’ criminal populations? Lujan, 504 U.S. at 564–67. A theory of injury grounded in rising
crime rates seems like it would “hinge” on third parties committing more crimes. Id. at 562. As
for the increased cost of public services, that requires showing, as noted, that the Guidance
would be the cause. With or without Massachusetts, there are many dubious justiciability
questions with respect to the States’ theory of standing—enough for us to be skeptical at this
stage of the case that they can bring the action.

Reviewability. We also doubt that the Administrative Procedure Act permits review of
the Guidance. The Act generally allows anyone “adversely affected or aggrieved by agency
action” to seek judicial review. 5 U.S.C. § 702. Even if we assume for the sake of this argument
that the Guidance has “adversely affected” the three States, the Act permits federal courts to
review only “final agency action.” Id. § 704. To qualify, the action must (1) “mark the
consummation of the agency’s decisionmaking process” and (2) be an action “by which rights or
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 13

obligations have been determined, or from which legal consequences will flow.” U.S. Army
Corps of Eng’rs v. Hawkes Co., 578 U.S. 590, 597 (2016) (quotation omitted). The Guidance
counts as the consummation of the Department’s decisionmaking, all sides agree. Today’s
dispute turns on the second inquiry.

We have some guidance and sharpening inquiries of our own on this point. Will the
agency’s action “impose liability” on a regulated party, create legal rights, or “mandate, bind, or
limit other government actors” in the future? Parsons v. U.S. Dep’t of Just., 878 F.3d 162, 169
(6th Cir. 2017). And will the agency’s action have “a sufficiently direct and immediate impact
on the aggrieved party and a direct effect on its day-to-day business”? Berry v. U.S. Dep’t of
Lab., 832 F.3d 627, 633 (6th Cir. 2016) (quotation omitted). If an action maintains officials’
“independent decisionmaking” and can be “discretionarily relied on,” it likely lacks legal effect.
Parsons, 878 F.3d at 170. Through it all, we will not overlook whether the agency’s action puts
a party to a “Catch-22,” stuck between heavy compliance costs or feared liability, neither of
which can be undone. Air Brake Sys. v. Mineta, 357 F.3d 632, 645 (6th Cir. 2004).

Viewed through the lens of these “legal effect” considerations, the Guidance likely is not
reviewable. Start with the revealing language of its action: “Guidelines for the Enforcement of
Civil Immigration Law.” R.4-1 at 1. “Guidelines” do not evoke binding legal effect. Consistent
with its label, the Guidance couches its instructions on lots of conditional language that preserves
officials’ discretion. The document provides a “not exhaustive” list of factors as “examples” of
what officials should consider. Id. at 3–4. It allows officials to make decisions “depending on
the facts.” Id. at 4. It cautions that it “does not compel an action to be taken or not taken” but
“leaves the exercise of prosecutorial discretion to the[ir] judgment.” Id. at 5. Even when the
Guidance uses the word “requires,” it does so in the context of “an assessment of the individual
and the totality of the facts and circumstances.” Id. at 3. Capping it off, the Guidance makes
clear that it “is not intended to, does not, and may not be relied upon to create any right or
benefit.” Id. at 7. These are telltale signs all of a nonbinding policy statement, not of reviewable
agency action.

Neither does the Guidance place the States in a regulatory Catch-22. Whatever costs the
Guidance creates for the States downstream arise only from officials who exercise their
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 14

discretion under the Guidance, confirming that those costs are not the Guidance’s “direct or
appreciable legal” consequences. Parsons, 878 F.3d at 170. For similar reasons, the Supreme
Court held that the base-closing commission’s recommendations (which the President had to
accept or reject) lacked finality because they were not the “action that will directly affect the
military bases.” Dalton v. Specter, 511 U.S. 462, 469 (1994) (quotation omitted). Even if the
Guidance creates some still-to-be-determined costs for the three States, it is well to remember
that “adverse economic effects accompany many forms of indisputably non-final government
action.” Air Brake Sys., 357 F.3d at 645.

The States call our attention to the Guidance’s section that explains that it “will become
effective in sixty (60) days” and that all “[a]gency leaders . . . will implement this guidance
accordingly.” R.4-1 at 6–7. The same would be true for any nonbinding policy statement.
Everything has a beginning and an end. Just as this interpretive and policy Guidance brings an
end to the former set of guidelines—“rescind[ing]” the prior interim guidance, id. at 6—this
Guidance surely will be modified at some point, whether by this administration or a future one.
The existence of effective dates does not tell us whether the Guidance has the kind of legal effect
that makes it reviewable.

In like vein, the States point to training to implement the Guidance, the transfer of agents
to the border under it, and the goal of producing uniformity in practice, all of which (they claim)
show that the Guidance will have a greater impact than its flexible and no-legal-effect language
suggests. But any policy statement could, indeed likely would, lead to training, reprioritization
of employees, and uniformity. All of this amounts to the kind of “practical consequences” that
come with most policy statements as opposed to the “direct or appreciable legal” consequences
that come with agency rules and other reviewable agency action. Parsons, 878 F.3d at 170.

The Supreme Court’s decision in Hawkes Co. does not say otherwise. In that case, the
U.S. Army Corps of Engineers issued a “jurisdictional determination.” 578 U.S. at 595–96
(quotation omitted). Such determinations definitively state whether the Clean Water Act applies
to a property and bind the Corps and the Environmental Protection Agency for five years. Id. at
594–95. The “definitive nature” of this decision, the Supreme Court explained, made it final
agency action, as the determination that the Act did not apply to a property bound the two
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 15

agencies and created “a five-year safe harbor from [enforcement] proceedings for a property
owner.” Id. at 598 (quotation omitted). In a similar vein, the Supreme Court recently addressed
whether the Department’s Memorandum rescinding its “Remain in Mexico” program was final
agency action. Biden v. Texas, 597 U.S. ___ , ___ (2022) (slip op. at 2, 18). Under the “Remain
in Mexico” rules, the Department returned certain noncitizens to Mexico pending their removal
proceedings, all with the cooperation of the Government of Mexico. Id. at ___ (slip op. at 2).
The Supreme Court held that the Department’s Memorandum formally rescinding the program
was final agency action because it “bound DHS staff by forbidding them to continue the program
in any way from that moment on” and thus determined legal rights and obligations. Biden v.
Texas, 597 U.S. ___ , ___ (2022) (slip op. at 20) (quotation omitted) (emphasis added). The
Department’s nonbinding Guidance, by contrast, does not rule out any enforcement action
against any noncitizen or create a safe harbor from enforcement of any kind.

Confirming that the Guidance lacks legal effect is the reality that it is difficult to see how
any noncitizen—or any person at all—could invoke it to establish legal protection. We are not
aware of any such instance with respect to the Guidance or its prior incarnations. The States
claim that a page on the Department’s website shows otherwise. But that is not true. The
website contains a page entitled “Contact ICE About an Immigration/Detention Case.” U.S.
Immigr. & Customs Enf’t, Contact ICE About an Immigration/Detention Case (updated Mar. 9,
2022), https://perma.cc/RP5T-NNA3. It says that if “a noncitizen or their representative believes
they do not meet DHS’ priorities for enforcement, they are encouraged to” contact the
Department “to request a case review.” Id. But a “Contact Us” page for requesting the exercise
of prosecutorial discretion does not show that the Guidance confers any legal right. Nothing
about this service suggests that the Guidance can be used to seek relief in court. Nor does it
undo the Guidance’s own disclaimer that it confers no legal rights or benefits.

What of the possibility that we should be skeptical in the other direction—that we should
consider the possibility that the Department labeled its directive “Guidance” and included
discretion-conferring language in order to avoid review of these policies? Labels, it is true, do
not control the inquiry. Legal effects do. But there is nothing in the record to indicate that the
Guidance has binding legal effects. The combined realities that the relevant statutes have many
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 16

moving parts, the Guidance leaves considerable discretion in implementing it, and the Guidance
does not create any legal rights for noncitizens all suggest it is not reviewable.

The Guidance is likely unreviewable on the additional ground that enforcement priorities
are “committed to agency discretion by law.” 5 U.S.C. § 701(a)(2). It may be true that agency
action is presumptively reviewable under the Administrative Procedure Act. Dep’t of Com., 139
S. Ct. at 2567. But the opposite is true for “[r]efusals to take enforcement steps.” Heckler v.
Chaney, 470 U.S. 821, 831 (1985). An agency’s choice “not to prosecute or enforce, whether
through civil or criminal process, is a decision generally committed” to an agency’s “discretion.”
Id. No doubt, that approach does not apply when Congress, as opposed to the agency, sets all of
the marching orders. Id. at 834–35. But as shown above and below, Congress did not remove all
discretion from the Department in making removal and detention decisions.

III.

Even if the States cleared these two justiciability hurdles, they are unlikely to succeed on
the merits of their claim that the Guidance violates the Administrative Procedure Act, whether on
the grounds that it is contrary to law, it is arbitrary or capricious, or it lacks a required notice and
comment. 5 U.S.C. §§ 706(2), 553.

Contrary to law. The States allege that the Guidance violates detention and removal
instructions in two immigration statutes by compelling officers to consider a host of factors
before making decisions that Congress made mandatory and precluded from turning on multiple
immigration-policy considerations. As the States emphasize, §§ 1226(c)(1) and 1231(a)(1)(A)
require the Department to arrest and remove certain noncitizens, but the Guidance “forbids field
officers from taking these actions based on an alien’s eligibility for removal under these statutes”
and instead lets officers do so “only if the extra-statutory factors in the Policy are satisfied.”
States’ Br. 40–41, 42.

But this claim must account for the considerable discretion already embedded in the
immigration system. “A principal feature of the removal system is the broad discretion exercised
by immigration officials.” Arizona, 567 U.S. at 396. Those officials, “as an initial matter, must
decide whether it makes sense to pursue removal at all.” Id. There are “various stages in the
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 17

deportation process,” and “[a]t each stage the Executive has discretion to abandon the endeavor.”
Reno v. American-Arab Anti-Discrimination Comm., 525 U.S. 471, 483 (1999). All of this
explains why Congress has charged the Department, not the federal courts, with “[e]stablishing
national immigration enforcement policies and priorities.” 6 U.S.C. § 202(5).

The question is not whether 8 U.S.C. §§ 1226(c)(1) and 1231(a)(1)(A) have mandatory
language. It is whether this mandatory language displaces the Department’s longstanding
discretion in enforcing the many moving parts of the nation’s immigration laws. We think it
unlikely that either statute creates a judicially enforceable mandate that the Department arrest or
remove certain noncitizens.

Look first at 8 U.S.C. § 1226(c)(1). It provides that the Department “shall take into
custody any alien who” is removable for specified reasons (including those convicted of crimes
with respect to terrorism, aggravated felonies, firearms, drugs, and moral turpitude) “when the
alien is released,” presumably meaning released from state or federal custody for a qualifying
offense. 8 U.S.C. § 1226(c)(1). Two problems undercut the bright line, judicially enforceable
rule that the States claim that § 1226(c)(1) creates. One turns on the reality that, by prioritizing
efforts to prevent terrorism, to protect public safety, and to ensure border security, the Guidance
does not necessarily violate a single word of the statute. After all, the identified crimes all fit
within the category of ensuring public safety and protecting against terrorism. Nor is it
problematic to prioritize anti-terrorism by name. Nothing seems to prevent the Department from
taking the five categories of crime and saying that we will prioritize not-unlimited detention and
removal resources in the order of these crimes: terrorism, aggravated felonies, firearm offenses,
drug offenses, and moral-turpitude offenses. The States acknowledge that, in a world with
scarce resources, the Department may guide individual officers’ discretion with prioritization
schemes. This Guidance does just that. And this would not be the first administration to use
triage in enforcing immigration laws.

The second problem is that, while the provision says that the Department must take
certain people, including the ones that commit these crimes, into custody, it does not say how
long they must remain in custody or even ensure they must immediately be taken into custody.
Keep in mind that this binding duty applies “pending a decision on whether the alien is to be
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 18

removed.” Id. § 1226(a). Immigration authorities, as the Supreme Court has made clear, have
considerable discretion over whom to arrest and remove. Arizona, 567 U.S. at 396. That means
the Department does not necessarily have an obligation to take a noncitizen into custody, or keep
them in custody, if they decide at that point or later on not to bring an enforcement action. At
that point, there would no longer be a “pending . . . decision on whether the alien is to be
removed.” 8 U.S.C. § 1226(a).

Move to 8 U.S.C. § 1231(a)(1)(A). “Except as otherwise provided,” it says that, “when
an alien is ordered removed, the Attorney General shall remove the alien from the United States
within a period of 90 days.” Id. But Congress itself appreciated that removal would not always
occur within 90 days. It permitted supervised release—release from custody—“[i]f the alien
does not leave or is not removed within the removal period.” Id. § 1231(a)(3). Combined with
the basic principle that “[a]t each stage” of the removal process, “the Executive has discretion to
abandon the endeavor” to remove someone, Reno, 525 U.S. at 483, all of this means that
immigration officials retain some discretion not to execute a final order of removal within 90
days. That explains why the Supreme Court expressed “doubt” that, “when Congress shortened
the removal period to 90 days in 1996 it believed that all reasonably foreseeable removals could
be accomplished in that time.” Zadvydas v. Davis, 533 U.S. 678, 701 (2001). Which
presidential administration since this law came into effect in 1996, it is fair to wonder, has come
close to removing all eligible noncitizens within 90 days, whether with respect to statutorily
permitted reasons or not?

Even so, both statutes say “shall,” the States insist, connoting a command, particularly
when contrasted with the use of “may” elsewhere in both statutes. But the use of “shall” does
not automatically create a judicially enforceable mandate, especially when criminal or civil law
enforcement is at issue. Town of Castle Rock v. Gonzales, 545 U.S. 748, 761–62 (2005) (statute
saying that officers “shall arrest” did not eliminate police discretion whether to arrest a violator).
“[C]ommon sense” dictates that law enforcement officers generally retain “deep-rooted”
discretion “even in the presence of seemingly mandatory legislative commands.” Id. at 761
(quotation omitted). Even an “express statutory deadline” does not necessarily mean “Congress
intended for courts to enforce the deadline.” See Nielsen v. Preap, 139 S. Ct. 954, 969 n.6
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 19

(2019); see also United States v. James Daniel Good Real Prop., 510 U.S. 43, 63 (1993). We
see no “stronger indication” from Congress in these statutes that “shall” creates a judicially
enforceable mandate. Castle Rock, 545 U.S. at 761. The context in fact cuts the other way.
There are many moving parts in immigration law, and we doubt these laws completely eliminate
the Department’s discretion to decide whom to charge, whom to remove, and when to do so.

The States, moreover, have a juxtaposition problem of their own. Another provision,
8 U.S.C. § 1231(a)(2), says that, “[d]uring the removal period, the Attorney General shall detain
the alien. Under no circumstance during the removal period shall the Attorney General release
an alien who has been found inadmissible” for the same types of offenses captured by § 1226(c).
Id. (emphasis added). Notably, the States do not complain that the Guidance violates this
provision, and the Department acknowledges that it must follow this under-no-circumstance
directive. Having argued that the juxtaposition between the “may” and “shall” language in the
two statutes supports it position, the States must acknowledge that the juxtaposition between the
“shall” and “under no circumstance” language supports the Department’s position. Congress’s
deployment of even more directive language in the same statutory scheme counsels against
finding the “stronger indication” of Congressional intent necessary to infer a judicially
enforceable mandate from the mere use of “shall.” Castle Rock, 545 U.S. at 761–62.

The States also suggest that the Supreme Court has ratified their view of these statutes.
True, some cases parrot the “shall” language in describing the obligations these statutes place on
the Department. See Nielsen, 139 S. Ct. at 966; Johnson v. Guzman Chavez, 141 S. Ct. 2271,
2281 (2021). But these cases “are ones in which detainees subject to enforcement action were
seeking their release.” Texas, 14 F.4th at 338. In explaining that detainees are not entitled to
bond hearings or release under these statutes, the Court had no occasion to consider whether the
statutes subject the Department to a judicially enforceable mandate to arrest and remove all
noncitizens covered by these provisions in the first place. See also Biden, 597 U.S. at ___ (slip
op. at 14 n.5, 18) (declining to decide whether a different statute providing that certain
noncitizens “shall be detained” creates a judicially enforceable mandate against the Department).
That question comes with complexities all its own, especially in the context of the light cast by
the Executive Branch’s traditional prosecutorial discretion and the reality of scarce resources.
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 20

Not every “shall” directive in a federal immigration statute, it turns out, necessarily
creates a judicially enforceable mandate. That is in part because the Executive Branch has
considerable enforcement discretion in deploying limited resources to address its policy
challenges. And that is in part to preserve bedrock separation of powers. It takes little
imagination to envision the difficulty the Judicial Branch would face in trying to ensure that
immigration officers enforce federal laws like these just the way some States would like them to.
If it is fair to worry from time to time about the risks when executive-branch agencies exercise
legislative and judicial power, it is equally fair to worry when judges are called into disputes that
turn principally on policy and resource debates between the First and Second Branches.

Arbitrary or capricious. The APA requires that “agency action be reasonable.” FCC v.
Prometheus Radio Project, 141 S. Ct. 1150, 1158 (2021). Review is “deferential,” and “a court
may not substitute its own policy judgment for that of the agency.” Id. We seek merely to
ensure that the agency considered relevant data points and offered a satisfactory explanation for
its decision. Id. At this stage, we cannot say that the Department failed to address relevant
concerns or explain itself. The agency considered the problem of recidivism. It explained that
the Guidance’s call for context-specific consideration of a noncitizen’s circumstances is meant to
assess “whether a noncitizen poses a current threat to public safety, including through a
meaningful risk of recidivism.” R.27-2 at 12. The agency also considered the effect of its
Guidance on the States. Its Considerations Memo includes an “Impact on States” section in
which it reasoned that the Guidance’s fiscal impact on States “would vary based on a range of
factors,” is “difficult to quantify,” and might be offset at least in part by cost savings for States
due to the “implementation of priorities guidance” aimed at public safety threats. Id. at 15. The
States do not suggest that the agency had to calculate the costs of its Guidance on States, and the
States themselves have not offered any concrete evidence of the Guidance’s fiscal effects on
each of them.

The Secretary also offered a satisfactory explanation for the priorities. He emphasized
that the Department needs to “make smart and strategic choices about how to utilize” limited
resources in enforcing the nation’s immigration laws. Id. at 5. And he added that the
Department’s mission “is not best served by simply pursuing the greatest overall number of
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 21

enforcement actions but is rather best advanced by directing resources to prioritize enforcement
against those noncitizens who most threaten the safety and security of the Nation.” Id. at 17.
This was not arbitrary or capricious—at least not likely so at this stage of the case.

Notice and comment. The Guidance did not need to go through notice and comment. By
statute, that requirement does not apply to “general statements of policy,” 5 U.S.C. § 553(b)(A),
which include “statements issued by an agency to advise the public prospectively of the manner
in which the agency proposes to exercise a discretionary power,” Lincoln v. Vigil, 508 U.S. 182,
197 (1993) (quotation omitted). This Guidance fits that bill.

True, an agency cannot avoid procedural requirements via a self-serving label, calling
something a “policy” or “guidance” when it amounts to a legislative rule. See Azar v. Allina
Health Servs., 139 S. Ct. 1804, 1812 (2019). The content of the agency’s action, not its name,
shapes the inquiry. Id. But, to repeat, the Guidance “does not compel” any action, “leaves the
exercise of prosecutorial discretion to the judgment of” federal personnel, and does not create
any “right or benefit . . . enforceable at law.” R.4-1 at 5, 7. It does not bear the hallmarks of a
substantive rule because it does not legally “affect[] individual rights and obligations.” Chrysler
Corp. v. Brown, 441 U.S. 281, 302 (1979) (quotation omitted).

IV.

The other pertinent factors counsel against preliminary injunctive relief as well. The
absence of a preliminary injunction should not irreparably injure the three States. Yes, if they
are right, the Guidance may impose costs on them that are difficult to recover. But the extent of
those costs is filled with ifs and maybes, particularly given the reality that the States concede that
the relevant federal statutes do not tell the Department how to deploy its resources, do not stop it
from setting prioritization categories, and do not prevent it from sending its enforcement agents
wherever it wishes. If federal law permits all of that, it is hard to see how vacating the
preliminary injunction will result in substantial and distinct injuries to the three States. The
preliminary injunction likely causes irreparable harm to the Department by interfering with its
authority to exercise enforcement discretion and allocate resources toward this administration’s
priorities. And in view of our doubts about the States’ claims under the Administrative
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 22

Procedure Act, the public interest favors reversal. Most of the States’ contrary arguments pivot
on the assumption that the Guidance is illegal, an assumption that we have not found persuasive
thus far. All in all, the equity scale tips in the National Government’s favor.

We reverse and remand for further proceedings.
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 23

_________________

CONCURRENCE
_________________

SUTTON, Chief Judge, concurring. The district court erred in issuing this preliminary
injunction for two additional reasons.

Reason One: It lacked the authority to do so under 8 U.S.C. § 1252(f)(1). “Regardless of
the nature of the action or claim or of the identity of the party” bringing the action, it says, “no
court (other than the Supreme Court) shall have jurisdiction or authority to enjoin or restrain the
operation of the provisions of” 8 U.S.C. §§ 1221–1232, “other than with respect to the
application of such provisions to an individual alien against whom proceedings under such part
have been initiated.” The provision prohibits “lower courts from entering injunctions that order
federal officials to take or to refrain from taking actions to enforce, implement, or otherwise
carry out the specified statutory provisions.” Garland v. Aleman Gonzalez, 596 U.S. ___, ___
(2022) (slip op. at 5).

The district court entered a preliminary injunction that “enjoined and restrained” the
Department from “[e]nforcing and implementing” the Guidance in a manner that violated the
district court’s interpretation of §§ 1226(c)(1) and 1231(a)(1)(A). R.44 at 78–79. In effect, the
district court’s remedy forbids the Department from implementing a guidance that the agency
views as consistent with those statutes and compels the Department to exercise its enforcement
authority according to the district court’s reading of the statutes. The order violates § 1252(f)(1)
because it “require[s] officials to take actions that (in the Government’s view) are not required
by [§§ 1226(c)(1) and 1231(a)(1)(A)] and to refrain from actions that (again in the Government’s
view) are allowed by [§§ 1226(c)(1) and 1231(a)(1)(A)].” Aleman Gonzalez, 596 U.S. at ___
(slip op. at 7).

Not so, the States resist. They say they did not sue to enjoin “the operation of either
statute” but to enjoin an administrative policy that “thwarts the operation of those statutes.”
States’ Br. 58–59. But § 1252(f)(1) has the same force even when the National Government
allegedly enforces the relevant statutes unlawfully. Else, it would not be much of a prohibition.
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 24

See Aleman Gonzalez, 596 U.S. at ___ (slip op. at 7) (observing that “it is very common to refer
to the ‘unlawful’ or ‘improper’ operation of whatever it is that is being operated”). The
Guidance represents the Department’s effort at implementing §§ 1226(c)(1) and 1231(a)(1)(A)
by prioritizing the use of scarce resources. By prohibiting the Department from relying on the
Guidance in making these decisions, the district court’s injunction “order[s] federal officials
to . . . refrain from taking actions to . . . implement . . . the specified statutory provisions.” Id. at
___ (slip op. at 5).

The States protest that the National Government forfeited this point by raising it for the
first time on appeal. The U.S. Supreme Court recently explained that § 1252(f)(1) “does not
deprive the lower courts of all subject matter jurisdiction over claims brought under sections
1221 through 1232 of the” Immigration and Nationality Act, but at the same time it declined to
decide whether the provision’s limitation on injunctive relief “is subject to forfeiture.” Biden v.
Texas, 597 U.S. ___ , ___ (2022) (slip op. at 9, 12 n.4). Even assuming that parties may forfeit
such a remedial challenge, “we have discretion to excuse forfeitures and compelling reasons to
do so here.” Moody v. United States, 958 F.3d 485, 493 (6th Cir. 2020); see Scottsdale Ins. Co.
v. Flowers, 513 F.3d 546, 552 (6th Cir. 2008) (noting that courts may excuse a forfeiture in
“exceptional cases or particular circumstances”). The district court awarded the States the
“extraordinary remedy” of a preliminary injunction, Winter v. Nat. Res. Def. Council, Inc., 555
U.S. 7, 24 (2008), and purported to apply that remedy nationwide. Section 1252(f)(1) plainly
prohibits that grant of relief. This is precisely the kind of exceptional case in which a forfeiture
should be excused.

The Supreme Court, I appreciate, has not addressed whether § 1252(f)(1) prohibits
declaratory relief or prevents district courts from setting agency action aside under the APA.
Aleman Gonzalez, 596 U.S. at ___ (slip op. at 7 n.2); Biden, 597 U.S. at ___ (slip op. at 12 n.4).
But because the district court entered preliminary injunctive relief here, I need not address these
other remedies. It suffices to say that the district court “exceeded [its] jurisdiction in awarding”
preliminary injunctive relief. Aleman Gonzalez, 596 U.S. at ___ (slip op. at 3). The preliminary
injunction should be vacated for this independent reason.
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 25

Reason Two: The scope of the district court’s remedy—universally enjoining the
National Government from enforcing the Guidance in any State in the country—also likely
exceeded its authority apart from the statutory restriction. I do not take issue with the court’s
decision to extend the remedy beyond the Southern District of Ohio as to the three state
claimants. When “exercising its equity powers,” a district court “may command persons
properly before it to cease or perform acts outside its territorial jurisdiction.” Steele v. Bulova
Watch Co., 344 U.S. 280, 289 (1952). But it is one thing to honor a federal court judgment
issued in favor of, say, Arizona by the Southern District of Ohio anywhere in the country. It is
quite another to do so for the 47 States that did not participate in the lawsuit. I am not the first to
question nationwide (or universal) injunctions (or remedies) that bar the federal government
from enforcing a law or regulation anywhere and against anyone. See, e.g., Trump v. Hawaii,
138 S. Ct. 2392, 2424–29 (2018) (Thomas J., concurring); Dep’t of Homeland Sec. v. New York,
140 S. Ct. 599, 599–601 (2020) (mem.) (Gorsuch, J., concurring); CASA de Md., Inc. v. Trump,
971 F.3d 220, 256–63 (4th Cir. 2020) (vacated on other grounds); Samuel Bray, Multiple
Chancellors: Reforming the National Injunction, 131 Harv. L. Rev. 417, 457–82 (2017).

I meet this concept with considerable skepticism. Article III grants the “judicial Power,”
which extends only to specified “Cases” and “Controversies.” U.S. Const., art. III, § 2. Standing
limitations, a prohibition on advisory opinions, distinctions between judgments and opinions all
grow out of this language and the history behind it.

The same is true of remedies, which emerge from a federal court’s equitable power. A
valid Article III remedy “operate[s] with respect to specific parties,” not with respect to a law “in
the abstract.” California v. Texas, 141 S. Ct. 2104, 2115 (2021) (quotation omitted). That is
why courts generally grant relief in a party-specific and injury-focused manner. See Gill v.
Whitford, 138 S. Ct. 1916, 1934 (2018). In this same way, we do not remove—“erase”—from
legislative codes unconstitutional provisions. Jonathan Mitchell, The Writ–of–Erasure Fallacy,
104 Va. L. Rev. 933, 1016–17 (2018). We merely refuse to enforce them in a case, thereby
exercising “the negative power to disregard an unconstitutional enactment.” Massachusetts v.
Mellon, 262 U.S. 447, 488 (1923). After a court has remedied a claimant’s injury, it is fair to ask
what controversy remains for a court to adjudicate or remedy.
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 26

Call them what you will—nationwide injunctions or universal remedies—they seem to
take the judicial power beyond its traditionally understood uses, permitting district courts to
order the government to act or refrain from acting toward nonparties in the case. The law
already has a mechanism for applying a judgment to third parties. That is the role of class
actions, and Civil Rule 23 carefully lays out the procedures for permitting a district court to bind
nonparties to an action. Nationwide injunctions sometimes give States victories they did not
earn and sometimes give States victories they do not want. They always sidestep Rule 23’s
requirements.

Such injunctions create practical problems too. The effect of them is to prevent the
National Government from enforcing a rule or executive order without (potentially) having to
prevail in all 94 district courts and all 12 regional courts of appeals. They incentivize forum
shopping. They short-circuit the decisionmaking benefits of having different courts weigh in on
vexing questions of law and allowing the best ideas to percolate to the top. They lead to rushes
to judgment. And all of this loads more and more carriage on the emergency dockets of the
federal courts, a necessary feature of any hierarchical court system but one designed for
occasional, not incessant, demands for relief.

Confirming the point, a district court considering the validity of this same Guidance has
recently assumed authority for the issue across the nation: It vacated the agency action
“universally.” Texas v. United States, No. 6:21-CV-00016, 2022 WL 2109204, at *46 (S.D. Tex.
June 10, 2022). Although that decision does not address the preliminary-injunction posture
presented here, the same insight applies. Reasonable jurists may sometimes reach different
outcomes on the merits of tough questions. Yet one decision against the government by one
judge potentially freezes the debate for all lower court judges.

At a minimum, a district court should think twice—and perhaps twice again—before
granting universal anti-enforcement injunctions against the federal government. Even if it turns
out that the three States in this case are entitled to relief, it is difficult to see why an injunction
applicable only to them would not do the trick.
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 27

The contrary arguments are unconvincing. The Administrative Procedure Act, it is true,
says that a reviewing court may “hold unlawful and set aside” agency actions that violate the
law. 5 U.S.C. § 706(2). But that raises a question; it does not answer it. The question is
whether Congress meant to upset the bedrock practice of case-by-case judgments with respect to
the parties in each case or create a new and far-reaching power through this unremarkable
language. We presume that statutes conform to longstanding remedial principles. Nken v.
Holder, 556 U.S. 418, 433 (2009); Weinberger v. Romero-Barcelo, 456 U.S. 305, 320 (1982).
And it is far from clear that Congress intended to make such a sweeping change. Compare Bray,
supra, at 438 n.121; and John Harrison, Section 706 of the Administrative Procedure Act Does
Not Call for Universal Injunctions or Other Universal Remedies, 37 Yale J. Reg. Bull. 37, 41–47
(2020); with Mila Sohoni, The Power to Vacate a Rule, 88 Geo. Wash. L. Rev. 1121, 1191–92
(2020). Use of the “setting aside” language does not seem to tell us one way or another whether
to nullify illegal administrative action or not to enforce it in the case with the named litigants.
Observe as well that the statute does not say against whom an unlawful agency action must be
“set aside.” In the context of a law authorizing identifiable “person[s]” “aggrieved by agency
action” to seek judicial review, 5 U.S.C. § 702, we should not lightly conclude that a court is
entitled to “set aside” agency action against persons not privy to the case before it. For that
reason, I would be inclined to stand by the long-understood view of equity—that courts issue
judgments that bind the parties in each case over whom they have personal jurisdiction.

What of the English courts’ use of a “bill of peace” at the founding? Does that permit a
contrary, and historically grounded, justification for nationwide injunctions? At common law,
this device allowed courts to resolve multiple suits involving common claims, say with several
tenants suing one [land]lord. Bray, supra, at 426. In one sense, it is true, bills of peace allowed
English courts to adjudicate the rights of members of dispersed groups without formally joining
them to a lawsuit through the usual procedures. See id.; Amanda Frost, In Defense of
Nationwide Injunctions, 93 N.Y.U. L. Rev. 1065, 1080–81 (2018). But this observation tees up
the question rather than answering it. Question: Does the bill of peace have a modern analog?
Answer: Yes, class actions under Civil Rule 23. Bray, supra, at 426. Nationwide injunctions
depart from, they do not embrace, traditional equity practices incorporated into the Federal Rules
of Civil Procedure. The domesticated animal known as a bill of peace looks nothing like the
 No. 22-3272 Arizona, et al. v. Biden, et al. Page 28

dragon of nationwide injunctions. A bill of peace applied to small, cohesive groups. And
because it was representative in nature, any decision, win or lose, bound any nonparty members
of the group to the judgment. Id.; Michael T. Morley, Disaggregating Nationwide Injunctions,
71 Ala. L. Rev. 1, 36–37 (2019). Not so for today’s asymmetric, applicable-against-the-world
injunctions. Use of the common law bill of peace to justify today’s nationwide injunction gives
analogy—and history—a bad name.

The district court separately feared that a narrower injunction “would create a patchwork
immigration enforcement system,” R.44 at 78, instead of a “comprehensive and unified” one,
Arizona v. United States, 567 U.S. 387, 401 (2012). But that justification lacks a limiting
principle and would make nationwide injunctions the rule rather than the exception with respect
to all actions of federal agencies. That is especially troubling in the domain of immigration law,
where the federal Legislative and Executive Branches, not the Judicial Branch, are the key
drivers of national policy.

What of the States’ fear that a state-specific injunction “will not prevent DHS from
releasing” covered noncitizens “upon crossing state lines,” making universal relief necessary to
“fully redress the[se] States’ injuries”? States’ Br. 58. That argument, again, would permit a
nationwide injunction for any immigration-related claim by any one State. No less importantly,
the States have not offered any evidence to back up the point or to concretely illustrate its
consequences. Indeed, the National Government acknowledges that it cannot release criminal
noncitizens “once they are detained in the first instance,” Reply Br. 25, suggesting that the States
have nothing to fear in this respect. Even if this alleged injury were not speculative, it is
doubtful that a nationwide remedy was the narrowest way to cure it. Relatedly, the district court
worried that the Guidance could not “be applied on a state-by-state basis.” R.44 at 78. But that
is initially the National Government’s problem, not ours, and it indeed acknowledged that
severed policy enforcement remains a feasible alternative.

All in all, nationwide injunctions have not been good for the rule of law. Left unchecked,
such nationwide injunctions have become a springing easement on the customary deliberative
process for dealing with issues of national importance. The sooner they are confined to discrete
settings or eliminated root and branch the better.