Wilkins, Circuit Judge, concurring

Brinton Lucas, Counsel to the Assistant Attorney General, U.S. Department of Justice, argued the cause for appellants. With him on the briefs were Hashim M. Mooppan, Deputy Assistant Attorney General, Marleigh D. Dover, and Tara S. Morrissey, Attorneys.

Jennifer Levi argued the cause for appellees. With her on the brief were Paul R.Q. Wolfson, Washington, DC, John T. Byrnes, Washington, DC, Kevin M. Lamb, Washington, DC, Alan E. Schoenfeld, New York, NY, Shannon P. Minter, San Francisco, CA, and Christopher Stoll.

Maura Healey, Attorney General, Office of the Attorney General for the Commonwealth of Massachusetts, Robert E. Toone, Assistant Attorney General, Janet T. Mills, Attorney General at the time the brief was filed, Office of the Attorney General for the State of Maine, Brian E. Frosh, Attorney General for the State of Maryland, Lori Swanson, Attorney General at the time the brief was filed, Office of the Attorney General for the State of Minnesota, Gurbir S. Grewal, Attorney General, Office of the Attorney General for the State of New Jersey, Hector Balderas, Attorney General, Office of the Attorney General for the State of New Mexico, Barbara Underwood, Attorney General at the time the brief was filed, Office of the Attorney General for the State of New York, Joshua H. Stein, Attorney General, Office of the Attorney General for the State of North Carolina, Xavier Becerra, Attorney General, Office of the Attorney General for the State of California, George Jepsen, Attorney General at the time the brief was filed, Office of the Attorney General for the State of Connecticut, Matthew P. Denn, Attorney General at the time the brief was filed, Office of the Attorney General for the State of Delaware, Karl A. Racine, Attorney General, Office of the Attorney General for the District of Columbia, Russell A. Suzuki, Attorney General at the time the brief was filed, Office of the Attorney General for the State of Hawaii, Lisa Madigan, Attorney General at the time the b rief was filed, Office of th e Attorney Gen eral for the S tate of Illino is, Tom Miller, Attorney General, Office of the Attorney General for the State of Iowa, Peter F. Kilmartin, Attorney General at the time the brief was filed, Office of the Attorney General for the State of Rhode Island, Thomas J. Donovan, Jr., Attorney General, Offi ce of the Attorne y General for the State of Vermont, Mark R. Herring, Attorney General, Office of the Attorney General for the Commonwealth of Virginia, Ellen F. Rosenblum, Attorney General, Office of the Attorney General for the State of Oregon, and Josh Shapiro, Attorney General, Office of the Attorney General for the Commonwealth of Pennsylvania, were on the brief for amici curiae The States of Massachusetts, et al. in support of appellees and affirmance of the District Court decision.

Peter C. Renn, Diana K. Flynn, Chicago, IL, and Tara L. Borelli, Los Angeles, CA, were on the brief for amici curiae National Center for Transgender Equality and Other Advocacy Organizations in support of plaintiffs-appellees and affirmance.

Andrew J. Ehrlich, New York, NY, George W. Kroup, New York, NY, Eric A. Felleman, New York, NY, and Craig A. Benson, Washington, DC, were on the brief for amici curiae American Veterans Alliance, et al. in support of plaintiffs-appellees and in support of affirmance.

Eamon P. Joyce, New York, NY, John T. Hebden, Christopher A. Eiswerth, Washington, DC, and Robert S. Chang, Seattle, WA, were on the brief for amici curiae Asian American Legal Defense and Education Fund, et al. in support of plaintiffs-appellees.

Cynthia Cook Robertson, Washington, DC, and Suzanne B. Goldberg were on the brief for amici curiae The National Organization for Women Foundation, et al. in support of plaintiffs-appellees.

Douglas H. Hallward-Driemeier, Washington, DC, Irina Finkel, New York, NY, and Douglas E. Brayley, Boston, MA, were on the brief for amici curiae The Organization of Historians and 47 Historians of the Military, National Security, and Foreign Relations supporting plaintiffs-appellees.

Elizabeth B. Wydra, San Francisco, CA, and Ashwin P. Phatak were on the brief for amicus curiae Constitutional Accountability Center in support of plaintiffs-appellees.

Steven G. Thompson Reed, Washington, DC, was on the brief for amici curiae The Service Women’s Action Network and Other Veterans Service Organizations and Veterans Advocacy Groups in support of affirmance.

Harold Hongju Koh and Phillip Spector were on the brief for amici curiae Retired Military Offices and Former National Security Officials in support of plaintiffs-appellees.

Daniel S. Harawa, Sherrilyn A. Ifill, New York, NY, and Janai S. Nelson were on the brief for amicus curiae NAACP Legal Defense & Educational Fund, Inc. in support of appellees and affirmance.

Devi M. Rao, Washington, DC, was on the brief for amici curiae American Medical Association and Seven Other HealthCare Organizations in support of plaintiffs-appellees and affirmance.

Stuart F. Delery, Washington, DC, was on the brief for amicus curiae The Trevor Project in support of plaintiffs-appellees and affirmance.

Susan Baker Manning and Stephanie Schuster, Washington, DC, were on the brief for amici curiae Vice Admiral Donald C. Arthur, USN (Ret.) et al. in support of plaintiffs-appellees and affirmance of the District Court decision.

Before: Griffith and Wilkins, Circuit Judges, and Williams, Senior Circuit Judge.

Concurring opinion filed by Circuit Judge Wilkins.

Opinion concurring in the result filed by Senior Circuit Judge Williams.

Wilkins, Circuit Judge, concurring:"[T]he passage of time frequently brings about changed circumstances — changes in the nature of the underlying problem, changes in governing law or its interpretation by the courts, and new policy insights — that warrant reexamination of the original judgment." Horne v. Flores , 557 U.S. 433, 448, 129 S.Ct. 2579, 174 L.Ed.2d 406 (2009). As we described in the Judgment, the District Court’s finding of no changed circumstances and its denial of the motion to dissolve the preliminary injunction was error. I write separately to elaborate on why I believe that was the case.

First, we must review some history and background.

I.

At issue in this case is the regulation of military service by transgender persons. As noted in the report of the Transgender Military Service Commission chaired by former Surgeon General Jocelyn Elders, a prevalent theme in this area is that "regulatory terminology that references transgender identity is inconsistent." J.A. 753. Thus, to avoid confusion, it is critically important to define terms and characterize military regulations carefully and precisely.

Based on the record, transgender persons are "individuals who identify with a gender different from the sex they were assigned at birth." J.A. 606; see also J.A. 263. Thus, while a transgender woman may have been assigned the male sex at birth, she nonetheless identifies with the female gender. Similarly, a transgender man was assigned the female sex at birth but identifies with the male gender.

As explained by amici American Medical Association and seven other healthcare organizations, "[e]very person has a gender identity, which cannot be altered voluntarily or necessarily ascertained immediately after birth." AMA Amicus Br. 6. A person "communicates gender identity to others through behavior, clothing, hairstyles, voice, or body characteristics," id. , and the extent to which transgender individuals express their gender identity varies from person to person. As found by the RAND Corporation in a report for Former Secretary of Defense Ash Carter, some transgender persons express their gender identity though "transitioning," which is "the act of living and working as a gender different from that assigned at birth." J.A. 606 (emphasis added). But according to the information that we have in the record, only "[a] subset of transgender individuals may choose to transition ..." J.A. 606. As explained in the Transgender Military Service Commission report:

Being transgender does not mean that one has already transitioned to a different gender, or that such a transition will occur in the future. It means recognizing that the gender one has always had does not match the physical gender that was assigned at birth. The transgender community includes people who have already transitioned to the other gender, those who have not yet transitioned but who plan to do so, those who identify with the other gender but do not wish to transition, and others.

J.A. 752.

Some, but not all, transgender persons develop gender dysphoria, which manifests as stress and anxiety caused by the incongruence between the sex assigned to the person at birth and the person’s preferred gender identity. J.A. 622-23. Gender dysphoria is recognized by the Diagnostic and Statistical Manual of Mental Disorders (DSM-5), and is often treated with "psychotherapy, hormone therapy, surgery, and changes to gender expression and role (i.e., how people present themselves to the world ...)." J.A. 623. Thus, many transgender persons with gender dysphoria require medical treatment, including transitioning to their preferred gender; however, "[n]ot all [transgender persons diagnosed with gender dysphoria ] will prefer or need all or any of those options." Id .

Prior to 2015, the Department of Defense (DoD) effectively banned all transgender persons from either joining or remaining in the military. J.A. 275; J.A. 734; J.A. 784. These accession and retention restrictions were enforced through medical standards that precluded applicants from joining the military if they had "defects of the genitalia" including but not limited to change of sex, and through mental health standards that disqualified persons with any history of certain psychosexual conditions, including "transsexualism" and "transvestism." J.A. 754. Thus, all transgender persons were essentially banned from military service, even if they had never undergone sex-reassignment surgery, even if they did not have gender dysphoria, and even if they had not transitioned to their preferred gender and were willing to serve pursuant to the military standards applicable to the sex assigned to them at birth.

Under the pre-2015 regime, the only thing that mattered to exclude a person from military service was the person’s transgender status: that the person did not identify with the gender assigned to them at birth.

Things began to change in 2015. On July 28, 2015, then-Secretary of Defense Ash Carter issued a memorandum to the secretaries of the military departments directing that "[e]ffective as of July 13, 2015, no Service member shall be involuntarily separated or denied reenlistment or continuation of active or reserve service on the basis of their gender identity, without the personal approval of the Under Secretary of Defense for Personnel and Readiness." J.A. 709. The memorandum further ordered a working group composed of senior representatives from each of the Military Departments, Joint Staff, and relevant components from the Office of the Secretary of Defense to formulate policy options for the DoD regarding the military service of transgender servicemembers. The working group commissioned the RAND Corporation’s National Defense Research Institute to conduct a study on the impact of permitting transgender servicemembers to serve openly.

The RAND Corporation subsequently issued the aforementioned 91-page report ("RAND Report") that found no evidence that allowing transgender individuals to serve would have any effect on "unit cohesion," and concluded that any related costs or impacts on readiness would be "exceedingly small," "marginal," or "negligible." J.A. 597-708. Based on all of the information it collected, the working group unanimously concluded that transgender people should be allowed to serve openly in the military. The group not only concluded that allowing transgender people to serve would not significantly affect military readiness or costs but also found that prohibiting transgender people from serving would undermine military effectiveness and readiness. Specifically, prohibiting transgender people from serving would exclude qualified individuals on a basis that has no relevance to one’s fitness to serve and create unexpected vacancies requiring expensive and time-consuming recruitment and training of replacements.

On June 30, 2016, Secretary Carter issued Directive-Type Memorandum 16-005 ("Carter Policy"), which announced "that service in the United States military should be open to all who can meet the rigorous standards for military service and readiness," and set forth a policy permitting service by qualified transgender individuals. J.A. 586. The Carter Policy took immediate effect with respect to retention, allowing current transgender servicemembers to serve under "the same standards" as cisgender (non-transgender) servicemembers and prohibiting the discharge of otherwise qualified servicemembers "solely on the basis of their gender identity." J.A. 588. The Carter Policy also allowed transgender servicemembers diagnosed with gender dysphoria to transition to their preferred gender if they so desired. J.A. 275. Upon joining the military, each servicemember is assigned a "gender marker" in the Defense Enrollment Eligibility Reporting System (DEERS), and the servicemember must follow all sex-based military regulations pertinent to that gender marker. J.A. 1045. Thus, under the Carter Policy, transgender servicemembers with a gender dysphoria diagnosis became eligible to change their gender marker in DEERS, and with this transition, begin following the sex-based military regulations for their preferred gender rather than for their biological sex. The sex-based standards and procedures that vary by gender include (1) uniform, grooming, physical fitness, body fat, and drug testing standards; (2) requirements for separate berthing, bathroom and shower facilities; and (3) different policies regulating military training and sports, such as boxing. J.A. 296-98; J.A. 1045.

The Carter Policy directed DoD to update its standards for persons entering the military (a process formally referred to as "accession") by July 1, 2017. Secretary of Defense James Mattis subsequently deferred the July 1, 2017 accession deadline to January 1, 2018 so that the services could review their accession plans and provide input on the impact to the readiness and lethality of the armed forces. J.A. 426.

On July 26, 2017, President Donald J. Trump issued a statement via Twitter announcing that "the United States Government will not accept or allow transgender individuals to serve in any capacity in the U.S. Military." J.A. 124. A formal Presidential Memorandum ("2017 Presidential Memorandum") followed on August 25, 2017. J.A. 406-07. The 2017 Presidential Memorandum reversed the Carter Policy. President Trump explained that "[s]hortly before President Obama left office, ... his Administration dismantled the Departments’ established framework by permitting transgender individuals to serve openly in the military, authorizing the use of the Departments’ resources to fund sex-reassignment surgical procedures, and permitting accession of such individuals after July 1, 2017 [later extended to January 1, 2018]." J.A. 406. President Trump further declared:

In my judgment, the previous Administration failed to identify a sufficient basis to conclude that terminating the Departments’ longstanding policy and practice would not hinder military effectiveness and lethality, disrupt unit cohesion, or tax military resources, and there remain meaningful concerns that further study is needed to ensure that continued implementation of last year’s policy change would not have those negative effects.

J.A. 406. Thus, President Trump directed Secretary Mattis "to return to the longstanding policy and practice on military service by transgender individuals that was in place prior to June 2016 until such time as a sufficient basis exists upon which to conclude that terminating that policy and practice would not have the negative effects discussed above." J.A. 406. President Trump ordered Secretary Mattis to submit a plan for implementing the policy directives of the 2017 Presidential Memorandum by February 2018. J.A. 406-07.

By ordering a return to pre-Carter policies, President Trump effectively reinstated the prior blanket ban on accession and retention in military service by all transgender persons.

The President specified that the ban on transgender accession would go into effect immediately and remain in place "until such time as the Secretary of Defense, after consulting with the Secretary of Homeland Security, provides a recommendation to the contrary that I find convincing." J.A. 406. The President also ordered an immediate stop to the "use of DoD or DHS resources to fund sex-reassignment surgical procedures for military personnel, except to the extent necessary to protect the health of an individual who has already begun a course of treatment to reassign his or her sex." J.A. 406. Other than the halt to sex-reassignment funding, the President did not make the transgender ban immediately effective with respect to retention standards, which would have led to the immediate discharge of transgender servicemembers. Instead, President Trump instructed Secretary Mattis to "determine how to address transgender individuals currently serving in the United States military" and ordered that "no action may be taken against such individuals" until that study was complete. J.A. 407.

On October 30, 2017, the District Court issued a preliminary injunction enjoining the implementation of the 2017 Presidential Memorandum, the effect of which was to reinstate the Carter Policy. As the District Court found, and as we agreed in denying a motion to stay the preliminary injunction, a number of factors "strongly suggest[ed]" that Plaintiffs were likely to succeed on their Fifth Amendment due process claim. Doe 1 v. Trump , 2017 WL 6553389, at *1 (D.C. Cir. Dec. 22, 2017). Those factors included the breadth of the exclusion (virtually a complete ban) ordered by the Memorandum, the unusual and abrupt initial announcement of the ban, the failure to provide any supporting facts for the ban, and the recent professional judgment by the military, after a thorough study, that the prior ban should be lifted. Id . Given that intemperate contemporaneous statements by policymakers, departures from normal procedures, and adoption of policies unsupported or contrary to data can be considered evidence that invidious discrimination was "a motivating factor" in the decision, Village of Arlington Heights v. Metro. Hou. Dev. Corp. , 429 U.S. 252, 266-68, 97 S.Ct. 555, 50 L.Ed.2d 450 (1977), the course of events leading up to the 2017 Presidential Memorandum had more than a whiff of the stench of arbitrariness and of a "bare ... desire to harm a politically unpopular group." U.S. Dept. of Agric. v. Moreno , 413 U.S. 528, 534, 93 S.Ct. 2821, 37 L.Ed.2d 782 (1973).

II.

But that is not the end of the story. In February 2018, as ordered by the 2017 Presidential Memorandum, Secretary Mattis presented a memorandum to the President that proposed a policy regarding transgender military service ("the Mattis Plan"). J.A. 263-65. The reasoning underlying the Mattis Plan is spelled out in a 44-page report prepared by a panel of senior military and civilian experts that was also submitted to the President in February 2018 ("the Panel Report"). J.A. 268-312. The Panel Report concludes that transgender persons with gender dysphoria or who have undergone or will require gender transition undermine the military. According to the Panel Report, these servicemembers are fundamentally incompatible with the military’s mental health standards, physical health standards, and sex-based standards and are a detriment to military readiness and unit cohesion. The Panel Report likens gender dysphoria to conditions such as "bipolar disorder, personality disorder, obsessive-compulsive disorder, suicidal behavior, and even body dysmorphic disorder." J.A. 288. It concludes that individuals with gender dysphoria are more likely to have other mental health conditions and substance abuse problems, and to commit suicide. J.A. 289. The Panel Report also states that these individuals impose "disproportionate costs" on the military, J.A. 309, and repeatedly cites "uncertainty" in the medical field about these individuals as a reason to urge that the military "proceed with caution," J.A. 274.

The Mattis Plan has three key planks. First "[t]ransgender persons without a history or diagnosis of gender dysphoria, who are otherwise qualified for service, may serve, like all other Service members, in their biological sex." J.A. 265. "Biological sex" is not defined, but it appears to mean that the DEERS gender marker for the servicemember is the same as the sex assigned to the person at birth based on a physical examination of the genital organs. J.A. 299. "Serving" in one’s biological sex is also not defined, but it appears to mean that the transgender servicemember must follow all sex-based military regulations pertinent to the person’s DEERS gender marker. J.A. 273-74. (These critical points should be fleshed out in further proceedings.) The Mattis Plan’s panel of experts says that this particular recommendation is consistent with the Carter Policy, which also required transgender persons without gender dysphoria to serve in their biological sex, J.A. 300, and that there are transgender persons who "have served, and are serving, with distinction under the standards for their biological sex," J.A. 274.

Second, transgender persons who require or have undergone gender transition are disqualified from service, so they may not join, and may not be retained in, the military. J.A. 273. (However, the Mattis Plan includes a grandfather clause, which exempts current servicemembers who were diagnosed with gender dysphoria and began their transition pursuant to the Carter Policy and prior to the effective date of the Mattis Plan. J.A. 264; J.A. 273-74; J.A. 311.)

Third, transgender persons are disqualified from accession or retention in military service if they have "a history or diagnosis of gender dysphoria and require, or have already undertaken, a course of treatment to change their gender." J.A. 264; see also J.A. 300. (As above, the grandfather clause in the Mattis Plan exempts current servicemembers who were diagnosed with gender dysphoria and began their transition pursuant to the Carter Policy and prior to the effective date of the Mattis Plan. J.A. 264; J.A. 273-74; J.A. 311.) Thus, unless grandfathered, a transgender person with a history or current diagnosis of gender dysphoria can remain in the military only if the person is willing and able to serve in his or her biological sex. Similarly, a transgender person with a current diagnosis of gender dysphoria cannot join the military, and a transgender person with a history of gender dysphoria can join only if the person can "demonstrate 36 consecutive months of stability (i.e. absence of gender dysphoria )," has not transitioned, and is willing and able to serve in his or her biological sex. J.A. 273.

On March 23, 2018, the President issued a 2018 Presidential Memorandum, which stated that he "revoke[s]" his 2017 Presidential Memorandum, "and any other directive [he] may have made with respect to military service by transgender individuals." J.A. 261. The President ordered that "[t]he Secretary of Defense, and the Secretary of Homeland Security, with respect to the U.S. Coast Guard, may exercise their authority to implement any appropriate policies concerning military service by transgender individuals." J.A. 261. That same day, the government moved to dissolve the October 2017 injunction so that the military could implement the Mattis Plan. On August 6, 2018, the District Court found that the Mattis Plan had not "genuinely changed" the circumstances in the case, J.A. 94, and thus denied the motion to dissolve its injunction, J.A. 64-98.

III.

As stated earlier, consistency and clarity in terminology and definitions are important, and that was a key problem below.

The District Court ruled that by definition, transgender persons "do not identify or live in accord with their biological sex." J.A. 69; J.A. 91. Plaintiffs urged this definition upon the District Court, just as they urge it upon us, by arguing that all transgender persons undergo a gender transition, Appellee Br. 19. The problem is that there is no record support for this definition of the term. As noted earlier, everything in the record, from the Elders Report, the RAND Study, the Carter Policy and the Mattis Plan consistently defines a transgender person as a person who does not "identify" with his or her biological sex. Indeed, Plaintiffs’ own expert witness, J.A. 1056, and the American Medical Association, amici in support of Plaintiffs, AMA Amicus Br. 3, define transgender in terms of identifying with, rather than identifying with and living in accord with , one’s preferred gender.

This subtle shift in definition by Plaintiffs is significant, because it provides the lynchpin for their argument that the Mattis Plan was no different than the policy of the 2017 Presidential Memorandum. Plaintiffs, and the District Court, swept aside all distinctions in the various elements of the Mattis Plan by characterizing it as a monolith. Under this construct, the Mattis Plan is a total ban on service by transgender persons because a transgender person cannot "liv[e] in accord with their gender identity" while serving in the military according to their biological sex. Appellee Br. 21; see also J.A. 69; J.A. 91. In this paradigm, the Mattis Plan is a blanket transgender ban because it does not permit each and every transgender person to serve in his or her preferred gender; unless grandfathered, the Mattis Plan requires that transgender persons serve in accord with their biological sex.

Plaintiffs acknowledge that the term transgender is often defined to include persons who identify with another gender but who do not wish to live or work in accordance with that preferred gender, but they blithely dismiss that definition with the argument that any "broader meaning [of transgender] is irrelevant to this case." Appellee Br. 23. How so? Plaintiffs want us to ignore the fact that they did not present evidence showing that all transgender persons necessarily "live" or "work" in their preferred gender rather than their biological sex. There were no findings below describing what it means to "serve in one’s biological sex," let alone findings that all transgender persons either currently serving or applying to join the military must suppress who they are to "serve in their biological sex" and are unwilling or unable to do so. Instead, Plaintiffs want us to adopt the position that transgender persons who desire to serve in their biological sex are not really transgender, at least for the purpose of this lawsuit. But we cannot simply substitute Plaintiffs’ ipse dixit for evidence.

In sum, the record — at least at this point — shows that the Mattis Plan does not exclude all transgender persons serving in the military (unless grandfathered), as Plaintiffs maintain and as the District Court found. Rather, the record supports the conclusion that the Mattis Plan excludes (unless grandfathered) transgender persons who desire to express their gender identity by transitioning — changing their gender marker in DEERS — and thereby serving in accordance with all of the military requirements for their preferred gender. To the extent that there are transgender servicemembers who desire to serve under the standards for their biological sex, under the Mattis Plan, they may identify as they wish with their preferred gender while complying with the service requirements applicable to the DEERS gender marker for their biological sex.

None of this is to say that the Mattis Plan is not a hardship for transgender servicemembers who wish to transition. Nor is this dispositive of whether the Mattis Plan targets only transgender persons or is instead facially neutral. But it does mean that the Mattis Plan does not target all transgender persons, at least on this record, and it was therefore error to conclude that the Mattis Plan was not a substantive change from the 2017 Presidential Memorandum.

IV.

As we observed in the Judgment, the separation of powers principles upon which our Constitution is based requires that courts defer to the reasoned, professional analysis of Congress and the Executive on matters strictly within the realm of military expertise. This is so because the Constitution vests authority over the armed forces with Congress and the Executive, because military policies devised by the political branches are accountable to the electorate while those imposed by judges are not, and because, generally speaking, judges do not have competence in matters of military personnel and training. See generally Rostker v. Goldberg , 453 U.S. 57, 77-79, 101 S.Ct. 2646, 69 L.Ed.2d 478 (1981) ; Goldman v. Weinberger , 475 U.S. 503, 507-10, 106 S.Ct. 1310, 89 L.Ed.2d 478 (1986).

But that does not mean that Congress and the Executive have a wholesale license to discriminate in matters of military policy. Over forty years ago, the Supreme Court struck down a statute that required married female servicemembers to prove the financial dependency of their husbands in order to receive increased housing and healthcare benefits, while married male servicemembers automatically received the increased benefits, even if their wives were not financially dependent upon them. Frontiero v. Richardson , 411 U.S. 677, 678-79, 93 S.Ct. 1764, 36 L.Ed.2d 583 (1973). The statute, justified as promoting "administrative convenience," was a classification based on sex and therefore subject to heightened scrutiny, the same as any non-military, sex-based statute or policy. Id . at 688-89, 93 S.Ct. 1764. To satisfy heightened scrutiny, the Court demanded "concrete evidence" that the law promoted the claimed justification of administrative convenience, and finding none, held that the statute violated due process protections. Id . at 689-91, 93 S.Ct. 1764.

There was no deference in Frontiero to the expertise or judgment of the Congress that enacted the law or of the military officials who administered it. The facially discriminatory military decree in Frontiero was simply struck down.

Plaintiffs argue that transgender persons are a suspect class for due process and equal protection purposes, that the Mattis Plan facially discriminates against transgender persons, and citing Frontiero , that heightened scrutiny is appropriate. Appellee Br. 29-33. The government responds that transgender persons are not a suspect class, that the Mattis Plan does not target transgender persons, and that even if it did, it should not receive heightened scrutiny because policies strictly of a military nature can never be subject to heightened scrutiny. Appellant Br. 19-23; Reply Br. 9-15. The government, citing our opinion in Goldman , argues that the only reason heightened scrutiny was applied in Frontiero was because the statute involved personnel benefits and was therefore not "purported to be a congressional judgment on a uniquely military matter." Goldman v. Sec’y of Defense , 734 F.2d 1531, 1537 (D.C. Cir. 1984).

The government thus has a point, but we also observed in Goldman that not every infringement of a constitutional right can be judged equally, distinguishing between the "freedom to believe" in one’s religion, which is "absolute," and the "freedom to act" in accordance with one’s religion, which is "regulable for a permissible reason, provided that the regulation is not unduly restrictive." Id . at 1540-41. It would follow that a military policy prohibiting a servicemember from stating his adherence to Jewish beliefs would receive more scrutiny than a military policy restricting the wearing of a yarmulke in observance of that faith. Similarly, we have stated that even in the military context, "[c]lassifications based on race or religion, of course, would trigger strict scrutiny," Steffan v. Perry , 41 F.3d 677, 689 n.9 (D.C. Cir. 1994) (en banc), so we presumably would give more scrutiny to a military policy that permits the wearing of any religious headgear except yarmulkes than we would give to a neutral policy that restricts all religious headgear.

Since Frontiero , facially discriminatory military policies have been upheld only when there was a showing that the two classes of individuals were not truly "similarly situated" or when the discrimination was not against members of a suspect class. See Schlesinger v. Ballard , 419 U.S. 498, 508-09, 95 S.Ct. 572, 42 L.Ed.2d 610 (1975) (distinguishing Frontiero and upholding tenure statute that differentiated between male and female naval officers because it applied only in those circumstances where men and women were not similarly situated); Rostker , 453 U.S. at 77-79, 101 S.Ct. 2646 (upholding sex-based draft-registration statute aimed at developing the pool of potential combat troops, where there was no challenge to the then-existing exclusion of women from combat, and thus the policy was "not invidious, but rather realistically reflect[ed] the fact that the sexes [were] not similarly situated"); Steffan , 41 F.3d at 684 n.3 (no heightened scrutiny of Naval Academy policy prohibiting attendance of homosexuals because homosexuals were not a suspect class under the unchallenged, then-existing military regulations criminalizing homosexual conduct — reasoning that is untenable after Lawrence v. Texas , 539 U.S. 558, 578-79, 123 S.Ct. 2472, 156 L.Ed.2d 508 (2003), which struck down a state statute criminalizing consensual homosexual sodomy). But cf . Trump v. Hawaii , ––– U.S. ––––, 138 S.Ct. 2392, 2419, 201 L.Ed.2d 775 (2018) (noting that in Fiallo v. Bell , 430 U.S. 787, 97 S.Ct. 1473, 52 L.Ed.2d 50 (1977), the Court upheld an immigration statute that included a " ‘categorical’ entry classification that discriminated on the basis of sex and legitimacy").

Furthermore, most constitutional challenges to military laws and regulations have involved policies that were facially neutral. See , e.g. , Greer v. Spock , 424 U.S. 828, 839-40, 96 S.Ct. 1211, 47 L.Ed.2d 505 (1976) (regulations restricting partisan political speeches, demonstrations, and literature distribution were "politically neutral," and there was no claim that the military had acted "irrationally, invidiously, or arbitrarily"); Goldman , 475 U.S. at 513, 106 S.Ct. 1310 (Stevens, J., concurring) (uniform regulation was "based on a neutral, completely objective standard — visibility"); In re Navy Chaplaincy , 738 F.3d 425, 430 (D.C. Cir. 2013) (personnel policies for promoting chaplains were "facially neutral" and there was "no showing of intent to discriminate" based upon religion). At a minimum, the observation that the military policy was facially neutral in each of those cases demonstrates that the presence or absence of neutrality must have some relevance to the analysis. Cf . Hawaii , 138 S.Ct. at 2417-18 (in resolving constitutional challenge to a Presidential Proclamation in the analogous context of national security, the Court’s first step was to determine that the Proclamation was facially neutral). Even when dealing with facially neutral policies, Congress and the Executive receive deference only where military policies are based upon the "considered professional judgment" of "appropriate military officials" and only after finding that the policies "reasonably and evenhandedly regulate" the matter at issue, Goldman , 475 U.S. at 509-10, 106 S.Ct. 1310 ; see also Rostker , 453 U.S. at 68, 101 S.Ct. 2646 (court will not substitute its "own evaluation of evidence for a reasonable evaluation by the Legislative Branch").

The point here is that determining the correct standard of review for military policies, and then properly applying it, is a complex venture. In Rostker , the Court declined the Government’s invitation to declare that rational basis scrutiny, rather than heightened scrutiny, necessarily applies to all military policies, even when the policy includes a facially discriminatory classification:

We do not think that the substantive guarantee of due process or certainty in the law will be advanced by any further "refinement" in the applicable tests as suggested by the Government. Announced degrees of "deference" to legislative judgments, just as levels of "scrutiny" which this Court announces that it applies to particular classifications made by a legislative body, may all too readily become facile abstractions used to justify a result. In this case the courts are called upon to decide whether Congress, acting under an explicit constitutional grant of authority, has by that action transgressed an explicit guarantee of individual rights which limits the authority so conferred. Simply labeling the legislative decision "military" on the one hand or "gender-based" on the other does not automatically guide a court to the correct constitutional result.

Rostker , 453 U.S. at 69-70, 101 S.Ct. 2646. The fact that a military policy is involved certainly counsels greater deference to Congress and the Executive, but Rostker , Goldman , and the other precedent cited above teach that the standard of review cannot be easily quantified using a specific degree of deference or level of scrutiny. Rather, our review involves the careful assessment of a number of factors, including whether the policy is facially neutral, whether it targets a suspect class, whether the class is similarly situated to others affected, whether the policy was motivated by animus, whether it infringes upon a fundamental right (and, if so, how), what military purposes are furthered by the policy, whether those purposes are legitimate, and whether Congress or the Executive used considered professional judgment and accommodated the servicemembers’ rights in a reasonable and evenhanded manner, given the rights at issue.

At this juncture, I express no views on these various factors, including whether the Mattis Plan was a product of comparable "considered professional judgment" as the policy in Goldman, other than to concur with the Judgment that the Plan and its accompanying Panel Report constituted changed circumstances that warrant vacating the preliminary injunction and assessing the Mattis Plan anew. This is especially true since the preliminary injunction relied heavily on the curious circumstances surrounding the President’s tweet and subsequent Memorandum, which reversed a policy supported by lengthy and careful study and replaced it with a policy lacking the apparent support of any contradictory study. Reassessment is also necessary to clarify whether requiring service in one’s biological sex impacts all servicemembers who identify as transgender in the same manner, because that impacts whether the Mattis Plan should be construed as one monolithic policy rather than analyzed separately into its constituent elements. Once those matters are understood, the District Court can properly review the policy or policies challenged by Plaintiffs. I express no views on the merits or the outcome of that reassessment. I write only to provide some guidance to the parties and the District Court so that this matter can proceed in an efficient manner. At this stage, and on this record, I believe it is premature to opine, let alone rule, on anything further.

My concurring colleague would go much further and not only reach the merits of the constitutional issues but also take the drastic step of precluding any discovery and dismissing the lawsuit. Op. at 706–07, 715–16, 735–37 (Williams, J., concurring). For several reasons, I disagree. To begin with, having found adequate non-constitutional reasons to vacate the preliminary injunction, it is inappropriate to seek out constitutional questions to decide in order to reach the same result. The concurrence strains to find a case on point, but all are inapposite; none holds that we are required to reach the merits of the claims in order to dissolve a preliminary injunction. Further, even if we have the discretion to reach the merits, that doesn’t mean we should, particularly when they are constitutional claims. "A fundamental and longstanding principle of judicial restraint requires that courts avoid reaching constitutional questions in advance of the necessity of deciding them." Lyng v. Northwest Indian Cemetery Protective Ass’n , 485 U.S. 439, 445, 108 S.Ct. 1319, 99 L.Ed.2d 534 (1988) ; see also Stillman v. C.I.A. , 319 F.3d 546, 548 (D.C. Cir. 2003) ; United States v. Rostenkowski , 59 F.3d 1291, 1302-03 (D.C. Cir. 1995).

I grant that the Court reached out to decide the constitutional issues in Munaf v. Geren , 553 U.S. 674, 128 S.Ct. 2207, 171 L.Ed.2d 1 (2008), but that was not only an extraordinary circumstance but also one where "[a]djudication of the merits ... rest[ed] on a question of law ..." id. at 691, 128 S.Ct. 2207. Here, the constitutional claims are not purely legal in nature, and the District Court has found that some factual issues are in dispute. The concurrence frets about "intrusions into executive decision making" and of the "President’s mental processes," Op. at 737 (Williams, J., concurring), but that does not justify shutting off all discovery. For instance, as explained in our Judgment and above, there is considerable confusion in this case about how the Mattis Plan impacts transgender servicemembers who must comply with the service requirements for their biological sex. Compelling military or executive officials to explain the operation and purpose of this requirement would not improperly intrude upon such mental processes. "It is the decision-making process that requires shielding from public scrutiny, not the decision itself once it has been acted on." 3 WEINSTEIN’S FEDERAL EVIDENCE § 509.23 (2019); see also In re Sealed Case , 121 F.3d 729, 737 (D.C. Cir. 1997) ("The deliberative process privilege does not shield documents that simply state or explain a decision the government has already made ...."); EDWARD J. IMWINKELREID, THE NEW WIGMORE : A TREATISE ON EVIDENCE § 7.7.2 (3d ed. 2019) ("Nor is there a privilege for a post-decision explanation of the decision."); 26A CHARLES ALAN WRIGHT ET AL. , FEDERAL PRACTICE AND PROCEDURE : EVIDENCE § 5680 (2018) ("The [deliberative process] privilege does not apply ... to explanations of existing regulations or of a past decision."). Now that the Mattis Plan has been adopted, there is nothing unduly intrusive about asking officials to testify about how the policy operates and what military purposes it serves, and we have required similar discovery in the past where military policies were at issue. See Waldie v. Schlesinger , 509 F.2d 508, 510 & n.1 (D.C. Cir. 1974) (denying summary judgment where affidavits of government witnesses were "ambiguous" and "conclusory" in explaining military policies). Indeed, this very type of discovery occurred in Goldman and Rostker without any apparent damage to the republic. See, e.g. , Brief for Respondents, Goldman v. Weinberger, 1985 WL 669077, at *4-5 ("At trial, Air Force witnesses explained why the extraordinarily detailed regulation of the official uniform worn by its personnel is deemed essential to the accomplishment of the Air Force’s mission."); Brief for Appellant, Rostker v. Goldberg, 1981 WL 390367, at *30 ("[H]igh level military personnel who testified in this case were of the view ... that developing methods of warfare employed by the United States and its allies would require more combat intensive forces than those that had been required during past military conflicts." (citing deposition testimony)).

My concurring colleague points out that Rostker and Goldman reversed the trial court rulings on the merits, Op. at 737 (Williams, J., concurring), but in doing so, the Court did not criticize discovery about how military policies operated or what interests they served. Rather, the Court noted that it was improper for lower courts to consider plaintiff expert testimony that contradicted the military experts about whether the policies at issue were justified under the circumstances. Rostker , 453 U.S. at 80-81, 101 S.Ct. 2646 ; Goldman , 475 U.S. at 509-10, 106 S.Ct. 1310. The concurrence cites no case that has adopted the astounding rule that the government is immune from all discovery explaining the operation and purpose of its military policies. Even in a facial challenge, discovery may be necessary where the impact of the regulation is unclear and disputed. Turner Broad. Sys., Inc. v. F.C.C. , 512 U.S. 622, 667-68, 114 S.Ct. 2445, 129 L.Ed.2d 497 (1994) (reversing summary judgment for the government in facial challenge because of factual disputes about the "actual effects" of the regulation). It is one thing to defer to the government’s justification for military policy; it is quite another not to require the government to explain fully, under oath, that justification. The former custom of deference is reasonable, but the latter is imprudent. Cf. United States v. Nixon , 418 U.S. 683, 710 n.18, 94 S.Ct. 3090, 41 L.Ed.2d 1039 (1974) ("Because of the key role of the testimony of witnesses in the judicial process, courts have historically been cautious about privileges."). The concurrence cites no reason why the current commander-in-chief is entitled to more deference or authorized to provide less explanation than any of his predecessors, and I am certainly not aware of any. It is for this reason that the panel judgment opted for regular order.

Williams, Senior Circuit Judge, concurring in the result:

The district court issued a nationwide injunction strictly limiting the authority of the current President and secretary of defense to alter or amend their predecessors’ military directives and policy judgments about the composition of the armed forces. In our January 4 judgment, we reversed the court’s refusal to dissolve the injunction; we also held that it abused its discretion to the extent it granted plaintiffs additional relief. See Doe 2 v. Shanahan , No. 18-5257, 2019 WL 102309, at *1 & n.1 (D.C. Cir. Jan. 4, 2019) (" Panel Judgment "), rev’g Doe 2 v. Trump , 315 F.Supp.3d 474 (D.D.C. 2018), cert. before judgment denied , No. 18-677, ––– U.S. ––––, 139 S.Ct. 946, ––– L.Ed.2d ––––, 2019 WL 272026 (U.S. Jan. 22, 2019). I joined the panel in dissolving the injunction—and explained that a separate opinion would follow. This is that opinion.

I write separately because I believe the record and the law require dismissal of plaintiffs’ claims. A correct resolution at this stage is important because the decisions of the district court reflect what in my view are wholly mistaken assumptions about the nature of constitutional review of military personnel policy—at least on the facts of this case. To be specific, the court contemplates a highly intrusive examination of the mental processes of the civilian and military leadership of a coordinate branch of government. See, e.g., Doe 2 v. Mattis , 322 F.Supp.3d 92, 101 (D.D.C. 2018) (declaring that plaintiffs "are entitled to complete discovery" regarding the executive’s "alleged deliberations"); Doe 2 v. Trump , 319 F.Supp.3d 539, 543 (D.D.C. 2018) (emphasizing that plaintiffs will "be able" to "seek discovery from" the President). These interrogations are to be aimed, the district court has said, at uncovering infirmities in a policy that the secretary of defense, in his "professional judgment," has determined is essential to putting our armed forces "in the strongest position to protect the American people." Memorandum from Secretary Mattis to President Trump 3 (Feb. 22, 2018) ("Mattis Memo "), J.A. 265. Under prevailing constitutional doctrine, however, this case can (and should) be resolved on the existing record.

I should start by clarifying what this case is actually about. It is not about a "transgender ban." See, e.g., Appellees’ Br. 2. The challenged policy expressly provides that "[t]ransgender persons ... may serve, like all other Service members ." Mattis Memo 3, J.A. 265 (emphasis added). It is instead about whether the Constitution requires the current administration to reinstate a policy created by the previous administration allowing certain transgender individuals to serve in their preferred gender rather than their biological sex, as all service members have for decades. Id . For those transgender persons for whom the value of serving otherwise than in their biological sex exceeds the value of being in the military, of course, the policy thwarts their wish to serve. The Constitution does not compel the military to yield to their preference.

Once we abstract away from the politically charged subject-matter—as we must—this is a straightforward legal case. Our "Constitution vests [t]he complex, subtle, and professional decisions as to the composition, training, equipping, and control of a military force exclusively in the legislative and executive branches." Panel Judgment *2 (alteration in original) (quoting Kreis v. Sec’y of Air Force , 866 F.2d 1508, 1511 (D.C. Cir. 1989) ) (internal quotation marks omitted); see U.S. Const. art. I, § 8, cls. 12 – 16 ; art. II, § 2, cl. 1. So, although the policy argument in favor of accommodating plaintiffs’ claimed entitlement to transition into their preferred gender may be compelling, the legal argument for requiring such accommodation is not.

The current administration’s policy—which is far more accommodating to transgender service than that of any administration from 1789 until 2016—easily passes constitutional muster. Given their biological differences, males and females have long been assigned to separate berthing, bathroom, and shower facilities, and subject to different sets of physical fitness, body fat, uniform, and grooming standards. This is unquestionably lawful; it is required (in part) by Congress, see, e.g., 10 U.S.C. § 7419, and has been approved (repeatedly) by the Supreme Court, see, e.g., United States v. Virginia , 518 U.S. 515, 550 n.19, 116 S.Ct. 2264, 135 L.Ed.2d 735 (1996) ; Rostker v. Goldberg , 453 U.S. 57, 81, 101 S.Ct. 2646, 69 L.Ed.2d 478 (1981). To put it simply, there is no constitutional right for, say, biological males who identify as female to live, sleep, shower, and train with biological females. Whether allowing such flexibility in military service is a good idea or not is of no concern to the courts; that is a question for the people acting through their elected representatives.

I

With the general issue in mind, I discuss the relevant terminology and the successive policies that lie at the root of the case, together with a timeline of the litigation; I then turn to the legal issues.

A

Transgender individuals, according to plaintiffs’ expert witness, have a "gender identity"—an "internalized, felt sense of who they are as male or female"—that does not align with "their assigned sex at birth." Decl. of George Richard Brown in Support of Plaintiffs ¶¶ 13–14 (Aug. 30, 2017) ("Brown 2017 Decl."), J.A. 1056; accord, e.g., Amici Curiae Am. Medical Ass’n et al. Br. ("AMA Br.") 3. In this way, "[t]ransgender people differ from non-transgender individuals, whose gender identity aligns with the sex assigned at birth." Id . at 4. In virtually all cases, the latter concept, "sex assigned at birth," lines up with an individual’s "biological sex ," as determined by "chromosomes, gonads, hormones, and genitals," Department of Defense Report and Recommendations on Military Service by Transgender Persons 7 n.10 (Feb. 2018) ("DoD Report "), J.A. 275 (emphasis added); see, e.g., Agnes Gereben Schaefer et al., Assessing the Implications of Allowing Transgender Personnel To Serve Openly 5 (2016) ("RAND Report "), J.A. 621 (explaining that "birth sex ... typically correlates with primary sex characteristics (e.g., genitalia)"). To be sure, some transgender individuals undergo sex reassignment surgery. But the rates for complete sex reassignment surgery "are exceedingly low—2% of transgender men and 10% of transgender women." DoD Report 31, J.A. 299.

Gender dysphoria , again according to plaintiffs’ experts, is a mental health condition from which only "a subset of transgender people" suffer. Decl. of George Richard Brown in Support of Opp’n to Defs.’ Mot. to Dismiss ¶ 9 (May 11, 2018), J.A. 839; accord, e.g., Decl. of Brad R. Carson in Support of Pls.’ Mot. for Preliminary Injunction ¶ 23 (Aug. 28, 2017), J.A. 995. It is a serious mental health condition that is recognized by the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders (or "DSM "). Brown 2017 Decl. ¶¶ 16–18, J.A. 1057; accord, e.g., DoD Report 13, J.A. 281; AMA Br. 7. The condition is "associated with clinically significant distress or impairment in social, occupational, or other important areas of functioning." DoD Report 13, J.A. 281 (quoting Am. Psychiatric Ass’n, Diagnostic and Statistical Manual of Mental Disorders 453 (5th ed. 2013)); accord, e.g., Brown 2017 Decl. ¶ 17, J.A. 1057; AMA Br. 7. "[U]ntreated," it "can cause debilitating distress, depression, impairment of function, self-mutilation to alter one’s genitals or secondary sex characteristics, other self-injurious behaviors, and suicide." Id . at 9.

"The recommended treatment for gender dysphoria includes assessment, counseling, and, as appropriate,"gender transition . AMA Br. at 11. Gender transition includes "social transition, hormone therapy, and surgical interventions to bring the body into alignment with one’s gender identity." Id .; accord, e.g., Brown 2017 Decl. ¶ 23, J.A. 1059. Social transition, the sole choice of many, consists simply of "living one’s life fully in accordance with one’s gender identity." AMA Br. 11. "This typically includes publicly identifying oneself as that gender through all of the ways that people signal their gender to others such as through their name, pronoun usage, dress, manner and appearance, and social interactions." Id .

Transition became relevant to military personnel administration for the first time in 2016, under the orders of then-Secretary of Defense Ashton Carter. The so-called Carter policy, which we’ll discuss in detail below, defined a transition as being "complete" when a service member had "completed the medical care identified or approved by a military medical provider in a documented medical treatment plan as necessary to treat" gender dysphoria. U.S. Dep’t of Defense, Transgender Service in the U.S. Military: An Implementation Handbook 11 (Sept. 30, 2016) ("Carter Memo Implementation Handbook "), J.A. 518; see also Memorandum from Secretary Carter to Secretaries of Military Departments, attachment at 1 (June 30, 2017) ("Carter Memo "), J.A. 588 (explaining that a "history of medical treatment associated with gender transition is disqualifying" unless a service member "has completed all medical treatment associated" with the transition).

B

We deal here with three successive policies on military service by transgender persons: (1) Pre-2016; (2) the Carter policy (so named after Secretary Carter); and (3) the Mattis policy (so named after James Mattis, Carter’s immediate successor as secretary of defense).

1. Pre-2016.

(A) Being transgender. The parties appear to agree that transgender persons with or without gender dysphoria (or the equivalent of this condition in earlier terminology) were barred from military service. See, e.g., Appellants’ Br. 12; Appellees’ Br. 3. I return to this below.

(B) Gender dysphoria was clearly a bar to military service (at least as clearly as one can say, given changes in nomenclature). Given the "unique mental and emotional stresses of military service," DoD Report 10, J.A. 278, the executive branch long presumptively disqualified from service individuals with "[m]ost mental health conditions," id . at 20, J.A. 288. The military leadership has traditionally aligned these disqualifying conditions with the conditions listed in the DSM . Id . at 34, J.A. 302.

For decades, then, the military’s "accession" standards—the standards governing induction into the armed forces—disqualified individuals with "psychosexual conditions," including "transsexualism." M. Joycelyn Elders et al., Medical Aspects of Transgender Military Service 3 (2014), J.A. 784. ("Transsexualism" first appeared in the DSM in 1980. DoD Report 10, J.A. 278.) Likewise, the military’s "retention" standards—the standards governing separation from the armed forces—traditionally permitted discharge of individuals with "sexual gender and identity disorders." Elders et al., supra , at 3, J.A. 784.

The parties seem generally to regard the historic accession and retention practices as effectively barring all "[t]ransgender individuals" from "enlist[ing] or serv[ing] in the US armed forces." Id . Even transgender individuals who did not "take hormones, have surgery, or undergo any other aspect of gender transition" were, under these "psychological" standards, ineligible to serve. Id . (emphasis added); see also Doe 1 v. Trump (Doe I ), 275 F.Supp.3d 167, 178 (D.D.C. 2017) (explaining that "transgender individuals" could not "obtain medical waivers for entrance into the military").

I frankly have some doubt whether this characterization of the pre-2016 world is fully accurate. Because many transgender persons live in their "biological sex," without apparent gender dysphoria, it seems improbable that the policies in fact excluded all transgender persons. As nothing before the court appears to turn on this, however, I will assume the correctness of the parties’ shared assumptions.

2. Carter Policy.

As the administration of President Barack Obama wound down, the military embarked on a reconsideration of its longstanding policy on "transsexualism." By that time, the DSM no longer included that term; it had been replaced with "gender dysphoria," detailed above. For this and other reasons, then-Secretary of Defense Ashton Carter adopted a new policy to govern the "accession" and "retention" of "transgender personnel." Carter Memo 1, J.A. 585.

(A) Transgender individuals not suffering from gender dysphoria or undergoing gender transition were eligible for service—only in their biological sex.

(B) For transgender individuals with gender dysphoria or who embarked on transition to their preferred gender, the story was more complicated.

(1) A "history of gender dysphoria" was "disqualifying" for accession purposes unless an individual had "been stable without clinically significant distress or impairment in social, occupational, or other important areas of functioning for 18 months." Carter Memo , attachment at 1, J.A. 588. Individuals diagnosed with gender dysphoria while serving could continue to serve provided that they met deployability standards, see id ., and were able to serve in their biological sex (unless and until they completed a gender transition, as discussed below), see Carter Memo Implementation Handbook 11–12, J.A. 518–19; supra p. 710 n.1.

(2) A history of treatment for "gender transition" was likewise "disqualifying" for accession purposes unless an individual had "completed all medical treatment associated with the ... gender transition" and had "been stable in the preferred gender for 18 months." Carter Memo , attachment at 1, J.A. 588.

Transgender persons who had completed "gender transition" as a treatment for "gender dysphoria" could access and serve in their "preferred gender" rather than their biological sex. Id .; see also Appellees’ Br. 19, 41, 42 (referring to transition as "treatment"). Because "gender transition" under the Carter policy included (as in general medical parlance) social transition, i.e., simply "living in the preferred gender," Carter Memo Implementation Handbook 31, J.A. 538, an individual could (for example) remain biologically male in every respect (at least from an external point of view) and, nonetheless, demand to be treated as in all respects female, including for "medical fitness, physical fitness, uniform and grooming, deployability and retention standards," as well as for assignment to "berthing, bathroom, and shower facilities," id . at 43–44, J.A. 550–51.

For retention purposes, individuals who embarked on gender transition while serving could continue to serve provided that they met the standards associated with their biological sex until their transition was "complete" (at which point, they could serve in their preferred gender). See Carter Memo Implementation Handbook 11–12, J.A. 518–19; supra p. 710 n.1.

3. Mattis Policy.

(A) Like the Carter policy before it, the Mattis policy allowed transgender persons without a history or diagnosis of gender dysphoria to serve in their biological sex. Mattis Memo 3, J.A. 265.

(B) Also like the Carter policy, the Mattis policy applied more complex rules for those with a history or diagnosis of gender dysphoria or involved in gender transition.

(1) The policy presumptively disqualified for accession purposes individuals with a "history" of "gender dysphoria" unless they were stable, without clinically significant symptoms, for some period of time—36 months under Mattis, id . 2, J.A. 264 (as compared with 18 months under Carter). For retention purposes, individuals "diagnosed with gender dysphoria after entering into service may be retained if they do not require a change of gender and remain deployable within applicable retention standards." Id .

(2) The Mattis policy diverged strongly from the Carter policy in its handling of gender transition. It reinstated the prior military practice of requiring that "all" individuals serve in their "biological sex"; those who had undergone "gender transition" were thus presumptively disqualified. Mattis Memo 2–3, J.A. 264–65. In recognition that some individuals had acted in reliance on the Carter policy, however, Secretary Mattis created a reliance exception: individuals who were "diagnosed with gender dysphoria" under the Carter policy could "continue to serve in their preferred gender" (rather than their biological sex). Id . at 2, J.A. 264.

The following chart summarizes the three policies for accession and retention of transgender personnel:?

C

Plaintiffs launch a comprehensive broadside against the Mattis policy and demand a judicial mandate ordering the executive branch to "revert to" the Carter policy. See Second Am. Compl. 20, J.A. 209. Plaintiffs claim support in the supposedly "unusual" circumstances surrounding the announcement of the new policy. In doing so, they resolutely focus on a presidential "tweet" from July 2017, an event that, they say, forever tainted the Mattis policy. But they pay little or no attention to Secretary Mattis’s prior order of June 30, 2017, deferring the start of accessions under the Carter policy. And while the sequence may have little real consequence under a proper legal analysis, plaintiffs’ persistent highlighting of the tweet suggests the need for a precise timeline:

July 28, 2015: Secretary Carter convenes a working group to study the "policy options" for allowing service by "transgender Service members." He instructs the group to "start with [a] presumption"—namely, "that transgender persons can serve openly without adverse impact." Memorandum from Secretary Carter to Secretaries of Military Departments (July 28, 2015), J.A. 709.

2016: The Carter working group enlists the RAND National Defense Research Institute to study the impacts of allowing "transgender service members" to serve openly. RAND finds (at, e.g., J.A. 646–48, 655–59) that the proposed policy change would have an adverse impact on health care utilization costs and readiness, but concludes (at, e.g., J.A. 685, 686) that the impact would be "negligible" or "marginal" because of the "small" number of transgender service members relative to the size of the military as a whole. See RAND Report , J.A. 597–708. RAND does not explain why overall military costs are a relevant comparator, as opposed, for example, to the benefits to be derived (by the military and the individuals in question) from increased transgender service. RAND further concludes, based on "limited publicly available data," that open transgender service would not have a "significant effect" on unit cohesion. Id . at 44, J.A. 660. But RAND does "not have direct survey evidence or other data to directly assess the impact on the U.S. military," id ., and identifies "reports of resistance" in the ranks to a similar policy change in the United Kingdom, id . at 45, J.A. 661.

June 30, 2016 : Following this review, Secretary Carter adopts a new policy on "Military Service of Transgender Service Members." The Carter policy "[e]stablishes policy ... for the standards for retention, accession, separation, in-service transition, and medical coverage for transgender personnel serving in the Military Services." See Carter Memo 1, J.A. 585.

September 30, 2016: The military issues a 71-page handbook to address "some of the issues" related to "transgender Americans serving openly in the military." See Carter Memo Implementation Handbook 8, 10, J.A. 515, 517.

June 30, 2017: On the day before the Carter policy’s accession directives were scheduled to go into effect, Secretary Mattis "defer[s] the start of accessions [under the Carter policy] for six months." He explains that his "intent is to ensure that [he] personally ha[s] the benefit of the views of the military leadership and of the senior civilian officials who are now arriving in the Department." The military will "use this additional time to evaluate more carefully the impact of such accessions on readiness and lethality." Memorandum from Secretary Mattis to Secretaries of Military Departments (June 30, 2017), J.A. 425.

July 26, 2017: The President, who appears not to have addressed the issue until now, expresses agreement with Secretary Mattis. He tweets:

After consultation with my Generals and military experts, please be advised that the United States Government will not accept or allow Transgender individuals to serve in any capacity in the U.S. Military. Our military must be focused on decisive and overwhelming victory and cannot be burdened with the tremendous medical costs and disruption that transgender in the military would entail. Thank you[.]

July 27, 2017: The Chairman of the Joint Chiefs of Staff announces that "no modifications" to policy will be made in response to the tweet. See Memorandum from Chairman Dunford to Chiefs of Military Services (July 27, 2017), J.A. 408.

August 9, 2017: Without mentioning Secretary Mattis’s June 30, 2017 deferral order or Chairman Dunford’s above announcement, plaintiffs file suit, claiming that the President’s tweet "reverse[d] the current policy"—referring to the Carter policy, even though Secretary Mattis had already put that on hold. Compl. ¶ 1 (Aug. 9, 2017), ECF No. 1.

August 25, 2017: The President issues a formal memorandum announcing that he, like Secretary Mattis, wants "further study" before any implementation of Secretary Carter’s "policy change." In the President’s "judgment, the previous Administration failed to identify a sufficient basis to conclude that terminating the Departments’ longstanding policy and practice would not hinder military effectiveness and lethality, disrupt unit cohesion, or tax military resources." Memorandum of August 25, 2017, 82 Fed. Reg. 41,319, 41,319 (Aug. 30, 2017) ("2017 Presidential Memorandum "). Accordingly, exercising the power entrusted to him as "Commander in Chief of the Army and Navy," U.S. Const. art. II, § 2, cl. 1, the President directs the Secretary of Defense "to return to the longstanding policy and practice on military service by transgender individuals that was in place prior to [former-Secretary Carter’s change]"—at least "until such time as a sufficient basis exists upon which to conclude that terminating that policy and practice would not have [ ] negative effects," 2017 Presidential Memorandum , 82 Fed. Reg. at 41,319. But the President provides that "no action may be taken" against "currently serving" transgender individuals until the Secretary "determine[s] how to address" the issue. Id . at 41,320. The President also makes clear that Secretary Mattis could "advise [him] at any time, in writing, that a change to this [longstanding] policy is warranted." Id . at 41,319.

September 14, 2017: Secretary Mattis convenes a panel of experts. He orders them to implement "a comprehensive, holistic, and objective approach to study military service by transgender individuals, focusing on military readiness, lethality, and unit cohesion, with due regard for budgetary constraints and consistent with applicable law." Memorandum from Secretary Mattis to Secretaries of Military Departments 2 (Sept. 14, 2017), J.A. 404.

October 30, 2017: At plaintiffs’ request, the district court preliminarily enjoins all branches of the military from enforcing two provisions of the 2017 Presidential Memorandum. See Doe I , 275 F.Supp.3d 167.

Specifically, the court "enjoins Defendants from enforcing the following directives of the [2017] Presidential Memorandum": (1) " ‘return to the longstanding policy and practice on military service by transgender individuals that was in place prior to [the Carter policy]’ "; and (2) " ‘maintain the [longstanding, pre-Carter] policy regarding accession of transgender individuals into military service.’ " Order 1–2 (Oct. 30, 2017), ECF No. 60, J.A. 188–89 (quoting 2017 Presidential Memorandum , 82 Fed. Reg. at 41,319).

The court characterizes the President’s directives as discriminating on the basis of "transgender identity," and as therefore subject to "intermediate scrutiny." 275 F.Supp.3d at 209. Applying such scrutiny, the court holds that the President’s directives likely fail because of several supposedly "unusual" factors associated with the President’s announcement—namely, that it was "contradicted" by the conclusions of a previous administration, lacked supporting "studies," and was issued "via Twitter," without "formality." Id . at 212–13 (emphasis removed).

November 21, 2017: The government files an interlocutory appeal of the preliminary injunction. See Defs.’ Notice of Appeal (Nov. 21, 2017), ECF No. 66.

November 27, 2017: The district court issues a "clarification." Although the October 2017 order enjoined implementation of only two specific provisions of the Presidential Memorandum, the court "clarif[ies]" that "[a]ny action by any of the Defendants that changes [the] status quo [i.e., the Carter policy] is preliminarily enjoined." Order 2 (Nov. 27, 2017) (emphasis in original), ECF No. 70, J.A. 110. The court, however, continues to view the original October 2017 injunction as the operative injunction. See, e.g., Doe 2 v. Trump (Doe II ), 315 F.Supp.3d 474, 498 (D.D.C. 2018) ("To avoid any possible need for clarification, the Court states expressly: enforcing the Mattis Implementation Plan would violate the Court’s October 30, 2017 preliminary injunction.").

January 4, 2018: In anticipation of the announcement of the Mattis policy (see February 22), the government voluntarily dismisses its appeal. See Doe 1 v. Trump , No. 17-52-67, 2018 WL 411236 (D.C. Cir. Jan. 4, 2018).

February 2018: The Department of Defense releases a 44-page report based on the recommendations of the panel of experts convened by Secretary Mattis. "The Panel made recommendations based on each Panel member’s independent military judgment." DoD Report 4, J.A. 272.

February 22, 2018: In "light of the Panel’s professional military judgment" and his own "professional judgment," Secretary Mattis recommends a new policy to the President. Noting that the President had "made clear" that the Secretary could "at any time, in writing," advise that a change in policy was "warranted," Secretary Mattis recommends that the President "revoke" his 2017 Presidential Memorandum, "thus allowing" the military to adopt the new policy. Mattis Memo 1, 3, J.A. 263, 265.

March 23, 2018: The President (again) agrees with Secretary Mattis, revoking all prior directives so that Secretary Mattis could "implement any appropriate policies." See Memorandum of March 23, 2018, 83 Fed. Reg. 13,367, 13,367 (Mar. 28, 2018) (revoking the 2017 Presidential Memorandum "and any other directive [the President] may have made with respect to military service by transgender individuals").

On the same day, the government, in accord with its prior dismissal of its appeal, seeks from the district court protection from a possible contempt citation. Though framing its motion as a request to the court to dissolve the injunction, the government "maintain[s] that the Court’s preliminary injunction, which addressed only certain provisions of the President’s 2017 Memorandum [see account of November 27, 2017 injunction], does not extend to the Department’s new policy." "But in an abundance of caution," the government urges the court "to dissolve the preliminary injunction in order to permit the military to implement the policy it believes will best ensure our Nation’s defense." Defs.’ Mot. to Dissolve the Preliminary Injunction 2 (Mar. 23, 2018), ECF No. 96.

April 6, 2018: Following an inquiry by the district court as to whether plaintiffs will be "amending the complaint" in light of the announcement of the new Mattis policy, see Tr. 5:16–17 (Mar. 28, 2018), ECF No. 102, plaintiffs amend their complaint, see Second Am. Compl. (Apr. 6, 2018), ECF No. 106, J.A. 190.

April 20, 2018: The government again asks the district court, "in an abundance of caution," to dissolve the preliminary injunction (again asserting that the original injunction did not apply to Secretary Mattis’s actual policy choice). Defs.’ Mot. to Dissolve the Preliminary Injunction 2 (Apr. 20, 2018), ECF No. 116.

August 6, 2018: The district court refuses to dissolve the injunction, holding that the October 2017 injunction still applies because the Mattis policy "merely implements the basic policy directives in the President’s 2017 tweet and memorandum." Doe II , 315 F.Supp.3d at 488 n.6. The court undertakes no detailed comparison between the Mattis policy and the military’s longstanding prior policy, or between the Mattis policy and the Carter policy, or between the Mattis policy and the outstanding preliminary injunction.

August 24, 2018: The district court denies the government’s motion for summary judgment, holding that plaintiffs are "entitled to complete discovery" regarding the executive branch’s "alleged deliberation" [sic] over the Mattis policy. 322 F.Supp.3d at 101.

November 30, 2018: The district court denies the government’s motion for a stay pending appeal. See Doe 2 v. Mattis , 344 F.Supp.3d 16 (D.D.C. 2018).

II

On January 4, 2019, we reversed, and I join my colleagues in dissolving the injunction. But this is where I must part ways.

A

The panel assumes that we are "reviewing" not a new or modified injunction, but a "motion to dissolve" a pre-existing one. Panel Judgment *1. This matters, the panel says, because our review of a motion to dissolve is particularly narrow: The only issue is whether the government has shown a "significant change" in circumstances such that continued enforcement of the original injunction may no longer be appropriate. Id ."The merits of the preliminary injunction," the panel concludes, "are not properly before us." Id . But on a sound view of the case, the merits are before us.

1. The merits are properly before us because we confront a necessarily new or modified injunction.

First, although the government styled its request for relief as a motion to dissolve the earlier preliminary injunction, it made clear its view that the actions of Secretary Mattis were so different from anything enjoined that the earlier injunction did not apply. See, e.g., Defs.’ Mot. to Dissolve the Preliminary Injunction 2 (Apr. 20, 2018), ECF No. 116; see also Appellants’ Br. 2, 3, 15, 16, 43, 47 (arguing that the district court extended the earlier, October 2017 injunction). When a district court uses a preliminary injunction addressed to certain conduct as a vehicle for barring completely different conduct, it has either "necessarily modified" the prior injunction or issued a "completely new" one. Int’l Ass’n of Machinists &Aerospace Workers v. Eastern Air Lines, Inc. , 849 F.2d 1481, 1486 (D.C. Cir. 1988). Either way, our review is standard and far from narrow: The precedents require us to assess "the district court’s legal conclusions as to each of the four" preliminary injunction factors—including plaintiffs’ likelihood of "success on the merits"—"de novo , and its weighing of them for abuse of discretion." Archdiocese of Wash. v. Wash. Metro. Area Transit Auth. , 897 F.3d 314, 321 (D.C. Cir. 2018) (quoting League of Women Voters v. Newby , 838 F.3d 1, 6–7 (D.C. Cir. 2016) ). Here, the likelihood of plaintiffs’ success on the merits turns on the probability that their constitutional claim is valid.

If not completely new, the order before us—the August 6, 2018 injunction—is a modified injunction. In that order, the district court purported to "clarif[y]" that the Mattis policy "violate[d] the Court’s [pre-existing] October 30, 2017 preliminary injunction." Doe II , 315 F.Supp.3d at 498. But "we are not governed by the district court’s own characterization of the order as [a] ... ‘clarification,’ as distinguished from a ‘modification.’ " United States v. Philip Morris USA Inc. , 686 F.3d 839, 844 (D.C. Cir. 2012) (quoting Birmingham Fire Fighters Ass’n 117 v. Jefferson Cnty. , 280 F.3d 1289, 1292 (11th Cir. 2002) ); see also Abbott v. Perez , ––– U.S. ––––, 138 S.Ct. 2305, 2320, 201 L.Ed.2d 714 (2018) (refusing to "allow[ ] district courts to ‘shield [their] orders from appellate review’ " with the flip of a "label" (alteration in original) (quoting Sampson v. Murray , 415 U.S. 61, 87, 94 S.Ct. 937, 39 L.Ed.2d 166 (1974) )). Rather, we must assess "the scope" of the October 2017 injunction with our own " ‘independent judgment.’ " Philip Morris , 686 F.3d at 844 (quoting Int’l Ass’n of Machinists , 849 F.2d at 1485 ).

The original October 2017 injunction "enjoin[ed] Defendants from enforcing [two] directives of the [2017] Presidential Memorandum": (1) " ‘return to the longstanding policy and practice on military service by transgender individuals that was in place prior to [the Carter policy]’ "; and (2) " ‘maintain the [longstanding, pre-Carter] policy regarding accession of transgender individuals into military service.’ " Order 1–2 (Oct. 30, 2017), ECF No. 60, J.A. 188–89 (quoting 2017 Presidential Memorandum , 82 Fed. Reg. at 41,319). But the Mattis policy does neither of the actions forbidden in the language quoted above. The "longstanding policy" was a "blanket ban on all ‘transgender individuals,’ " Doe II , 315 F.Supp.3d at 481 ; whereas, the Mattis policy, as the panel correctly concludes, "allows some transgender persons barred under the ... [longstanding policy] to join and serve in the military," Panel Judgment *2 ; see also id . ("[T]he District Court erred in finding that the Mattis Plan was a blanket transgender ban."). Because the defendants were not "already bound by the" October 2017 injunction not to pursue the Mattis policy, Washington Metro. Area Trans. Comm’n v. Reliable limousine Serv. LLC, 776 F.3d 1, 9 (D.C. Cir. 2015), the August 2018 order "change[d] the legal relationship of the parties," id . (citing Philip Morris , 686 F.3d at 844 ). The August 2018 injunction is thus fully subject to conventional review.

An alternative analysis leads to the same result. The August 2018 order is subject to full review because at the time of its issuance the plaintiffs’ lawsuit challenging the President’s 2017 directives—the basis of the October 2017 preliminary injunction—was moot. The district court, to be sure, disagreed, see Doe II , 315 F.Supp.3d at 492–96, but, again, we review its "interpretation of [the prior] injunction de novo," United States ex rel. Yelverton v. Fed. Ins. Co. , 831 F.3d 585, 587 (D.C. Cir. 2016) (citing Int’l Ass’n of Machinists , 849 F.2d at 1485 ). And that October 2017 injunction addressed the 2017 Presidential Memorandum, which the President revoked , see 83 Fed. Reg. at 13,367, well before the district court issued its August 2018 injunction. Although such a repeal and replace does not always moot a claim, it does so where the challenged policy is "changed substantially," NE Fla. Chapter of Associated Gen.Contractors of Am. v. City of Jacksonville , 508 U.S. 656, 662 n.3, 113 S.Ct. 2297, 124 L.Ed.2d 586 (1993), as the Presidential Memorandum undoubtedly was, see Panel Judgment *2 ("The government took substantial steps to cure the procedural deficiencies the court identified in the enjoined 2017 Presidential Memorandum."). See also Trump v. Int’l Refugee Assistance , ––– U.S. ––––, 138 S.Ct. 353, 353, 199 L.Ed.2d 203 (2017) (dismissing as moot a challenge to an expired presidential order that suspended the entry of certain aliens, even though the President had already issued a new, related order); Trump v. Hawaii , ––– U.S. ––––, 138 S.Ct. 377, 377, 199 L.Ed.2d 275 (2017) (same). Thus, the district court "necessarily ... issued a completely new injunction." Int’l Ass’n of Machinists , 849 F.2d at 1486. Again, the history of this case puts the merits fully before us.

2. Even viewed solely as a motion to dissolve, the case calls for review of the Mattis policy’s constitutionality.

In any event, even if we were simply reviewing a motion to dissolve, we still must resolve the constitutionality of the Mattis policy. A "significant change" in circumstances is only a "threshold " showing to dissolve an injunction. Salazar ex rel. Salazar v. District of Columbia , 896 F.3d 489, 492 (D.C. Cir. 2018) (emphasis added) (citing Rufo v. Inmates of Suffolk Cnty. Jail , 502 U.S. 367, 383, 112 S.Ct. 748, 116 L.Ed.2d 867 (1992) ). Once that threshold is met, as here, we must decide whether the change in circumstances "renders continued enforcement" of the injunction "detrimental to the public interest." Petties ex rel. Martin v. District of Columbia , 662 F.3d 564, 569 (D.C. Cir. 2011) (quoting Horne v. Flores , 557 U.S. 433, 447, 129 S.Ct. 2579, 174 L.Ed.2d 406 (2009) ). To make this determination, we "need[ ] to ascertain whether ongoing enforcement of the original [preliminary injunction] [is] supported by an ongoing violation of federal law ." Horne , 557 U.S. at 454, 129 S.Ct. 2579 (emphasis added); but cf. O p. 704–05 (Wilkins, J., concurring) (declining to address whether the Mattis policy violated federal law). And we do that by assessing whether constitutional "compliance has been achieved" by the Mattis policy—that is, we decide whether the Mattis policy is constitutional. Id . at 452, 129 S.Ct. 2579 (citing Frew ex rel. Frew v. Hawkins , 540 U.S. 431, 442, 124 S.Ct. 899, 157 L.Ed.2d 855 (2004) ).

Proceeding with that assessment is especially appropriate where, as here, the district court moved well beyond "changed circumstances" and directly ruled "[o]n the merits" of the Mattis policy. See Doe II , 315 F.Supp.3d at 497 (addressing all the preliminary injunction factors: "the merits," "irreparable injury," "the balance of equities," and the "public interest"); see also id . at 492 (ruling that "even if" the plaintiffs’ claims against the 2017 Presidential Memorandum were "moot," the court would still enjoin implementation of the Mattis policy). The panel decision quite properly reflects this aspect of the record, citing the district court’s discussion of "the merits," id . at 497–98, cited in Panel Judgment *1 n.1, and "revers[ing]," "[t]o the extent that the District Court granted preliminary relief ... apart from its refusal to dissolve the existing injunction," Panel Judgment *1 n.1 (emphasis added). Accordingly, I proceed to the question of whether the injunction "was supported by an ongoing violation of federal law." Horne , 557 U.S. at 454, 129 S.Ct. 2579.

B

More generally, "once jurisdiction is properly invoked under § 1292(a)(1)," as it is here, our review permits "disposition of all matters appropriately raised by the record, including entry of final judgment." Hartman v. Duffey , 19 F.3d 1459, 1464 (D.C. Cir. 1994) (quoting Wagner v. Taylor , 836 F.2d 578, 585 (D.C. Cir. 1987) ). "This has long been the rule: ‘By the ordinary practice in equity ...,’ a reviewing court has the power on appeal from an interlocutory order ‘to examine the merits of the case ... and upon deciding them in favor of the defendant to dismiss the bill." Munaf v. Geren , 553 U.S. 674, 691, 128 S.Ct. 2207, 171 L.Ed.2d 1 (2008) (second alteration in original) (quoting N.Carolina R. Co. v. Story , 268 U.S. 288, 292, 45 S.Ct. 531, 69 L.Ed. 959 (1925) ); see also, e.g., Ark. Dairy Co-op Ass’n v. U.S. Dep’t of Agriculture , 573 F.3d 815, 833 (D.C. Cir. 2009) ("Review of a preliminary injunction ‘is not confined to the act of granting [or denying] the injunctio[n], but extends as well to determining whether there is any insuperable objection, in point of jurisdiction or merits, to the maintenance of [the] bill, and if so, to direct a final decree dismissing it.’ " (alterations in original) (quoting Munaf , 553 U.S. at 691, 128 S.Ct. 2207 )).

III

I turn now to the merits.

A

In our government of separate and coordinate powers, it is a "basic principle" that "one branch of the Government may not intrude upon the central prerogatives of another." Loving v. United States , 517 U.S. 748, 757, 116 S.Ct. 1737, 135 L.Ed.2d 36 (1996). This restraint serves an important purpose: it preserves the government of our people’s design. Those who drafted and established our Constitution did not dole out federal power haphazardly; rather, to create "a National Government that is both effective and accountable," they painstakingly allocated "specific powers and responsibilities" to the branches of government best "fitted to the task." Id .

To the branches "periodically subject to electoral accountability," Gilligan v. Morgan , 413 U.S. 1, 10, 93 S.Ct. 2440, 37 L.Ed.2d 407 (1973), went "authorities essential to the common defense," The Federalist No. 23 , at 153 (Alexander Hamilton) (Clinton Rossiter ed., 1961); see U.S. Const. art I, § 8, cl. 11 (declare war); cl. 12 (raise and support armies); cl. 13 (provide and maintain a navy); cl. 14 (make rules for the land and naval forces); cl. 15 (provide for calling forth the militia); cl. 16 (provide for organizing the militia); art. II, § 2, cl. 1 (commander in chief). For good reason. It is "difficult to conceive of an area of governmental activity in which the courts have less competence." Gilligan , 413 U.S. at 10, 93 S.Ct. 2440. The political branches are far better equipped (and more accountable to the people) for making the types of decisions that arise in the military setting—decisions that must be made quickly and based on limited, often secret, information.

Our Constitution thus vests the "power of oversight and control of military force" in Congress and the President. Id . And it does so, as Hamilton put it,

without limitation, because it is impossible to foresee or define the extent and variety of national exigencies, and the correspond[ing] extent and variety of the means which may be necessary to satisfy them . The circumstances that endanger the safety of nations are infinite, and for this reason no constitutional shackles can wisely be imposed on the power to which the care of it is committed.

The Federalist No. 23 , at 153 (emphasis in original). "The later-added Bill of Rights," of course, "limited this power to some degree," but "did not alter" the people’s basic "allocation" of authority, Loving , 517 U.S. at 767, 116 S.Ct. 1737 (citations omitted); Congress and the President—not the courts—retained the primary responsibility for the "delicate task of balancing the rights of servicemen against the needs of the military," id . (quoting Solorio v. United States , 483 U.S. 435, 447–48, 107 S.Ct. 2924, 97 L.Ed.2d 364 (1987) ); see also, e.g., United States v. Eliason , 41 U.S. (16 Pet.) 291, 301, 10 L.Ed. 968 (1842) ("The power of the executive to establish rules and regulations for the government of the army, is undoubted.").

Judges have long respected this allocation of powers. Gilligan , 413 U.S. at 11, 93 S.Ct. 2440. Thus, "courts traditionally have been reluctant to intrude upon the authority of the Executive in military and national security affairs," Dep’t of Navy v. Egan , 484 U.S. 518, 530, 108 S.Ct. 818, 98 L.Ed.2d 918 (1988), and have "hesitate[d] long" before tampering with the "unique structure of the military establishment," Chappell v. Wallace , 462 U.S. 296, 300, 103 S.Ct. 2362, 76 L.Ed.2d 586 (1983). In the military realm, citing separation of powers principles, the Supreme Court has rejected numerous attacks on the President’s authority, e.g., to define factors for imposition of the death penalty in military trials, see Loving , 517 U.S. at 768–69, 116 S.Ct. 1737 ; to "control access to information bearing on national security," see Egan , 484 U.S. at 527, 108 S.Ct. 818 ; and to make commissioning of officers dependent on "whatever facts [he] thinks" relevant, see Orloff v. Willoughby , 345 U.S. 83, 91, 73 S.Ct. 534, 97 L.Ed. 842 (1953). The issues were, in the words of Loving , within with the "central prerogative[ ]" of the executive, not the courts. 517 U.S. at 757, 116 S.Ct. 1737 ; see also Orloff , 345 U.S. at 93, 73 S.Ct. 534 ("[J]udges are not given the task of running the Army.").

Even addressing laws or regulations that would be subject to heightened scrutiny if applied to civilian society (as plaintiffs assume for the Mattis policy, see Br. 29–32), our "review of military regulations ... is far more deferential," Goldman v. Weinberger , 475 U.S. 503, 507–08, 106 S.Ct. 1310, 89 L.Ed.2d 478 (1986), and "the tests and limitations to be applied may differ because of the military context," Rostker , 453 U.S. at 67, 101 S.Ct. 2646.

Although the Supreme Court has eschewed a "label[ ]," id . at 69, 101 S.Ct. 2646, our standard of review reflects the separation of powers principles detailed above, see id . at 67, 101 S.Ct. 2646 (recognizing "that the Constitution itself requires such deference"). Thus—even when reviewing a facially discriminatory, gender-based classification, such as an exclusion of women from draft registration (see Rostker )—it is "quite wrong" and "palpably exceed[s]" our authority to "undertak[e] an independent evaluation of [ ] evidence, rather than adopting an appropriately deferential examination" of the political branches’ "evaluation of that evidence." Id . at 81, 82–83, 101 S.Ct. 2646 ; accord, e.g., Trump v. Hawaii , ––– U.S. ––––, 138 S.Ct. 2392, 2421–22, 201 L.Ed.2d 775 (2018) (declining in matters of national security to "substitute" the Court’s own "predictive judgments," or its own "evaluation of the underlying facts," for those of the President); Goldman , 475 U.S. at 509, 106 S.Ct. 1310.

Under these principles, I believe the Mattis policy survives constitutional scrutiny.

B

To begin, plaintiffs do not—and cannot—seriously challenge the Mattis policy’s reliance on gender dysphoria for limiting the access of those suffering from that medical condition to military service. They, after all, sought and the district court ordered, see Second Am. Compl. 20, J.A. 209; Doe II , 315 F.Supp.3d at 480, 484, a return to the Carter policy, which uses the same condition (a "history of gender dysphoria") to presumptively disqualify persons from accession into the military, subject to some exceptions. Carter Memo , attachment at 1, J.A. 588; cf. Mattis Memo 2, J.A. 264.

In any event, to the extent plaintiffs contend that the Mattis policy’s focus on gender dysphoria springs from hostility to transgender persons, this aspect of the policy is far from suspect. Gender dysphoria affects "[o]nly a subset of " (or "some ") "transgender people," as even plaintiffs’ experts repeatedly admit. Decl. of George Richard Brown in Support of Opp’n to Defs.’ Mot. to Dismiss ¶ 9 (May 11, 2018), J.A. 839 (emphasis added); Decl. of Brad R. Carson in Support of Pls.’ Mot. for Preliminary Injunction ¶ 23 (Aug. 28, 2017), J.A. 995 (emphasis added); accord, e.g., Carter Memo Implementation Handbook 13, J.A. 520 ("some transgender individuals"); DoD Report 19, J.A. 287 ("subset of transgender persons"); id . at 20, J.A. 288 ("[n]ot all transgender people" (alteration in original) (quoting American Psychiatric Association)); RAND Report 7, J.A. 623 ("[s]ome transgender individuals (again, the proportion is largely unknown)"); id . at x, J.A. 606 ("transgender status alone does not constitute a medical condition"; "only transgender individuals who experience significant related distress are considered to have" gender dysphoria ); id . at 6, J.A. 622 (similar); AMA Br. 7 (similar).

Gender dysphoria is also a serious mental health condition recognized by the American Psychiatric Association. It is "associated with clinically significant distress or impairment in social, occupational, or other important areas of functioning." DoD Report 13, J.A. 281 (quoting Am. Psychiatric Ass’n, Diagnostic and Statistical Manual of Mental Disorders 453 (5th ed. 2013)); accord, e.g., Geoffrey M. Reed et al., Disorders Related to Sexuality and Gender Identity in the ICD-11 , 15 World Psychiatry 205, 212 (2016), J.A. 934 ("distress and dysfunction [are] integral aspects of the condition"); RAND Report at 6, J.A. 622 ("significant related distress"); Report of the Transgender Military Service Commission 10 (Mar. 2014), J.A. 757 ("clinically significant distress"); AMA Br. 2 (same); see also Carter Memo , attachment at 1, J.A. 588; Carter Memo Implementation Handbook 12, J.A. 519. Untreated, it "can cause debilitating distress, depression, impairment of function, self-mutilation to alter one’s genital or secondary sex characteristics, other self-injurious behaviors, and suicide." AMA Br. 9.

The limits on accession and retention of those with gender dysphoria, imposed by the military (under both Secretaries Carter and Mattis), are thus akin to the many demanding selection practices that render the "vast majority" of military-age Americans—a full 71%— presumptively ineligible. DoD Report 6, J.A. 274; see DoD Instruction 6130.03: Medical Standards for Appointment, Enlistment, or Induction Into the Military Services (Mar. 30, 2018), J.A. 212 (providing 49 pages of medical guidelines). This includes disqualification for "[a]ny" DSM -recognized condition that (like gender dysphoria ) is associated with "residual symptoms, or medication side effects, which impair social or occupational performance." DoD Report 34, J.A. 302; see, e.g., DoD Instruction 6130.03 , supra , at 44–46, J.A. 255–57 (disqualifying those with a "[h]istory of learning disorders after the 14th birthday," a "[h]istory of obsessive-compulsive disorder," or a "[r]epeated inability to maintain reasonable adjustment in school"). In presumptively disqualifying individuals with gender dysphoria, the Mattis policy, like the Carter policy before it, serves the same legitimate interests as other disqualifications: ensuring that the armed forces consist "of qualified, effective, and able-bodied persons." 10 U.S.C. § 505(a).

C

Plaintiffs, therefore, hitch their wagon to a different star: the accommodations (or lack thereof) for "gender transition."

Gender transition, in their view, is "the defining characteristic of being transgender." Appellees’ Br. 20. As plaintiffs see it, "transgender people," "[b]y definition," do not "live in accord with their biological sex." Id . at 19 (emphasis added) (quoting Doe II , 315 F.Supp.3d at 482 ). Rather, "transgender people undergo a gender transition in order to live consistently with their gender identity." Id . As we’ll soon see, the record thoroughly belies this claim.

Working from this premise, plaintiffs contend, and district court agreed, that the Mattis policy "is a ban on transgender service." Id . (quoting Doe II , 315 F.Supp.3d at 495 (quoting, in turn, Pls.’ Opp’n to Mot. to Dismiss 10 (May 11, 2018), ECF No. 130)). That is so, they say, because the Mattis policy, unlike the Carter policy, does not accommodate gender transition. Whereas the Carter policy allows individuals to undergo a gender transition (ranging, as we have seen, from purely "social" changes to complete transformational surgery) and to then serve in their preferred gender (rather than their biological sex), Carter Memo , attachment at 1–2, J.A. 588–89; Carter Memo Implementation Handbook 43, J.A. 550, the Mattis policy sticks to longstanding military practice: "all" individuals must serve in their "biological sex." Thus, individuals who have undergone gender transition (to live outside their biological sex) are ineligible, Mattis Memo 2–3, J.A. 264–65 (unless they qualify for a waiver or the reliance exception).

Plaintiffs’ argument that this policy difference is of constitutional magnitude is fundamentally flawed—factually and legally.

1. The facts relating to transition for transgender persons and to the polices.

At oral argument, plaintiffs doubled down on their claim that a failure to accommodate gender transition (and life outside one’s biological sex) is, in effect, a ban on transgender service. Their counsel’s first substantive words were as follows:

[The Mattis policy] bans transgender people from the military, it bans only transgender people, and all transgender people. [The Mattis policy] requires anyone who serves to do so in their biological sex, [and] not living in a person’s biological sex is the defining characteristic of what it means to be transgender.

Oral Arg. Tr. 19:10–16.

That is wrong. By requiring most individuals to serve in their biological sex, the Mattis policy does not "translate[ ] into a ban on transgender persons." Appellees’ Br. 20 (quoting Doe II , 315 F.Supp.3d at 494 ). This is so for at least three reasons.

First, the factual reality: As the panel found, there is "nothing in the record" to support plaintiffs’ assertion that the "defining" characteristic of being "transgender" is not living in one’s biological sex. Panel Judgment *2. To the contrary, "everything in the record, from the Elders Report, the RAND Study, the Carter Memo and the Mattis Memo consistently define a transgender person as a person who does not ‘identify ’ with [his or her] biological sex." O p. 701 (Wilkins, J., concurring) (emphasis added); see, e.g., DoD Report 7 n.10, J.A. 275; RAND Report at x, J.A. 606; Report of the Transgender Military Service Commission 5 (Mar. 2014), J.A. 752; Carter Memo Implementation Handbook 13, J.A. 520. Even plaintiffs’ own expert, along with the American Medical Association (in support of plaintiffs), define "transgender" in terms of how one "identi[fies] ," not how one lives . See Brown 2017 Decl. ¶ 13, J.A. 1056 (emphasis added); AMA Br. 3 (same).

There is, moreover, no doubt that many transgender individuals can, do, and wish to serve in their biological sex . Contrary to plaintiffs’ assertion, the transgender community is not a monolith in which every person wants to take steps necessary to live in accord with his or her preferred gender (rather than his or her biological sex). Quite the opposite. The implementation handbook for the Carter policy, for example, explains that only "some" transgender individuals "feel compelled" to "transition to the preferred gender." Carter Memo Implementation Handbook 13, J.A. 52. The Elders Report adds that the "transgender community includes people" who do not even "wish to transition." Report of the Transgender Military Service Commission 5, J.A. 752. And the RAND Report specifies that only 55% of "transgender individuals" reported "living" in their preferred gender; of the remainder, only half—27% of the total—even "wished to transition at some point in the future." RAND Report 20 & n.2, J.A. 636 & n.2. This leaves but one conclusion, as the military has already found: transgender individuals can serve and are "serving[ ] with distinction ... [in] their biological sex." Panel Judgment *2 (quoting DoD Report 6, J.A. 274).

Second, the Mattis policy’s exceptions to the requirement of service in one’s biological sex: Again, as the panel found, the policy "contains a reliance exception," Panel Judgment *2, that expressly permits certain transgender individuals "to serve in their preferred gender," rather than their biological sex, Mattis Memo 2, J.A. 264. This exception covers around 1000 individuals, see DoD Report 22, J.A. 290, so that even on the false premise that transgender must be equated with transition, the policy would not be a "ban" on "all transgender people." It was "clear error" for the district court to hold otherwise. Panel Judgment *2. Moreover, the military reserves the authority—as it does for virtually all standards, DoD Report 10, J.A. 278—to issue waivers for persons who would otherwise be ineligible, id . at 5, J.A. 273.

Third, the internal inconsistency in plaintiffs’ position: Plaintiffs demand a return to the Carter policy. See Second Am. Compl. 20, J.A. 209; Doe I , 275 F.Supp.3d at 177. But even that policy, plaintiffs admit, requires transgender persons to serve in their biological sex until their gender transition is "complete." Appellees’ Br. 45; see Carter Memo Implementation Handbook 11, J.A. 518. So, if "a requirement to serve in one’s biological sex is a transgender ban," Appellees’ Br. 15, plaintiffs’ logic seems to doom the very remedy they now seek.

To put it simply: The Mattis policy—by declining to make the same accommodations for gender transition as the Carter policy—does not translate into a "transgender ban." See Panel Judgment *2 (holding that the district court "clear[ly] err[ed]" in "finding that the Mattis Plan was the equivalent of a blanket ban on transgender service"). And no amount of discovery will change "this fact." But cf. O p. 705 n.* (Wilkins, J., concurring).

2. The executive branch relied on an abundance of legitimate military concerns.

In reasoning about the policy issues it confronted, the executive branch relied on an abundance—some might say a superabundance—of concerns. Given the "biological differences" between males and females, military "policy and practice has long maintained a clear line between" the biological sexes. DoD Report, 41, J.A. 309. The military, for example, maintains separate berthing, bathroom, and shower facilities, along with different sets of physical fitness, body fat, uniform, and grooming standards, for biological males and biological females, id .—as required (in part) by Congress, see 10 U.S.C. §§ 7419, 7420, 8431, 8432, 9419. The DoD Report drafted at Secretary Mattis’s request found that these "clear sex-differentiated lines" preserve unit cohesion by, among things, protecting reasonable expectations of privacy, avoiding unfairness (or perceptions thereof), ensuring physical safety, and minimizing leadership challenges that would otherwise arise. DoD Report 40, J.A. 308.

The Supreme Court has expressly acknowledged the (rather widely shared) understanding of a key premise of the military judgment: that, in the military setting, "[p]hysical differences between men and women ... are enduring: ‘[T]he two sexes are not fungible[.]’ " Virginia , 518 U.S. at 533, 116 S.Ct. 2264 (quoting Ballard v. United States , 329 U.S. 187, 193, 67 S.Ct. 261, 91 L.Ed. 181 (1946) ). For that reason, "[s]ex classifications may be used[.]" Id . It is, the Court has said, "undoubtedly ... necessary to afford members of each sex privacy from the other sex in living arrangements" and to "adjust aspects of the physical training programs" in light of the biological differences between males and females. Id . at 550, 116 S.Ct. 2264 n.19 ; see also Rostker , 453 U.S. at 81, 101 S.Ct. 2646 (recognizing the need for "different treatment" of males and females "with regard to dependency, hardship, and physical standards").

Unsurprisingly, then, plaintiffs "do not challenge the military’s ability to maintain any of its sex-based standards," Appellees’ Br. 44; instead they try to work around it with a linguistic device. They contend that "the military can maintain sex-based standards" by applying to each service member the standard (male or female) that happens to match his or her "gender marker in [the military’s] personnel database." Id . at 45 (citing Carter Memo Implementation Handbook 11, J.A. 518). But for transgender persons who have exercised the asserted right to transition (whether socially, hormonally or surgically), that "marker" would only randomly relate to the physical differences between biological males and biological females (as differentiated by "chromosomes, gonads, hormones, and genitals," DoD Report 7 n.10, J.A. 275). Those physical differences, however, are what the military’s sex-based standards are actually based upon. And they are what the Supreme Court’s approval of those standards, quoted in the paragraph immediately above, actually rested on.

Plaintiffs’ plan would demolish the military’s sex-based standards. Under their view, individuals could switch genders (and adjust their gender marker) by simply "living in the preferred gender." Carter Memo Implementation Handbook 31, J.A. 538. In other words, a person need not undergo any "sex reassignment surgery," id . at 50, J.A. 557, or even "cross-sex hormone therapy," id . at 37, J.A. 544, in order to be recognized as, and thus subject to the standards associated with, his or her "preferred gender," id . at 43, J.A. 550. Instead, any person who remained (for example) biologically male in every respect could demand to be treated, in all respects, as female, including for "medical fitness, physical fitness, uniform and grooming, deployability and retention standards," as well as for assignment to "berthing, bathroom, and shower facilities." Id . at 43–44, J.A. 550–51.

Plaintiffs’ "approach[ ]" may be wise (or equitable). Appellees’ Br. 24. But it in no way responds to the concerns over unit cohesion, unit readiness, and related burdens relied on by the executive:

Unit cohesion. Consider first service members’ "reasonable expectations of privacy." DoD Report 37, J.A. 305. The military explained that "[g]iven the unique nature of military service," service members "of the same biological sex are often required to live in extremely close proximity to one another when sleeping, undressing, showering, and using the bathroom." Id . In these circumstances, permitting an individual with (for example) "male genitalia" to live, sleep, and shower with biological females would surely, as the military concluded, create some tension within the ranks—as even the Carter Memo Implementation Handbook admits, see pp. 63–64, J.A. 570–71 (explaining that biological females may be "uncomfortable"); see also id . at 29, J.A. 536 (discussing "privacy" concerns about "showers, bathrooms, or other shared spaces"). And while reasonable people might disagree, cf. id . at 65, J.A. 572 (ruling that biological females’ "lack of comfort" must give way to more egalitarian concerns), the executive branch "was certainly entitled, in the exercise of its constitutional powers ..., to focus on the question of military need rather than ‘equity,’ " Rostker , 453 U.S. at 80, 101 S.Ct. 2646 ; cf. Virginia , 518 U.S. at 550 n.19, 116 S.Ct. 2264 (recognizing that it is "necessary to afford members of each sex privacy from the other sex in living arrangements").

The military’s privacy concerns are grounded in—and partially mandated by—statute. The military is statutorily required to physically separate "males" and "females" in various contexts. See 10 U.S.C. §§ 7419, 7420, 8431, 8432, 9419 ; see also id . § 9420 ("same sex"). Plaintiffs have not challenged—or even mentioned—those provisions. Thus we must examine plaintiffs’ constitutional objections with these unchallenged "restrictions firmly in mind." Rostker , 453 U.S. at 77, 101 S.Ct. 2646 ; see also Schlesinger v. Ballard , 419 U.S. 498, 508, 95 S.Ct. 572, 42 L.Ed.2d 610 (1975). The very "existence" of the "restrictions clearly indicates" a legitimate basis for the policy’s bright-line separation of biological males and biological females, Rostker , 453 U.S. at 77, 101 S.Ct. 2646 —i.e., compliance with a congressional directive, see DoD Report 29 & n.108, 37, J.A. 297, 305. We must interpret statutory words "consistent with their ‘ordinary meaning ... at the time Congress enacted the statute.’ " Wisconsin Central Ltd. v. United States , ––– U.S. ––––, 138 S.Ct. 2067, 2070, 201 L.Ed.2d 490 (2018) (alteration in original) (quoting Perrin v. United States , 444 U.S. 37, 42, 100 S.Ct. 311, 62 L.Ed.2d 199 (1979) ). And at the time Congress enacted the privacy statutes at issue here—in 1998—there can be little doubt what "male" and "female" meant in ordinary parlance. See Pub. L. No. 105-261, §§ 521, 522, 112 Stat. 1920, 2009–13 (1998). Those words plainly referred to—and thus required separation of individuals on the basis of—biological sex , not preferred gender. See, e.g., Webster’s Third New International Dictionary 1366 (unabr. ed. 1993) (defining "male" in terms of "being the sex" that "perform[s] the fertilization function"); The American Heritage Dictionary 759 (2d coll. ed. 1985) ("the sex that has organs to produce spermatozoa"); The Randomhouse College Dictionary 809 (rev. ed. 1980) ("the sex that begets young by fertilizing the female"); see also Black’s Law Dictionary 1379 (7th ed. 1999) (defining "sex" in terms of the "structure[s] and function[s] that distinguish a male from a female organism").

Safety . The military explained that "combat remains a physical endeavor." DoD Report 37, J.A. 305. So, naturally, "vigorous competition, especially physical competition," remains "central to the military life and is indispensable to the training and preparation of warriors." Id . at 36, J.A. 304. But in such "physically violent training and competition," the military concluded, "pitting biological females against biological males who identify as female, and vice versa"—as plaintiffs demand, see, e.g., Appellees’ Br. 45 (demanding treatment in preferred gender "for all purposes")—would create a "serious safety risk." DoD Report 36, J.A. 304. That is a risk that the military is not constitutionally required to accept (or force upon its members). See Virginia , 518 U.S. at 550 n.19, 116 S.Ct. 2264 (recognizing that it is "necessary" to "adjust aspects of the physical training programs" for biological males and biological females).

Unfairness or the perception thereof also provides a legitimate basis for the Mattis policy—and easily falls within our precedent. The military concluded that it "could be perceived as discriminatory to apply different biologically-based standards to persons of the same biological sex based on gender identity, which is irrelevant to standards grounded in physical biology." DoD Report 36, J.A. 304. "For example, it unfairly discriminates against biological males who identify as male and are held to male standards to allow biological males who identify as female to be held to female standards, especially where the transgender female retains many of the biological characteristics and capabilities of a male." Id . It would also "result in perceived unfairness by biological females who identify as female" and would "be required to compete against" biological males who identify as female "in training and athletic competition." Id .

In the military’s view, plaintiffs’ demands would pit the interests of transgender service members against the interests of a potentially vast array of other service members who will be placed in an awkward, potentially unfair, and possibly bitterness-inducing posture of having to compete against those who are either far stronger or weaker than they are. It may turn out that these worries are overblown. Or it may be that wisdom counsels running the risks in order to accommodate transgender service members’ preferences. But to pretend that the military cannot rationally be concerned about this novel source of potential social conflict, resentment, and intra-unit strife is untenable. The concern is at least reasonable, and the courts cannot second-guess the military’s decision.

Indeed, if, as we’ve previously held, the military could refuse to exempt certain religious males from male uniform standards, so as to avoid "incurring resentment from those who are compelled to adhere to the rules strictly," Goldman v. Sec’y of Defense , 734 F.2d 1531, 1540 (D.C. Cir. 1984), aff’d sub nom. Goldman v. Weinberger , 475 U.S. 503, 106 S.Ct. 1310, 89 L.Ed.2d 478, then surely the military could, for the same reason, refuse to exempt certain biological males from male dress standards (not to mention body composition assessments and physical readiness testing), DoD Report 31, J.A. 299. After all, the ability to express one’s "religious ... identity," Goldman , 475 U.S. at 517, 106 S.Ct. 1310 (Brennan, J., dissenting), is surely no less protected than the ability to express one’s "gender identity," Appellees’ Br. 21; cf. U.S. Const. amend. I.

Unit readiness. Gender transition is available under the Carter policy only as a treatment for gender dysphoria. See, e.g., Carter Memo Implementation Handbook 11, 14, J.A. 518, 521; Carter Memo Implementation Handbook: Navy Supplement 2, J.A. 461; see also Appellees’ Br. 19, 41, 42 (referring to transition as "treatment"). The military concluded here that accommodating requests for transition would "present a significant challenge for unit readiness." DoD Report 35, J.A. 303. As its report noted, "there is considerable scientific uncertainty concerning whether" gender transition treatments (e.g., sex reassignment surgery) "fully remedy ... the mental health problems associated with gender dysphoria." Id . at 32, J.A. 300; see, e.g., id . at 24, J.A. 292 (reporting the conclusion of the Centers for Medicare and Medicaid Services "that there was ‘not enough high quality evidence to determine whether gender reassignment surgery improves health outcomes for Medicare beneficiaries with gender dysphoria ’ "); RAND Report 10, J.A. 626 ("[I]t is important to note that none of these studies [on gender-transition related healthcare] were randomized controlled trials (the gold standard for determining treatment efficacy). In the absence of quality randomized trial evidence, it is difficult to fully assess the outcomes of treatment for [gender dysphoria ]."). Those mental health problems, if un-remedied, are substantial: Gender dysphoria is undisputedly associated with "clinically significant distress," e.g., Decl. of George Richard Brown in Support of Plaintiffs ¶ 8 (May 11, 2018), J.A. 838; Report of the Transgender Military Service Commission 10 (Mar. 2014), J.A. 757; AMA Br. 2; see also Appellees’ Br. 19 (noting the "distress"), and can cause "depression, impairment of function, self-mutilation to alter one’s genital or secondary sex characteristics, other self-injurious behaviors, and suicide," AMA Br. 9. Any "persistence of these problems is," as the military reasonably concluded, "a risk for readiness." DoD Report 32, J.A. 300. The military’s decision to "proceed cautiously" was thus reasonable. Id . at 27, J.A. 295. We cannot, after all, "substitute our own assessment for the Executive’s predictive judgments" on matters of "national security." Hawaii , 138 S.Ct. at 2421.

Administrative burdens . Such burdens are "not for this Court to dismiss." Rostker , 453 U.S. at 81, 101 S.Ct. 2646 (upholding gender-based classification where "Congress simply did not consider it worth the added [administrative] burdens of including women in draft and registration plans"). As the military recognized, the "unique leadership challenges arising from gender transition are evident in the [ ] handbook implementing the Carter policy." DoD Report 38, J.A. 306. This 71-page handbook—actually touted by plaintiffs and the district court as evidence of the Carter policy’s feasibility, see, e.g., Doe 2 v. Mattis , 344 F.Supp.3d 16, 28–29 (D.D.C. 2018) ; Decl. of Deborah Lee James in Support of Pls.’ Mot. for Preliminary Injunction ¶ 34 (Aug. 29, 2017), J.A. 1008–09—provides guidance on "some of the issues" that commanders will face under the Carter policy. Carter Memo Implementation Handbook 8, J.A. 515. Consider one example—a simple swim test, required semi-annually throughout the military: "[A] female to male transgender Service member who has fully transitioned, but did not undergo surgical change, would like to wear a male swimsuit for the test with no shirt or other top coverage." Id . at 63, J.A. 570. The commanding officer has much to do. First, consult the handbook for his options: "counsel the individual," "address the unit," and consider "additional options (e.g., requiring all personnel to wear shirts)." Id . But the handbook is not "directive in nature"; it provides only a "general discussion." Id . at 48, J.A. 555. So, the commander is "reminded" to also "consult with [his] Chain of Command, ... Service, and DoD guidelines before determining the best course(s) of action." Id . And, as always, he should "[c]onsult with [the] SCCC [the Service Central Coordination Cell]," id . at 63, J.A. 570; accord id . 48, J.A. 555—a service-level "cell of experts created to provide multi-disciplinary (e.g., medical, legal) advice and assistance to commanders with regard to service by transgender Service members," id . at 12, J.A. 519. And this is only a swim test.

A multitude of similar puzzles—e.g., issues concerning showers ("Scenario 11"), sleeping ("Scenario 15"), and social events ("Scenario 16")—would call on the commander to re-consult the handbook, his "Chain of Command, SCCC, Service, and DoD guidelines before determining the best course(s) of action." Id . at 48, J.A. 555; see id . at 60–65, J.A. 567–72. If the commander found time for military issues—after, say, adjusting the "timing" of the shower facilities to avoid any conflicts and ensuring that each "shower stall[ ]" contains "curtains" and "clothing hooks," id . at 60–61, J.A. 567–68—then matters become "more complicated" still, id . at 69, J.A. 576. Consider deployments. Because "[s]ome nations view transgender people as culturally unacceptable and will not recognize [an] individual’s preferred gender," the commander must "[c]onduct a thorough analysis of the" destination country by reviewing "the U.S. State Department’s country specific website" and the "DoD Foreign Clearance Guide"—as well as, of course, "discuss[ing] the situation with [his] chain of command and the SCCC." Id . After that, the commander must also ensure that any "male" who happened to "have female anatomical characteristics" was "screened for pregnancy." Id . at 51, J.A. 558. But there’s a wrinkle: Because the Carter policy provided and plaintiffs demand that a person’s "gender marker in DOD’s personnel database" line up with his or her preferred gender (as opposed to biological sex), see Appellees’ Br. 45, the commander must remember (because it’s not recorded in the system) which of his "male" subordinates "maintained female anatomy," see Carter Memo Implementation Handbook 51, J.A. 558.

Whether these accommodations are "worth the added burdens" or not is "not for this Court to" say. Rostker , 453 U.S. at 81, 101 S.Ct. 2646. It is for the executive.

3. We assess plaintiffs’ empirical attacks on the military’s justifications with extreme deference.

Plaintiffs lob empirical attacks at the military’s justifications. They contend that the justifications are mere ipse dixit , with no "indication" or "data" in the record to support them, and are "contrary" to both "expert" testimony and "the consensus of the medical community." Appellees’ Br. at 42. But when the Supreme Court instructed that we in the judiciary "must be particularly careful not to substitute ... our own evaluation of evidence for a reasonable evaluation by" the political branches, it meant it. Rostker , 453 U.S. at 68, 101 S.Ct. 2646. Given the military’s reasonable evaluations, recounted above, the Court has written off arguments like plaintiffs’ as "quite beside the point." Goldman , 475 U.S. at 509, 106 S.Ct. 1310.

Take the military’s decision, citing unit cohesion issues, to apply different dress regulations to biological males and biological females, without exception. In Goldman , a plaintiff attacked comparable dress standards, arguing, as plaintiffs do here, that the military’s justifications for them were "mere ipse dixit , with no support from actual experience or a scientific study in the record, and [were] contradicted by expert testimony." 475 U.S. at 509, 106 S.Ct. 1310. The Court, however, would have none of it. Even though the regulations would have been subject to heightened scrutiny if a government had sought to apply them in the civilian world, see id . at 506, 106 S.Ct. 1310 (citing Sherbert v. Verner , 374 U.S. 398, 406, 83 S.Ct. 1790, 10 L.Ed.2d 965 (1963) ), the Court simply held that the "desirability of dress regulations in the military [was] decided by the appropriate military officials," who were "under no constitutional mandate to abandon their considered professional judgment," id . at 509, 106 S.Ct. 1310. So too here.

Or consider the military’s decision to presumptively disqualify individuals who have undergone gender transition, citing "scientific uncertainty." DoD Report 32, J.A. 300. The military directly acknowledged plaintiffs’ argument—that the "prevailing judgment of mental health practitioners is that gender dysphoria can be treated with [ ] transition-related care"—but explained that none of the relevant "studies account for the added stress of military life, deployments, and combat." Id . at 24, J.A. 292. I recognize, of course, that a service member who has surgically transitioned to his or her preferred gender does not present many of the anomalies and confusions applicable to social or hormonal transition. But as I’ve already mentioned, the studies leave uncertainty even for civilian life, and the military could reasonably rest on special characteristics of military life. As in Rostker , a court would "palpably exceed[ ] its authority" if it "ignored" this "considered response." 453 U.S. at 81, 101 S.Ct. 2646.

D

Given the "healthy deference" that we owe the political branches "in the area of military affairs," Rostker , 453 U.S. at 66, 101 S.Ct. 2646, plaintiffs look elsewhere to support their claim—namely, to Twitter. As did the district court, they write off the entire Mattis policy (along with the extensive supporting study) as fruit of the poisonous tweet. As plaintiffs and the district court see it, "unusual factors" supposedly surrounding the President’s July 2017 tweet continue to taint all that has happened since (and, evidently, even the Mattis memo of June 30, 2017, which predated the President’s intervention). The "post hoc processes and rationales," they assert, "appear to have been constrained by, and not truly independent from, the President’s initial policy decisions." Appellees’ Br. 26 (quoting Doe II , 315 F.Supp.3d at 497 ). In fact, it’s hard to discern anything especially "unusual" in the process.

1. The executive change of policy was not unusual.

What of the circumstances supposedly surrounding the "President’s initial policy decisions"—the July-August 2017 tweet and memorandum? Appellees’ Br. 26 (quoting Doe II , 315 F.Supp.3d at 497 ). As the district court saw it, the "2017 Presidential directives" represented "an abrupt reversal in policy" announced without "formality" (never defined by the court) and contrary to the conclusions of the military itself. See Doe II , 315 F.Supp.3d at 497 ; see also Doe I , 275 F. Supp. 3d at 212–13. Not exactly.

First , President Trump’s July-August 2017 directives came a month after , and were consistent with , Secretary Mattis’s prior memorandum. Compare Memorandum from Secretary Mattis, to Secretaries of Military Departments (June 30, 2017), J.A. 425 (delaying implementation of Secretary Carter’s policy change in order to have "additional time to evaluate more carefully the impact"), with 2017 Presidential Memorandum , 82 Fed. Reg. at 41,319 (Aug. 25, 2017) (delaying implementation of Secretary Carter’s "policy change" and, among other things, calling for "further study"). What formality was lacking? After Chairman Dunford of the Joint Chiefs made clear that a tweet was an inadequate ground for action, see Memorandum from Chairman Dunford to Chiefs of Military Services (July 27, 2017), J.A. 408, the President issued a formal memorandum, published in the Federal Register. See Nat’l Inst. of Military Justice v. U.S. Dep’t of Defense , 512 F.3d 677, 692 (D.C. Cir. 2008) (Tatel, J., dissenting) (Instructions "published in the Federal Register would be ‘formal.’ ").

Second , even if the "2017 Presidential directives" "abrupt[ly]" reversed the policy of the prior administration and were issued without "formality," Doe II , 315 F.Supp.3d at 497, "there’s nothing unusual about a new" administration "coming to office inclined to favor a different policy direction, ... disagreeing with staff, or cutting through red tape," In re Dep’t of Commerce , ––– U.S. ––––, 139 S.Ct. 16, 17, 202 L.Ed.2d 306 (2018) (Gorsuch, J., concurring in part, dissenting in part); cf. Motor Vehicle Mfrs. Ass’n of U.S., Inc. v. State Farm Mut. Auto. Ins. Co. , 463 U.S. 29, 59, 103 S.Ct. 2856, 77 L.Ed.2d 443 (1983) (Rehnquist, J., concurring in part, dissenting in part) ("A change in administration brought about by the people casting their votes is a perfectly reasonable basis for an executive agency’s reappraisal of the costs and benefits of its programs and regulations."); Elena Kagan, Presidential Administration , 114 Harv. L. Rev. 2245, 2378 (2001) ("[N]ew administrative interpretations following new presidential elections should provide a reason to think deference appropriate rather than the opposite.").

2. The military need not be independent of the President.

Consider next plaintiffs’ contention, accepted by the district court, that "[d]eference to military decisionmaking ... depends on the actual exercise of independent military judgment." Appellees’ Br. 32; accord, e.g., 322 F.Supp.3d at 100–01 ; Doe II , 315 F.Supp.3d at 497 ; Doe I , 275 F. Supp. 3d at 214. Accordingly, plaintiffs and the district court argue, "deference does not apply here" because the Mattis policy did not result from a process that was " ‘truly independent’ " of the " ‘President’s initial policy decisions.’ " Appellees’ Br. 32 (quoting Doe II , 315 F.Supp.3d at 497 ).

Forget the rhetorical disregard of the true sequence of events. The "Constitution itself requires" deference to the military choices of the political branches. Rostker , 453 U.S. at 67, 101 S.Ct. 2646 ; see, e.g., U.S. Const. art. II, § 1, cl. 1. This includes deference to the judgments of military officials—not because of their expertise—but because "the military authorities have been charged by the" constitutionally responsible branches—that is, "the Executive and Legislative Branches"—"with carrying out our Nation’s military policy." Goldman , 475 U.S. at 507–08, 106 S.Ct. 1310.

Plaintiffs’ implication that military policies are suspect unless they are (somehow) "independent" of the views of the Commander in Chief, Appellees’ Br. 32; Doe II , 315 F.Supp.3d at 497, verges on weird. We did, after all, rest our claim to national independence in part on anathematizing George III’s attempts to "render the Military independent of and superior to the Civil Power." The Declaration of Independence para. 14 (U.S. 1776) (emphasis added). And we endeavored in our Constitution to make sure that such independence was never exercised again. See, e.g., U.S. Const. art. II, § 2, cl. 1. That is why "decisions as to the composition, training, equipping, and control of a military force are ... subject always to civilian control of the Legislative and Executive Branches." Gilligan , 413 U.S. at 10, 93 S.Ct. 2440 (second emphasis added). "It is this power of oversight and control of military force by elected representatives and officials which underlies our entire constitutional system." Id . (emphasis added).

In any case, the record belies plaintiffs’ claim that the Mattis policy and the supporting study were "constrained by ... the President’s initial policy decisions." Appellees’ Br. 26 (quoting Doe II , 315 F.Supp.3d at 497 ). Consider the events following Secretary Mattis’s initial order to delay effectiveness of the Carter plan and the President’s later tweet:

The President himself called for a comprehensive study: "The Secretary of Defense ... may advise me at any time, in writing, that a change to [the pre-Carter] policy is warranted." 2017 Presidential Memorandum , 82 Fed. Reg. at 41,319.

Secretary Mattis ordered a comprehensive study: "The Panel and designated support personnel shall bring a comprehensive, holistic, and objective approach to study military service by transgender individuals, focusing on military readiness, lethality, and unit cohesion, with due regard for budgetary constraints and consistent with applicable law." Memorandum from Secretary Mattis to Secretaries of Military Departments 2 (Sept. 14, 2017), J.A. 404.

The Department of Defense understood the panel of experts that Secretary Mattis convened to have conducted a comprehensive study: "To fulfill its mandate, the Panel addressed three questions:" (i) "Should the Department of Defense access transgender individuals?" (ii) "Should the Department allow transgender individuals to transition gender while serving, and if so, what treatment should be authorized?" (iii) "How should the Department address transgender individuals who are currently serving?" DoD Report 18, J.A. 286 (emphasis added).

The Department’s report and recommendations were, in fact, comprehensive—and contrary to the President’s initial policy decisions: Whereas the President’s July-August 2017 directives

called for "generally prohibit[ing] openly transgender individuals" from serving, 2017 Presidential Memorandum , 82 Fed. Reg. at 41,319, the Department "conclude[d] that transgender persons should not be disqualified from service solely on account of their transgender status," DoD Report 19, J.A. 287.

When transmitting the proposed policy to the President, Secretary Mattis confirmed that the review was independent: "I charged the Panel to provide its best military advise, based on increasing the lethality and readiness of America’s armed forces, without regard to any external factors." Mattis Memo 1, J.A. 263.

In the same transmittal, Secretary Mattis asked the President to "revoke" his initial policy directives: "I [ ] respectfully recommend that you revoke your memorandum of August 25, 2017, regarding Military Service by Transgender Individuals, thus allowing me ... to implement appropriate policies concerning military service by transgender persons." Mattis Memo 3, J.A. 265.

The President acknowledged Secretary Mattis’s "independent judgment," and "revoke[d]" his initial policy decisions: "I hereby revoke my memorandum of August 25, 2017, ‘Military Service by Transgender Individuals,’ and any other directives I may have made with respect to military service by transgender individuals. ... The Secretary of Defense ... may exercise [his] authority to implement any appropriate policies concerning military service by transgender individuals." 83 Fed. Reg. at 13,367.

Especially given Secretary Mattis’s initial call for a "review" of "all relevant considerations" a month before the President tweeted anything, see Memorandum from Secretary Mattis to Secretaries of Military Departments (June 30, 2017), J.A. 425—this is hardly a fait accompli by tweet, much less a coup d’état, oxymoronic as that would be in context.

3. Our review of military decision-making is limited.

Plaintiffs and the district court may, to be sure, regard the entire decisionmaking record as a Potemkin village, designed to pull the wool over the eyes of simple-minded observers (including reviewing courts). Of course the plausibility of such a scheme tends to unravel as we try to imagine the dozens of participants—the "Cabinet members and other officials," Hawaii , 138 S.Ct. at 2405 —who would have been needed for its realization.

But apart from implausibility, plaintiffs’ approach rests on a fundamental misunderstanding of the nature of our review. In Trump v. Hawaii , the Supreme Court addressed an executive order that had been preceded by presidential tweets that the dissenters characterized as "expressing animus toward Islam," 138 S.Ct. at 2435 (Sotomayor, J., dissenting); accord id . at 2433 (Breyer, J., dissenting), a claim never directly refuted by the majority. Refusing to see the issue as "whether to denounce the statements," and rejecting the exhortation to "probe the sincerity of the stated justifications for the policy by reference to extrinsic statements," the Court saw its task as being to "consider not only the statements of a particular President, but also the authority of the Presidency itself." 138 S.Ct. at 2418 (majority opinion). Although it might be permissible to "consider plaintiffs’ extrinsic evidence" (in Hawaii , the government conceded this point), the Court made clear that courts must nevertheless "uphold the [challenged] policy so long as it can reasonably be understood to result from a justification independent of unconstitutional grounds." Id . at 2420. And for all the reasons discussed above, the Mattis policy can easily be understood to have arisen out of perfectly legitimate considerations.

This "narrow standard of review ‘has particular force’ " where, as here, the case "overlap[s] with ‘the area of national security.’ " Id . at 2419 (quoting Kerry v. Din , ––– U.S. ––––, 135 S.Ct. 2128, 2140, 192 L.Ed.2d 183 (2015) (Kennedy, J., concurring)). The Hawaii Court offered two justifications for this standard of review—both of which readily apply here. "For one, ‘[j]udicial inquiry into the national-security realm raises concerns for the separation of powers’ by intruding on the President’s constitutional responsibilities[.]" Id . (alteration in original) (quoting Ziglar v. Abbasi , ––– U.S. ––––, 137 S.Ct. 1843, 1861, 198 L.Ed.2d 290 (2017) ); see Gilligan , 413 U.S. at 10–11, 93 S.Ct. 2440. "For another, ‘when it comes to collecting evidence and drawing inferences’ on questions of national security, ‘the lack of competence on the part of the courts is marked.’ " Hawaii , 138 S.Ct. at 2419 (quoting Holder v. Humanitarian Law Project , 561 U.S. 1, 34, 130 S.Ct. 2705, 177 L.Ed.2d 355 (2010) ). For both reasons, our review into matters of "national security is highly constrained." Id . at 2420 ; see also id . at 2420 n.5 (rejecting a "more free-ranging inquiry" in the "national security" context).

This highly deferential approach is evident in the Court’s review—and affirmance—of two facially discriminatory, gender-based classifications that would unquestionably have fallen had any government attempted to apply them in the civilian world. In Schlesinger the Court upheld a statute entitling male and female Navy lieutenants to different periods of tenure, 419 U.S. at 499–500, 95 S.Ct. 572 —that is, a statute "premised solely upon" a "[s]uspect" classification, id . at 511, 95 S.Ct. 572 (Brennan, J., dissenting). Although the dissent focused (at length) on what "in fact" was "behind" Congress’s decision, id . at 520, 95 S.Ct. 572, the majority made quick work of the challenge, looking, instead, to what "Congress may ... quite rationally have believed ," id . at 508, 95 S.Ct. 572 (majority opinion) (emphasis added). This was not a searching inquiry. See id . at 511, 95 S.Ct. 572 (Brennan, J., dissenting) ("[T]he Court goes far to conjure up a legislative purpose which may have underlain the gender-based distinction here attacked."). Rather, the majority’s approach was one grounded in—and thus constrained by—a proper understanding of the separation of powers. See id . at 510–11, 95 S.Ct. 572 (majority opinion) (discussing U.S. Const. art. I, § 8, cls. 12 – 14 ; art. II, § 2, cl. 1 ).

Rostker —upholding a statutory exclusion of females from draft registration—pursued a similarly constrained analysis. As in Schlesinger , the dissenting justices homed in on what they thought "Congress itself [had] concluded." Rostker , 453 U.S. at 83, 101 S.Ct. 2646 (White, J., dissenting); see also, e.g., id . at 106–07, 101 S.Ct. 2646 (Marshall, J., dissenting) (finding "nothing in the Senate Report" to support the majority’s representation of what "Congress believed"). Yet, again, the majority declined to embark on a searching inquiry of congressional motives. Rather, invoking Schlesinger , see 453 U.S. at 70, 101 S.Ct. 2646 (majority opinion) (describing it as "perhaps" the "best instance[ ]" of proper "reconciliation between the deference due Congress" and the Court’s own duty), the Court simply identified conclusions or facts in the record that justified Congress’s policy choice—the whole time scrupulously resisting the urge to substitute its own "evaluation of evidence" for that of Congress, id . at 68, 101 S.Ct. 2646.

In the words of Hawaii , the Court upheld these policies because they could "reasonably be understood to result from a justification independent of unconstitutional grounds," 138 S.Ct. at 2420 —as could the Mattis policy.

E

Confronted with Hawaii , Goldman , Rostker , and Schlesinger , plaintiffs contend that "heightened scrutiny" nevertheless applies. Appellees’ Br. 32; accord, e.g., Doe I , 275 F.Supp.3d at 211 (applying "heightened scrutiny"). This is so, they claim, because the Mattis policy "facially" discriminates against a "suspect" class (i.e., transgender persons). See, e.g., Appellees’ Br. 16. But (1) the Mattis policy is facially neutral, and (2) it would not trigger heightened scrutiny even if it were not.

The Mattis policy, like the Carter policy before it, turns not on transgender status, but on a medical condition called gender dysphoria. See Mattis Memo 2–3, J.A. 264–65; Carter Memo , attachment at 1, J.A. 588. In fact, the Mattis policy expressly provides that "[t]ransgender persons without a history or diagnosis of gender dysphoria, who are otherwise qualified for service, may serve, like all other Service members , in their biological sex." Mattis Memo 3, J.A. 265 (emphasis added).

Plaintiffs, of course, object to the requirement that all must serve in their biological sex. That is central to their claim. See Oral Arg. Tr. 19:12–16 (arguing that the Mattis policy "requires anyone who serves to do so in their biological sex," but that "not living in a person’s biological sex is the defining characteristic of what it means to be transgender"). But the requirement is nevertheless facially neutral; "all" means "all." Transgender or non-transgender; gender dysphoria or non-gender dysphoria ; "all" service members must serve "in their biological sex." Mattis Memo 3, J.A. 265. This can’t be facially discriminatory as to transgender persons; military officials need not know an individual’s transgender status in order to enforce the policy—knowledge of physical characteristics unrelated to gender preference is both necessary and sufficient. Cf. Crandall v. Paralyzed Veterans of Am. , 146 F.3d 894, 897 (D.C. Cir. 1998) (observing that an employer can’t discriminate on the basis of a disability without an actual "awareness of the disability itself").

To be sure, plaintiffs (wrongly) maintain that the biological-based sex standards operate as a complete ban on transgender persons. Panel Judgment *2 (This is "clear error."). But the effect of these standards on transgender persons (O p. 705 n.* (Wilkins, J., concurring)) is no different from that of a regulation barring headgear (and thus yarmulkes) on Orthodox Jews. See Goldman , 475 U.S. at 514, 106 S.Ct. 1310 (Brennan, J., dissenting) ("It sets up an almost absolute bar to the fulfillment of a religious duty."). Even if both policies require "suppressi[on] [of] the characteristic that defines [a person’s] identity," Appellees’ Br. 21—be it "transgender identity," id ., or "religious ... identity," Goldman , 475 U.S. at 517, 106 S.Ct. 1310 (Brennan, J., dissenting)—the magnitude of the impact does nothing to transform a facially neutral policy into a facially discriminatory one, see id . at 510, 106 S.Ct. 1310 (majority opinion) (describing the headgear policy as "reasonabl[e]" and "evenhanded[ ]" "even though [its] effect is to restrict ... [expression] required by [ ] religious beliefs"); id . at 513, 106 S.Ct. 1310 (Stevens, J., concurring) (agreeing that the headgear policy is "neutral, completely objective").

Plaintiffs, nevertheless, maintain that the Mattis policy is invalid because it "hinges on a person’s transgender status; it subjects transgender people—and only transgender people—to a ‘special additional exclusionary rule.’ " Appellees’ Br. 20 (quoting Doe II , 315 F.Supp.3d at 497 ); accord, e.g., Stockman v. Trump , 331 F.Supp.3d 990, 999–1000 (C.D. Cal. 2018), appeal docketed , No. 18-56539 (9th Cir. Nov. 16, 2018), stayed , No. 18A627, ––– U.S. ––––, 139 S.Ct. 950, ––– L.Ed.2d ––––, 2019 WL 271946 (U.S. Jan. 22, 2019), and cert. before judgment denied , No. 18-678, ––– U.S. ––––, 139 S.Ct. 946, ––– L.Ed.2d ––––, 2019 WL 272027 (U.S. Jan. 22, 2019). The argument (it seems) proceeds as follows: Whereas both transgender and "non -transgender people can experience gender dysphoria," Report of the Transgender Military Service Commission 24 n.31, J.A. 771 (emphasis added); accord DoD Report 20 n.57, J.A. 288, the Mattis policy presumptively disqualifies only "[t]ransgender persons with a history or diagnosis of gender dysphoria," Appellees’ Br. 20 (quoting Mattis Memo 2, J.A. 264).

But in reality non -transgender people with gender dysphoria are no better off than their transgender compatriots in terms of rules for accession and retention. And this would be true even if the Mattis policy on gender dysphoria were applied only to transgender persons, as plaintiffs claim, but the record appears to bely. See DoD Report 42, J.A. 310 (describing the Mattis policy’s gender dysphoria disqualification in terms of all "persons"). Non-transgender individuals who develop gender dysphoria do so as a consequence of grievous injuries or diseases, see DoD Report 20 n.57, J.A. 288 (citing Report of the Transgender Military Service Commission 24 n.31, J.A. 771), that are themselves presumptively disqualifying. The only examples in the record involve horrific testicular wounds or mastectomies because of cancer. But, absent waiver, no individual—transgender or not—may join the military with a history of such conditions. See DoD Instruction 6130.03: Medical Standards for Appointment, Enlistment, or Induction into the Military Services 25 (May 6, 2018), J.A. 236 (disqualifying those without testicles); id . at 46, J.A. 257 (same for those with history of malignancy). And if an individual were already in the military when he or she began experiencing gender dysphoria, then—again, regardless of transgender status—that individual could "be retained" so long as he or she did "not require a change of gender and remain[ed] deployable within applicable retention standards." Mattis Memo 2, J.A. 26. There is thus no daylight between the opportunities available to transgender individuals with gender dysphoria and non-transgender individuals with gender dysphoria.

And even if there were a difference—suppose there were, theoretically, a non-transgender person who suffered from gender dysphoria and who was eligible to serve without waiver—plaintiffs’ attack on the Mattis policy would still be untenable. Plaintiffs cannot have their cake and eat it, too: they cannot seek to invalidate the Mattis policy on the basis that it "subjects transgender people—and only transgender people—to a ‘special additional exclusionary rule,’ " Appellees’ Br. 20, while, at the same time, seeking a remedy (a return to the Carter policy, see Second Am. Compl. 20, J.A. 209) that runs afoul of precisely the principle they invoke against the Mattis policy: it too subjects "transgender" people—and only "transgender" people—to a special additional exclusionary rule, see Carter Memo 1, J.A. 585 ("This DTM [Directive-type Memorandum]: Establishes policy ... for the standards for retention, accession ... for transgender personnel[.]"). Cf. United States v. Bethlehem Steel Corp. , 446 F.2d 652, 661 (2d Cir. 1971) (rejecting objection to one remedy where party didn’t object to a different remedy that was "logically ... subject to the same [alleged] defect").Further, the context refutes plaintiffs’ inference of biased intent. To the extent that Secretary Mattis addressed "transgender persons" (as opposed to all persons), this focus seems more a sign of bureaucratic orderliness than animus. Recall that Secretary Carter first commissioned studies regarding service by "transgender Service members" and, ultimately, established a new policy for service by "transgender personnel." See Memorandum from Secretary Carter to Secretaries of Military Departments (July 28, 2015), J.A. 709; Carter Memo , J.A. 585. It is only natural that his successor would follow suit.

In addition, there appear to be genuine bureaucratic imperatives that swayed Secretary Carter as much as Secretary Mattis. For example, it seems quite rational to handle on a case-by-case basis (rather than with a formal policy) the evidently rare instances in which non -transgender persons develop gender dysphoria. The examples in the record seem in their nature to call for case-specific treatment—grievous battlefield injuries (i.e., "genital wounds") and drastic surgeries (i.e., "mastectom[ies] because of breast cancer"). DoD Report 20 n.57, J.A. 288. To handle these on an individualized basis seems more a matter of common sense than animus.

In any case, even if the Mattis policy were facially discriminatory, this fact would not compel a different result. The policy disputed in Rostker , after all, "was not facially neutral," as counsel for plaintiffs conceded. See Oral Arg. Tr. 26:16–20 (conceding that " Rostker wasn’t" facially neutral). Nor is plaintiffs’ comparison to Frontiero v. Richardson , 411 U.S. 677, 93 S.Ct. 1764, 36 L.Ed.2d 583 (1973) (plurality opinion), apt. See Appellees’ Br. 33. As we have explained, the only reason that the Court applied heightened scrutiny to the disputed congressional rules on military fringe benefits was because, unlike here, the challenged policies "never purported to be a congressional judgment on a uniquely military matter." Goldman , 734 F.2d at 1537.

Here, as in Rostker , the Mattis policy’s classifications are "not invidious, but rather realistically reflect[ ] the fact that the sexes are not similarly situated." Rostker , 453 U.S. at 79, 101 S.Ct. 2646. To say, for example, that a biological male (who identifies as female) is similarly situated to a biological female is, to put it courteously, a "gesture[ ] of superficial equality." Id . Whatever the arguments for such a gesture, it is not constitutionally required in the military context. Rostker , 453 U.S. at 79, 101 S.Ct. 2646.

And while some transgender individuals (an exceedingly small percentage) undergo sex-reassignment surgery, which might, in some instances, closely align a person’s biological sex to his preferred gender, the fact is that an anatomical male from birth is simply not the same as an anatomical male from surgery: one has undergone a rare medical intervention; the other has not. The military may legitimately decide that the two are "not similarly situated," Rostker , 453 U.S. at 78, 101 S.Ct. 2646, and the executive branch is under no constitutional obligation to roll the dice, hoping—against its own considered assessment of the evidence, see, e.g., DoD Report 24, J.A. 292; id . at 32, 130 S.Ct. 2705, J.A. 300—that the efficacy of such surgeries will reliably withstand the rigors of combat. Hawaii , 138 S.Ct. at 2421 ("[W]e cannot substitute our own assessment for the Executive’s predictive judgments on" matters of "national security.").

IV

What I have said thus far is enough to reverse the decision of the district court and dissolve the preliminary injunction. Having reached this point, a court of appeals would typically end its review; with the injunction dissolved, the case would then proceed in the district court. (Under my approach, the district court would, of course, be constrained by this court’s analysis of the merits.)

But there are occasions where it is appropriate to proceed further and "definitively decid[e] the merits." U.S. Ass’n of Reptile Keepers, Inc. v. Zinke , 852 F.3d 1131, 1134–35 (D.C. Cir. 2017) (emphasis added). This is one of them. As I said at the outset, see supra Part II.B, "once jurisdiction is properly invoked under [ 28 U.S.C.] § 1292(a)(1)," we may " ‘dispos[e] of all matters appropriately raised by the record, including entry of final judgment.’ " Hartman , 19 F.3d at 1464 (quoting Wagner , 836 F.2d at 585 ); accord, e.g., Munaf , 553 U.S. at 691, 128 S.Ct. 2207 ("[A] reviewing court has the power on appeal from an interlocutory order ‘to examine the merits of the case ... and upon deciding them in favor of the defendant to dismiss the bill.’ " (second alteration in original) (quoting Story , 268 U.S. at 292, 45 S.Ct. 531 )).

The present case implicates sensitive separation of powers concerns—all in the context of military preparedness. See, e.g., Munaf , 553 U.S. at 692, 128 S.Ct. 2207 (resolving the merits and dismissing the case, which had arisen "in the context of ongoing military operations"); see also Youngstown Sheet & Tube Co. v. Sawyer , 343 U.S. 579, 584–85, 72 S.Ct. 863, 96 L.Ed. 1153 (1952) (resolving the merits on appeal from stay of preliminary injunction). And, for the reasons discussed above, the government "is entitled to judgment as a matter of law." Munaf , 553 U.S. at 692, 128 S.Ct. 2207. The district court’s finding that "some factual issues are in dispute," O p. 705 (Wilkins, J., concurring), does not warrant a different conclusion. As explained above, those "facts"—"about the process leading up to the development of the Mattis [policy]," 322 F.Supp.3d at 100 ; see also Doe II , 315 F.Supp.3d at 497 —are irrelevant to the judicial analysis of military personnel policy dictated by Supreme Court authority, see supra Part III.D.3. Nor does any "confusion" concerning the Mattis policy’s "requirements" for service in accordance with the standards applicable to one’s "biological sex," O p. 705 (Wilkins, J., concurring), call for further discovery. It’s hard to imagine how more detail on that subject could have any impact on plaintiffs’ "facial challenge" (Appellees’ Br. 53) to the Mattis policy , especially as the Carter policy , which plaintiffs want the executive branch to "revert to," see Second Am. Compl. 20, J.A. 209, similarly required service in one’s biological sex (at least until a service member’s gender transition was "complete," see Appellees’ Br. 45).

Any further proceedings—including a highly intrusive examination of the President’s mental processes, see, e.g., 322 F.Supp.3d at 101 ; 319 F.Supp.3d at 543 —would thus "be idle, or worse," Orloff , 345 U.S. at 92, 73 S.Ct. 534, see, e.g., Reptile Keepers , 852 F.3d at 1135 ("reach[ing] a definitive judgment ... in order to ‘save the parties the expense of future litigation’ " (quoting Thornburgh v. Am. Coll. of Obstetricians & Gynecologists , 476 U.S. 747, 756–57, 106 S.Ct. 2169, 90 L.Ed.2d 779 (1986) )). The district courts in Rostker and Goldman , to be sure, may have permitted such intrusions into executive decision making, see O p. 706 (Wilkins, J., concurring)—but, of course, those courts were reversed, see Goldman , 475 U.S. at 504, 106 S.Ct. 1310 ("The Court of Appeals ... reversed ... and [we] now affirm."); Rostker , 453 U.S. at 83, 101 S.Ct. 2646 ("The decision of the District Court ... is accordingly Reversed ."). And it does not appear that the Supreme Court considered any evidence uncovered by those trial courts’ explorations to be necessary—or even pertinent—to its disposition of the cases. To the contrary, Goldman dismissed plaintiff’s "expert testimony" as "quite beside the point," 475 U.S. at 509, 106 S.Ct. 1310, and Rostker chastised the district court for "palpably exceed[ing] its authority" in "relying on [such] testimony," 453 U.S. at 81, 101 S.Ct. 2646. Compare, e.g., Goldberg v. Rostker , 509 F.Supp. 586, 600–02 & nn.25–26 (E.D. Pa. 1980) (three-judge court) (citing deposition testimony that 80,000 females inductees could fill noncombat roles "and release men for immediate deployment into combat"), and Rostker , 453 U.S. at 84, 101 S.Ct. 2646 (White, J., dissenting) (similar), with Rostker , 453 U.S. at 81, 101 S.Ct. 2646 (majority opinion) (criticizing the district court for relying on precisely that testimony). I do not see inviting a district court to go down the pointless path of the Goldman and Rostker district courts as a sign of judicial restraint. But cf. O p. 705, 706 (Wilkins, J., concurring).

This is not a matter of privilege, see O p. 705 (Wilkins, J., concurring), or of shielding the government "from all discovery," id . at 706, but of respect for the judiciary’s proper place in our democracy—not to mention common sense: Where, as here, plaintiffs cannot save their claims with any further discovery because the law so clearly forecloses their demands—both on the current record and with any additions that can plausibly be imagined—the court should not bless (or invite) a futile fishing expedition into the executive’s decisionmaking—especially of the intrusive sort contemplated by the district court. The court should say what the law is and be done with it.

Here, we have the benefit of the Supreme Court’s views in Rostker and Goldman —which leave little doubt that further proceedings would do nothing but harm. After all, "judicial inquiries into ... executive motivation represent a substantial intrusion into the workings of [a coordinate] branch[ ] of government." Vill. of Arlington Heights v. Metro. Housing Dev. Corp. , 429 U.S. 252, 268 n.18, 97 S.Ct. 555, 50 L.Ed.2d 450 (1977). And litigation itself "might distract [the executive branch] from the energetic performance of its constitutional duties"—a distraction that we must guard against, as a matter of "paramount necessity." Cheney v. U.S. Dist. Court for D.C. , 542 U.S. 367, 382, 124 S.Ct. 2576, 159 L.Ed.2d 459 (2004). That is why, where, as here, the executive branch has acted "within its appropriate sphere," its action "must be promptly recognized, and ... [the] delay and inconvenience of a prolonged litigation must be avoided by prompt termination of the proceedings in the district court." Id . (brackets omitted; ellipsis in original) (quoting Ex parte Republic of Peru , 318 U.S. 578, 587, 63 S.Ct. 793, 87 L.Ed. 1014 (1943) ).

In this case, the executive branch has reasonably "drawn the line" for military service, Goldman , 475 U.S. at 510, 106 S.Ct. 1310, "balancing" the needs of "servicemen against the needs of the military" itself, Loving , 517 U.S. at 767, 116 S.Ct. 1737. Plaintiffs’ claims, by contrast, are fundamentally flawed in almost every respect. They give short shrift to the findings of a panel of military experts commissioned by the secretary of defense. They never grapple with the fact that the presidential tweet, on which they place so much weight, post-dates —rather than ante-dates—the decision of the secretary to reevaluate the previous administration’s policies. Their theory of the case requires recharacterizing the policy adopted by Secretary Mattis as a "ban" on service by transgender persons, which it is not, and pretending that the military could comply with its existing (partially congressionally mandated) sex-based standards by allowing persons of one biological sex to conform to the rules applicable to the opposite biological sex.

In sum, plaintiffs cannot demonstrate a likelihood of success on the merits. For that reason, I "need not consider the other [preliminary injunction] factors," Greater New Orleans Fair Hous. Action Ctr. v. U.S. Dep’t of Hous. & Urban Dev. , 639 F.3d 1078, 1089 (D.C. Cir. 2011) ; see also Munaf , 553 U.S. at 690, 128 S.Ct. 2207 ("[A] party seeking a preliminary injunction must demonstrate, among other things, ‘a likelihood of success on the merits.’ " (quoting Gonzales v. O Centro EspiritaBeneficente Uniao Do Vegetal , 546 U.S. 418, 428, 126 S.Ct. 1211, 163 L.Ed.2d 1017 (2006) )). Plaintiffs cannot prevail—so, as in Munaf , "the wisest course" is to "terminate the litigation now." 553 U.S. at 692, 128 S.Ct. 2207.

Appendix: Standing and the Nationwide Injunction

Because at least one plaintiff has standing, the rather obvious defects of the others’ standing need not be resolved on this appeal. And because there would be ample ground to vacate the nationwide injunction even if it ran (properly) only in favor of the parties before the court, the same is true of the injunction’s breadth. But because the district court’s use of the narrowness of the group before it as a justification for the breadth of its injunction is so extraordinary, see 344 F.Supp.3d at 24 (claiming that "nationwide injunctions are proper, and sometimes necessary, in circumstances where class certification may be impossible"), and because the plaintiffs’ theories of standing (accepted by the district court) are so flatly in contradiction of binding precedent, a brief discussion of these issues seems worthwhile—especially as the case is set to proceed before the district court. See Air Line Pilots Ass’n v. UAL Corp. , 897 F.2d 1394, 1397 (7th Cir. 1990) (an appellate court’s reversal on one ground does not moot alternative grounds for reversal).

Standing. To begin, a majority of the plaintiffs unquestionably lack standing. Plaintiffs Regan Kibby, Jane Does 2 through 5, and John Doe 1 were diagnosed with gender dysphoria in reliance on the Carter policy, see Doe II , 315 F.Supp.3d at 486 ; Second Am. Compl. ¶¶ 7–21, 27–30, 33–34, J.A. 191–96, and now demand that the military "revert to" the standards established under that policy, see id . at 20, J.A. 209. But Secretary Mattis agreed with them: Pursuant to the reliance exception, these service members "may continue to serve in their preferred gender and receive medically necessary treatment for gender dysphoria." Mattis Memo 2, J.A. 264. This is not in dispute. See, e.g., Doe II , 315 F.Supp.3d at 486 ; Appellees’ Br. 49. What, then, are they doing here? These plaintiffs have already secured agreement to exactly what they want; there is thus nothing left for the court to give them; and their claims must be dismissed.

Plaintiffs respond that they still need judicial intervention because the Mattis policy—even if it were applied only to others—would, nonetheless, stigmatize them as members of "an inherently inferior class of service members." 344 F.Supp.3d at 25 ; see Appellees’ Br. 49–50. But this argument is frivolous. A military policy that does not apply to plaintiffs, but "makes them feel like second-class citizens," does not give rise to a judicially cognizable injury. In re Navy Chaplaincy , 534 F.3d 756, 763 (D.C. Cir. 2008) (Kavanaugh, J.). Period. The district court’s continued insistence that it does is baffling. Compare, e.g., Doe II , 315 F.Supp.3d at 487–88 ("[T]he ‘stigmatic’ aspects of Plaintiffs’ injuries were ... alone sufficient to confer standing[.]"), with Allen v. Wright , 468 U.S. 737, 755, 104 S.Ct. 3315, 82 L.Ed.2d 556 (1984) (Plaintiffs cannot establish "standing to litigate their claims based on the[ir] stigmatizing injury[.]").

As an alternative ground for standing, plaintiffs contend that the Mattis policy—again, even if it were applied only to others—would, nonetheless, "imperil [their] military careers by reducing their ‘opportunities for assignments, promotion, training, and deployment.’ " Appellees’ Br. 50 (quoting Doe II , 315 F.Supp.3d at 488 ). That is so, plaintiffs claim, not because of any particular provision of the policy, but because its mere existence (along with its "lengthy supporting memorandum" and, of course, the President’s tweets) would spur military officers to "place less trust in Plaintiffs." Doe II , 315 F.Supp.3d at 488.

How so? Let’s assume (against reason) that plaintiffs are right—that military officers will so over-zealously implement an implied instruction (i.e., "that Plaintiffs’ service is harmful to the military"), that these obsessive followers of "military hierarchy," id ., will completely overlook an explicit instruction to the contrary (i.e., respect the military’s "commitment to these Service members," DoD Report 43, J.A. 311). How would an injunction against enforcement of the Mattis policy bring plaintiffs relief? Whether or not the military could enforce the Mattis policy has nothing to do with how various unspecified officers (who, by the way, are not parties before the court) would react to the policy’s "lengthy supporting memorandum" or to the President’s tweets. See Doe II , 315 F.Supp.3d at 488. Those allegedly malign directives would still exist: The district court cannot order the Department of Defense to drop all copies of the military’s study down the memory hole; nor can it force the secretary of defense to intone some court-approved truth on the Pentagon’s steps. So, whether plaintiffs will end up with "an unfavorable work detail" or not, id ., enforcement of the Mattis policy will have nothing to do with it; their claims, if any ever arise, will be against not the policy, but the officer who (unlawfully) discriminates.

With no plausible claim to standing, then, the claims of plaintiffs Regan Kibby, Jane Does 2 through 5, and John Doe 1 must be dismissed.

Nationwide Injunction. Plaintiffs demanded that a single district judge "bar" a coordinate branch of government from "enforc[ing]" a challenged "policy" directive—not merely against the ten plaintiffs, but against any individual to whom it might be applied, whether they are before the court or not. Appellees’ Br. 53. This so-called "modern" remedy, 344 F.Supp.3d at 24, is highly unusual, to put it mildly.

Just last term, the Supreme Court reiterated a critical message about "the proper—and properly limited—role of the courts in [our] democratic society." Gill v. Whitford , ––– U.S. ––––, 138 S.Ct. 1916, 1929, 201 L.Ed.2d 313 (2018) (quoting Allen , 468 U.S. at 750, 104 S.Ct. 3315 ). "The Court’s constitutionally prescribed role is to vindicate the individual rights of the people appearing before it." Id . at 1933. Nothing more. "[I]t is not the role of courts, but that of the political branches, to shape the institutions of government in such fashion as to comply with the laws and the Constitution." Lewis v. Casey , 518 U.S. 343, 349, 116 S.Ct. 2174, 135 L.Ed.2d 606 (1996). Thus, the Court "caution[ed]," a plaintiff’s judicial "remedy must be tailored to redress the plaintiff’s particular injury." Gill , 138 S.Ct. at 1934 (emphasis added) (quoting DaimlerChrysler Corp. v. Cuno , 547 U.S. 332, 353, 126 S.Ct. 1854, 164 L.Ed.2d 589 (2006) ).

Such judicial restraint is especially appropriate where, as here, the case concerns the internal operations of the military. "Our constitutional framework ‘requires that the judiciary be as scrupulous not to interfere with legitimate Army matters as the Army must be scrupulous not to intervene in judicial matters.’ " Munaf , 553 U.S. at 700, 128 S.Ct. 2207 (quoting Orloff , 345 U.S. at 94, 73 S.Ct. 534 ). Thus, even if a remedy were appropriate in this case—and it’s not—it should in no event extend beyond the actual plaintiffs. Cf. Dep’t of Defense v. Meinhold , 510 U.S. 939, 939, 114 S.Ct. 374, 126 L.Ed.2d 324 (1993) (staying injunction against military personnel policy to the extent the injunction "grant[ed] relief to persons other than [the plaintiff]"); Schlesinger v. Reservists Comm. to Stop the War , 418 U.S. 208, 222, 94 S.Ct. 2925, 41 L.Ed.2d 706 (1974) (explaining that the "framing of relief no broader than required by the precise facts" of the case is "especially important when the relief sought produces a confrontation with one of the coordinate branches of the Government").

Indeed, when asked at oral argument for any case "in the history of the Republic in which the judiciary has decided military-wide rules about accession and retention," plaintiffs’ counsel offered only Crawford v. Cushman , 531 F.2d 1114 (2d Cir. 1976). Oral Arg. Tr. 28:21–29:14. But although Crawford used casual language suggesting readiness to direct an order "declaring the challenged regulation to be unconstitutional," id . at 1126, 1127, it never issued any such order, and, in any event, the Second Circuit has broadly rejected Crawford ’s entire want of deference to military judgments "in light of the intervening Supreme Court opinion in Rostker [ ]," see Mack v. Rumsfeld , 784 F.2d 438, 439 (2d Cir. 1986) ; see also Goldman , 734 F.2d at 1537–38 (offering similar criticism).

^* My concurring colleague questions the relevance of this fact since Plaintiffs seek a reversion to the Carter Plan, and it, like the Mattis Plan, required transgender servicemembers to serve in accordance with the requirements for their biological sex, Op. at 736–37 (Williams, J., concurring). Fair enough, but Plaintiffs’ complaint also seeks to enjoin "the categorical exclusion of transgender people from military service," J.A. 209, and Plaintiffs clearly argue that the requirement of serving in one’s biological sex has the effect of a categorical exclusion. Ultimately, this is a matter best sorted out by the District Court in the first instance.