dissenting:
Feres v. United States, 340 U.S. 135, 71 S.Ct. 153, 95 L.Ed. 152 (1950), holds that the Federal Torts Claims Act (“FTCA”) waiver of sovereign immunity is inapplicable to suits by uniformed military personnel for injuries arising out of activity “incident to service.” The Feres doctrine is under something of a cloud. See, e.g., United States v. Johnson, 481 U.S. 681, 700, 107 S.Ct. 2063, 2074, 95 L.Ed.2d 648 (1987) (Scalia, J., dissenting) (“Feres was wrongly decided and heartily deserves the ‘widespread, almost universal criticism’ it has received.”). Nonetheless, the Supreme Court, this court and our fellow circuits have all consistently extended Feres beyond the FTCA. That they have done so is unsurprising for damage actions that carry a genuine risk of disrupting the hierarchical command relations among military personnel — after all, that is the risk that ultimately underlies Feres. Comparing Privacy Act damage actions with other claims that the Supreme Court and this court have found barred by Feres, I see neither any greater hint from Congress that Feres should not govern, nor any indication that Privacy Act damage claims pose less risk of interference with command relations. This leads me to the conclusion that Cummings’s claim is precluded.
It might seem at first blush that the Privacy Act’s explicit provisions addressing the military might justify non-application of Feres. The Act purports to govern “any executive department, military department, Government corporation, Government controlled corporation, or other establishment in the executive branch of the government ... or any independent regulatory agency.” 5 U.S.C. § 552(f) (incorporated into the Privacy Act by 5 U.S.C. § 552a(a)(l) & § 552a note (References in Text)). But Congress similarly specified the military in the FTCA. “Federal agency,” the FTCA’s basic unit for liability, see 28 U.S.C. § 2675, is defined to include “the military departments,” id. § 2671.
The Privacy Act also exempts certain military documents from protection under limited circumstances. See, e.g., 5 U.S.C. § 552a(k)(5) (discussing “investigatory material compiled solely for the purpose of determining suitability, eligibility, or qualifications for ... military service”); 5 U.S.C. § 552a(k)(7) (addressing “evaluation material used to determine potential for promotion in the armed services”). But these provisions provide no basis for refusing to apply Feres-, once again, the FTCA has similar language. It exempts claims “arising out of the combatant activities of the military or naval force, or the Coast Guard, during time of war,” 28 *1059U.S.C. § 2680; yet Feres applies to all claims, combat-related or not, wartime or peacetime. See, e.g., United States v. Shearer, 473 U.S. 52, 105 S.Ct. 3039, 87 L.Ed.2d 38 (1985) (applying Feres to claim involving crime off duty and off base).
Indeed, any inferences against Feres drawn from a statute’s explicit coverage and partial exemptions of the military are weaker for Privacy Act damage actions than for the FTCA. The Privacy Act creates a range of specific, non-damage remedies, see 5 U.S.C. § 552a(g)(2) & (3), which the government concedes remain fully applicable. Navy Br. at 25. Cf. Jorden v. National Guard Bureau, 799 F.2d 99, 110-11 (3d Cir.1986) (distinguishing between monetary and injunctive relief in applying Feres to § 1983). But the FTCA provides only damage actions. Thus there is a broad world to which the military-specific language of the Privacy Act can apply, while Feres leaves the FTCA’s similar language with little or no role.
More generally, we have understood Feres to create the equivalent of a “clear statement” rule. In finding it applicable to claims under 42 U.S.C. § 1985(3), we said:
... Feres itself represents a refusal to read statutes with their ordinary sweep. The unique setting of the military led the Feres Court to resist bringing the armed services within the coverage of a remedial statute in the absence of an express Congressional command.
Bois v. Marsh, 801 F.2d 462, 469 n. 13 (D.C.Cir.1986) (emphasis added); see also Coffman v. State of Michigan, 120 F.3d 57, 59 (6th Cir.1997) (holding that ADA and Rehabilitation Act do not apply to the military absent a “clear direction” from Congress (emphasis added)); Roper v. Department of Army, 832 F.2d 247, 248 (2d Cir.1987) (holding that in the “absence of some express indication” from Congress, Title VII does not apply to the military). In the Privacy Act there is no more of an “express Congressional command” than there is in the FTCA.
The Supreme Court has itself extended Feres to all Bivens actions. United States v. Stanley, 483 U.S. 669, 681-84, 107 S.Ct. 3054, 3062-64, 97 L.Ed.2d 550 (1987); Chappell v. Wallace, 462 U.S. 296, 103 S.Ct. 2362, 76 L.Ed.2d 586 (1983). Bivens, of course, imposes liability on individual officials, as indeed does the liability under 42 U.S.C. § 1985(3) that in Bois we found subject to Feres. Thus one might write off both Stanley and Chappell, and Bois, as extensions addressed to some special anxiety that might flow from individual exposure to liability. Compare Maj. Op. at 1056 n.4. That is, of course, conceivable. But it is a rather odd spin on Feres itself, which after all involved only government liability. There is no basis in the Supreme Court opinions for some idea that the discipline issues explaining Feres depend materially on the nominal defendant. That is not surprising, in view of the widespread government provision for reimbursement of officers found liable for acts committed in the service of government. See, e.g., Huang v. Johnson, 251 F.3d 65, 70 (2d Cir.2001) (noting that the “vast majority” of § 1983 claims that the Second Circuit handles involve state indemnification of the individual defendants); see also Board of County Commissioners of Bryan County v. Brown, 520 U.S. 397, 436, 117 S.Ct. 1382, 1404, 137 L.Ed.2d 626 (1997) (Breyer, J., dissenting) (listing various state statutes “authorizing indemnification of employees found liable under § 1983 for actions within the scope of their employment”).
Alternatively, one might write Stanley/Chappell off as simply an exercise of judicial restraint when implying damage actions (such implication is, of course, the only basis for Bivens). See Stanley, 483 U.S. at 679, 107 S.Ct. at 3061. But we *1060have already rejected this as too narrow a view. In Bois, confronting exactly such an argument by the dissenting judge, we said:
But Chappell will not reasonably bear such a crabbed interpretation. To the contrary, the Supreme Court expressly held in Chappell that the same analysis, based on concern with the disruption of military discipline, applies regardless of whether a court is asked to entertain an intramilitary suit under the Constitution or a statute.
801 F.2d at 470 n. 13 (citing explicit passages from Chappell).
Thus we must address the kinds of risks to military discipline that damage actions under the Privacy Act pose. For damage actions under the Act (in contrast with claims for specific relief), the plaintiff must show that the agency action was “intentional or willful.” 5 U.S.C. § 552a(g)(4). Proof that such an intent underlies the alleged breach (here an impermissible release of data) is likely to take the factfinder deep into the breach’s context. Indeed, the complaint here affirmatively asserts that the release arose out of sharply contested views about Cummings’s performance as a trainee on “Hornet” aircraft. See Complaint ¶ ¶ 5-43; see generally Maj. Op. at 1052-53. How does this stack up against the plain vanilla sex discrimination claim involved in Bois? I cannot imagine why it would less involve “second guessing of military decisions by civilian courts,” Bois, 801 F.2d at 470, or less “require testimony by military personnel about command decisions,” id., or less “tend to pit a plaintiffs superiors against one another,” id. Indeed, damage actions under § 552a(g)(4) seem quite commonly to yield evidence of prolonged workplace donnybrooks. See, e.g., Tomasello v. Rubin, 167 F.3d 612, 614-17 (D.C.Cir.1999) (involving government employee’s claim of various unauthorized disclosures in conjunction with an alleged pattern of ethnic and age discrimination); Mount v. U.S. Postal Service, 79 F.3d 531 (6th Cir.1996) (involving employee’s claim that certain disclosures of medical and other records to director and union official were motivated by a desire to retaliate for complaints about work conditions); Henson v. NASA, 14 F.3d 1143, 1146 (6th Cir.1994) (involving employee’s claim that superior released confidential medical information, “intentionally or negligently inflicted severe emotional distress ... [and] engaged in an intentional, reckless, malicious, and tortious pattern of abusive management” as part of a pattern of retaliation). So Bois, I believe, controls.
In sharp contrast is Hunt v. United States, 636 F.2d 580 (D.C.Cir.1980), where we found Feres inapplicable to the “substituted liability” regime established under the Swine Flu Act. Id. at 583 n. 3. To encourage flu vaccine production, the government had assumed responsibility for product liability suits against the manufacturers, id. at 591-92, though assigning the United States a possible right of later indemnification, id. at 595-96. Suits under the Swine Flu Act therefore involved claims of tortious conduct only by the manufacturer, conduct that seems as remote from matters of military discipline and command as one can imagine.
I dissent primarily because the court’s decision seems to me inconsistent with the principles laid down by the Supreme Court in its Feres jurisprudence, and with our circuit’s clear understanding of those principles. But it’s worth noting that other circuits have carried Feres well beyond its FTCA origins, applying it to § 1983 claims,1 to various antidiscrimination stat*1061utes such as Title VII,2 the Americans with Disabilities Act,3 and the Age Discrimination in Employment Act.4 And it has been applied to various admiralty waivers of immunity, such as those in the Suits in Admiralty Act, 46 U.S.C.App. § 742 et seq., and the Public Vessels Act, 46 U.S.CApp. § 781 et seq.5 As the Ninth Circuit has observed, holding that Feres barred a slew of common law, constitutional and even RICO complaints, the Feres “bar has been interpreted broadly,” so that “practically any suit that ‘implicates ... military judgments and decisions’ runs the risk of colliding with Feres.” Bowen, 125 F.3d at 803 (case citations omitted).
Feres is, as I observed earlier, under a cloud. The most plausible solutions seem to be (1) consistent application of its prinei-pie; (2) a rule rather arbitrarily cutting it off with the exact applications already found by the Supreme court and no more; and (3) complete abandonment. As the second and third options are available only to the Supreme Court, Rodriguez de Quijas v. Shearson/American Express, Inc., 490 U.S. 477, 484, 109 S.Ct. 1917, 1922-23, 104 L.Ed.2d 526 (1989); Khan v. State Oil Co., 93 F.3d 1358, 1364 (7th Cir.1996) (Posner, C.J.) (“We have been told by our judicial superiors not to read the sibylline leaves of the U.S. Reports for prophetic clues to overruling.”), I would proceed with the first.
I respectfully dissent.
. E.g., Bowen v. Oistead, 125 F.3d 800, 803 n. 2 (9th Cir.1997); Knutson v. Wisconsin Air National Guard, 995 F.2d 765, 769 (7th Cir.1993); Watson v. Arkansas National Guard, 886 F.2d 1004 (8th Cir.1989); Walden v. Bartlett, 840 F.2d 771, 773 (10th Cir.1988); Craw*1061ford v. Texas Army National Guard, 794 F.2d 1034 (5th Cir.1986); Penagaricano v. Llenza, 747 F.2d 55 (1st Cir.1984).
. E.g., Roper, 832 F.2d at 247-48; Gonzalez v. Department of the Army, 718 F.2d 926, 927-28 (9th Cir.1983); Mackey v. United States, 226 F.3d 773, 776 (6th Cir.2000).
. Baldwin v. United States Army, 223 F.3d 100 (2d Cir.2000); Coffman, 120 F.3d at 57-59.
. Spain v. Ball, 928 F.2d 61 (2d Cir.1991) (dictum).
. See, e.g., Blakey v. U.S.S. Iowa, 991 F.2d 148 (4th Cir.1993); Potts v. United States, 723 F.2d 20 (6th Cir.1983); Charland v. United States, 615 F.2d 508 (9th Cir.1980); Beaucoudray v. United States, 490 F.2d 86 (5th Cir.1974).