The panel has voted to deny the petition for rehearing and to reject the suggestion for a rehearing en banc.

The full court has been advised of the suggestion for en banc rehearing, and a majority of the judges of the court has voted against it. Fed.R.App.P. 35(b).

The petition for rehearing is denied and the suggestion for a rehearing en banc is rejected.

CANBY, Circuit Judge, joined by Circuit Judge NORRIS, dissenting from denial of rehearing en banc:

We have made a grave error in failing to rehear this case en banc. A panel of this court has held that our government may discriminate against homosexuals whenever it is able to put forth a rational basis for doing so. High Tech Gays v. Defense Industrial Security Clearance Office, 895 F.2d 563 (9th Cir.1990). That decision is wrong, and it will have tragic results. The case should have gone en banc because of its sheer importance. It also should have gone en banc because the panel's opinion skews equal protection analysis as ordained by the Supreme Court.

The panel decision upholds the Defense Department's practice of subjecting gay men and lesbians, automatically and as a class, to more burdensome security clearance procedures than it imposes on any other class of workers in private industry who require access to classified materials. The panel holds that the classification of homosexuals lacks the indicia of a suspect or quasi-suspect category, and that accordingly the government need only come forth with a rational basis to sustain its classification. Finally, the panel opinion finds a rational basis for the classification in the KGB's purported targeting of homosexuals, among others, as potential traitors. All of these holdings are infected with constitutional error.

* It is important to understand what this case is not about. The plaintiffs do not contend that the Department of Defense is without power to investigate fully persons seeking a clearance who it has reason to believe may be unstable or unreliable. Nothing in the district court's order, which the panel reversed, precluded the Department from so proceeding on a case-by-case basis. Plaintiffs also do not contend that the Department is improperly denying clearances at the conclusion of its investigations; indeed, the Department grants clearances to most homosexuals. What plaintiffs challenge is the Department's practice of subjecting all homosexuals, automatically and as a class, to expanded and time-consuming security clearance procedures that are required of no other class. Homosexuals are never granted clearance, as many applicants are, after a brief initial investigation; they are always subjected to an expanded investigation, and they are always referred for further review to a separate office charged with adjudicating questionable cases. The district court found that this practice inflicted injury on the plaintiffs, both in the discriminatory treatment itself, and in the loss of employment opportunities because of the excessive delay in obtaining clearances.

The first thing that ought to be clear about the panel's opinion is that it applies the wrong standard of review. The class of "homosexuals" clearly qualifies as a suspect category, triggering strict judicial scrutiny of any governmental discrimination against them. The applicable criteria are properly described but improperly applied by the panel: "To be a 'suspect' or 'quasi-suspect' class, homosexuals must (1) have suffered a history of discrimination; (2) exhibit obvious immutable, or distinguishing characteristics that define them as a discrete group; and (3) show that they are a minority or politically powerless...." 895 F.2d at 573.

The panel agrees that the first criterion is met; homosexuals have suffered a history of discrimination. Id. This point should not be put quickly out of mind, however, for this history of discrimination makes it far more likely that differential treatment is simply a resort to old prejudices. As the district court said, "[l]esbians and gays have been the object of some of the deepest prejudice and hatred in American society." High Tech Gays v. Defense Industrial Security Clearance Office, 668 F. Supp. 1361 at 1369 (N.D.Cal.1987). That fact tends to make discrimination against them all too easy. We should be careful not to endorse that tendency.

With regard to the second criterion, the panel's opinion states:

Homosexuality is not an immutable characteristic; it is behavioral and hence is fundamentally different from traits such as race, gender, or alienage, which define already existing suspect and quasi-suspect classes.

895 F.2d at 573. There are several problems with this conclusion. In the first place, the criterion quoted earlier by the panel required that the class, to be suspect, "exhibit obvious, immutable, or distinguishing characteristics that define them as a discrete group." Id. (emphasis added). The Supreme Court has more than once recited the characteristics of a suspect class without mentioning immutability. See, e.g., City of Cleburne v. Cleburne Living Center, 473 U.S. 432, 440-41, 105 S. Ct. 3249, 3254-55, 87 L. Ed. 2d 313 (1985); Massachusetts Bd. of Retirement v. Murgia, 427 U.S. 307, 313, 96 S. Ct. 2562, 2566-67, 49 L. Ed. 2d 520 (1976); San Antonio School Dist. v. Rodriguez, 411 U.S. 1, at 28, 93 S. Ct. 1278, at 1294, 36 L. Ed. 2d 16 (1973). Aliens, for example, constitute a suspect category, but the condition is not immutable. See Graham v. Richardson, 403 U.S. 365, 371-72, 91 S. Ct. 1848, 1851-52, 29 L. Ed. 2d 534 (1971). The real question is whether discrimination on the basis of the class's distinguishing characteristic amounts to an unfair branding or resort to prejudice, not necessarily whether the characteristic is immutable.

Immutability, of course, does make discrimination more clearly unfair. There is every reason to regard homosexuality as an immutable characteristic for equal protection purposes. It is not enough to say that the category is "behavioral." One can make "behavioral" classes out of persons who go to church on Saturday, persons who speak Spanish, or persons who walk with crutches. The question is, what causes the behavior? Does it arise from the kind of a characteristic that belongs peculiarly to a group that the equal protection clause should specially protect?

Homosexuals are physically attracted to members of their own sex. That is the source of the behavior that we notice about them. Did they choose to be attracted by members of their own sex, rather than by members of the opposite sex? The answer, by the overwhelming weight of respectable authority, is "no." Sexual identity is established at a very early age; it is not a matter of conscious or controllable choice. See Gay Rights Coalition v. Georgetown University, 536 A.2d 1, 34 (D.C.App.1987); A. Bell, M. Weinberg, & F. Hammersmith, Sexual Preference--Its Development in Men and Women 166-67, 190, 211, 222 (1981); L. Tribe, American Constitutional Law 944-45 n. 17 (1978). Can homosexuals change their orientation? Again, from everything we now know, the answer is "no." See id. At least they cannot change it without immense difficulty. As Judge Norris has asked, what would it take to get any one of us to change his or her sexual orientation? See Watkins v. U.S. Army, 875 F.2d 699, 726 (9th Cir.1989) (en banc) (concurring opinion).

For practical and constitutional purposes, then, homosexuality is an immutable characteristic, and the panel's opinion offers nothing to the contrary except its bare conclusion. When the government discriminates against homosexuals, it is discriminating against persons because of what they are, through no choice of their own, and what they are unable to change. Preventing such unfair discrimination is what the equal protection clause is all about.

The panel's opinion also concludes that homosexuals are not politically powerless. 895 F.2d at 574. Its support for this proposition is that one state broadly bars employment discrimination against homosexuals, two other states more narrowly bar discrimination against homosexuals,1 and a few cities bar some types of discrimination. Id. That showing is clearly insufficient to deprive homosexuals of the status of a suspect classification. Compare the situation with that of blacks, who clearly constitute a suspect category for equal protection purposes. Blacks are protected by three federal constitutional amendments, major federal Civil Rights Acts of 1866, 1870, 1871, 1875 (ill-fated though it was), 1957, 1960, 1964, 1965, and 1968, as well as by antidiscrimination laws in 48 of the states. By that comparison, and by absolute standards as well, homosexuals are politically powerless. They are so because of their numbers, which most estimates put at around 10 per cent of the population, and by the fact that many of them keep their status secret to avoid discrimination. That secrecy inhibits organization of homosexuals as a pressure group. Certainly homosexuals as a class wield less political power than blacks, a suspect classification, or women, a quasi-suspect one. One can easily find examples of major political parties' openly tailoring their positions to appeal to black voters, and to female voters. One cannot find comparable examples of appeals to homosexual voters; homosexuals are regarded by the national parties as political pariahs.

Homosexuals, then, are exactly the kind of class that should trigger strict scrutiny when the government discriminates against them. The panel's opinion, by ruling that such discrimination may be justified merely by a rational basis, creates the opportunity for immense abuse. While one may hope that future courts faced with discrimination against homosexuals would apply the rather "active" rational basis review exemplified by City of Cleburne v. Cleburne Living Center, 473 U.S. 432, 439, 105 S. Ct. 3249, 3254, 87 L. Ed. 2d 313 (1985), there is no guarantee that this will happen. There is ample authority defining "rational basis" in such permissive terms that it really means no scrutiny at all. E.g., McGowan v. Maryland, 366 U.S. 420, 426, 81 S. Ct. 1101, 1105, 6 L. Ed. 2d 393 (1961) ("A statutory discrimination will not be set aside if any state of facts reasonably may be conceived to justify it"). Thus the danger is great that an already much-vilified group will be subject to further governmental discrimination on the slightest of justifications. For this reason alone, this case should have been taken en banc, so that a more protective standard of review could have been established for our circuit.

The panel's opinion seems to suggest that even if homosexuals otherwise qualify as a suspect class, they are precluded from being so because of Bowers v. Hardwick, 478 U.S. 186, 106 S. Ct. 2841, 92 L. Ed. 2d 140 (1986). The suggestion is completely unfounded. In Hardwick, the Supreme Court held that the application of state sodomy laws to homosexuals did not violate any fundamental right of privacy protected by the due process clause. The Hardwick opinion pointed out, however, that no equal protection claim had been presented, id. at 196 n. 8, 106 S. Ct. at 2847 n. 8, and Hardwick is clearly not controlling on the suspect class issue.

The panel's opinion also appears to suggest that Hardwick 's conclusion that homosexuals have no fundamental right to engage in sodomy somehow by itself precludes a heightened level of equal protection scrutiny. 895 F.2d at 570-71. But there are two alternative routes to higher levels of scrutiny under the equal protection clause; a higher level of scrutiny is employed if (1) the classification impinges on a fundamental right, or (2) the classification itself is suspect or quasi-suspect. See, e.g., San Antonio Indep. School Dist. v. Rodriguez, 411 U.S. 1, 17, 93 S. Ct. 1278, 1288, 36 L. Ed. 2d 16 (1973). The panel's opinion seems to collapse the two separate routes into one. It appears to justify that approach because, the Department of Defense being federal, equal protection in this case comes through the Due Process Clause of the Fifth Amendment. But that makes no difference; as the panel itself recognizes, the " '[Supreme] Court's approach to Fifth Amendment equal protection claims has always been precisely the same as to equal protection claims under the Fourteenth Amendment.' " 895 F.2d at 571 (quoting Weinberger v. Wiesenfeld, 420 U.S. 636, 638 n. 2, 95 S. Ct. 1225, 1228 n. 2, 43 L. Ed. 2d 514 (1975)). The panel thus errs when it rejects an equal protection claim on due process grounds. See Sunstein, Sexual Orientation and the Constitution: A Note on the Relationship Between Due Process and Equal Protection, 55 U.Chi.L.Rev. 161 (1988).

Under a standard equal protection approach, the absence of a fundamental right does not preclude the possibility that a suspect classification demands heightened scrutiny. The Supreme Court has made it clear that there is no fundamental right to education, for equal protection purposes. San Antonio School Dist., 411 U.S. at 33-37, 93 S.Ct. at 1296-99. Yet racial segregation in federal schools is subject to the same strict scrutiny that applies to state racial segregation. Bolling v. Sharpe, 347 U.S. 497, 74 S. Ct. 693, 98 L. Ed. 884 (1954). The suspect category triggers the highest level of scrutiny regardless of whether the government regulation impinges on a fundamental right.

The panel's opinion draws an additional point from Hardwick, however. It states: "because homosexual conduct can ... be criminalized, homosexuals cannot constitute a suspect or quasi-suspect class entitled to greater than rational basis review." 895 F.2d at 571 (citing Ben-Shalom v. Marsh, 881 F.2d 454, 464-65 (7th Cir.1989); Woodward v. United States, 871 F.2d 1068, 1076 (Fed.Cir.1989); and Padula v. Webster, 822 F.2d 97, 103 (D.C.Cir.1987)). There are at least three problems with this proposition; the first two are important enough but the third is fundamental to the whole subject.

At the simplest level, it is no refutation of a claim of governmental discrimination that the government has the right to criminalize the activity in question. As one of the amici points out, a city would trigger the highest level of scrutiny if it jailed all blacks convicted of speeding, while it only fined whites similarly convicted. Brief of ACLU Foundation in support of rehearing, p. 13. The classification is suspect even though the city is free to criminalize speeding.

It is not a proper answer to this observation to say: "yes, but Hardwick authorizes a state to select homosexual sodomy for prosecution." Hardwick establishes no such proposition. Hardwick involved the application of a state statute that prohibited sodomy, not homosexual sodomy. Hardwick, a homosexual who had been charged with violating the statute, brought an action contending that application of the statute violated his fundamental rights. As the Supreme Court stated, "The issue presented is whether the Federal Constitution confers a fundamental right upon homosexuals to engage in sodomy and hence invalidates the laws of the many States that still make such conduct illegal and have done so for a very long time." Hardwick, 478 U.S. at 190, 106 S.Ct. at 2843 (emphasis added). The Court then held that there was no such fundamental right. It also rejected the argument that homosexuals possessed a fundamental right to engage in sodomy so long as it was practiced in the privacy of a home. "[I]t would be difficult, except by fiat, to limit the claimed right to homosexual conduct while leaving exposed to prosecution adultery, incest, and other sexual crimes even though they are committed in the home." Id. at 195-96, 106 S.Ct. at 2846.

All Hardwick established was that a homosexual had no fundamental right to violate the sodomy laws--laws that presumably apply (as they are written) to others not protected by some marital right. In this respect, the case is analogous to Employment Division v. Smith, --- U.S. ----, 110 S. Ct. 1595, 108 L. Ed. 2d 876 (1990), in which the Supreme Court held that members of the Native American Church had no free exercise right to violate Oregon's law against use of peyote. Certainly that decision does not establish the proposition that governments are now free to legislate against members of the Native American Church, or against religious peyote users as a class, without triggering a heightened level of scrutiny. The equal protection question of a suspect classification cannot be answered by the mere absence of a fundamental right; they are different issues. The equal protection issue can be fully addressed only by examining the class discriminated against, to see whether it bears the traditional indicia of suspectness.

Even if Hardwick is read more expansively, however, it goes no farther than to hold that sodomy committed by homosexuals may be made a crime. But the most that can be said for that fact is that it authorizes a form of discrimination (criminal punishment) for the act that is made a crime. But when the act is used to define a class for the purpose of imposing disabilities not designed as punishment for the commission of the criminally-proscribed act, then there is no logical reason to preclude the examination of the class to see whether it exhibits the characteristics of a suspect category. Particularly is this so when, in half of the states, the act of sodomy has not in fact been made criminal and there is no crime to punish. When the government defines a class, the nature of the class should be examined.

The final difficulty with the panel's proposition drawn from Hardwick is crucial. It assumes that the class of homosexuals is entirely defined by conduct that may be criminalized. The panel's error here is fundamental. In the first place, it is not proper to assume generally that "homosexual conduct ... can be criminalized." 895 F.2d at 571. There are many varieties of conduct that might be characterized as homosexual, from hand-holding to sodomy. Hardwick establishes only that the latter may be criminalized. Yet the Department of Defense invokes its expanded security clearance procedures when there is information indicating "deviant behavior," which it defines to include "homosexuality." DIS Manual 20-1-M at 4-5, quoted by the district court, 668 F. Supp. 1361, 1364. If there is "homosexual activity" within the past 15 years, the expanded security procedure is followed. Nothing in the record suggests that the Department is confining its view of homosexual conduct to sodomy.

Perhaps the panel would regard this discussion as quibbling, however, and would say that sodomy is very commonly the behavior that triggers an expanded investigation. That fact, if true, still does not justify the Department's discrimination against homosexuals. It is an error of massive proportions to define the entire class of homosexuals by sodomy. I will be the first to admit that homosexuals, in sexually expressing their affection for persons of their own sex, frequently engage in sodomy, as do heterosexuals sexually expressing their affection for persons of the opposite sex. Homosexuals and heterosexuals also engage in other affective conduct, criminalized nowhere. But homosexuality, like heterosexuality, is a status. As an amicus points out, one is a homosexual or a heterosexual while playing bridge just as much as while engaging in sexual activity. National Gay Rights Advocates brief p. 16. And the Department of Defense is discriminating against homosexuals for what they are, not what they do. The Department is not trying to send anyone to jail for sodomy. It is not asserting that acts of sodomy endanger national security. It is making the unsupported assumption that homosexuals are more likely to betray their country than other classes of persons, and it is discriminating against them because of that assumption. That is the key to this case.

The panel that decided this case erred fundamentally, then, in not selecting a higher standard of scrutiny. But it also erred in applying the standard of review that it did select. Indeed, the dangers of rational basis review are well illustrated by this decision, for the truth of the matter is that nothing in the record justifies a conclusion that the Defense Department's discrimination is rational. The panel concludes that the Department's discrimination against homosexuals is rationally based because the KGB "targets" homosexuals in attempting to gain access to classified materials. 895 F.2d at 575-78 & n. 13. There are at least two major problems with this conclusion.

The most fundamental reason not to accept the "targeting" rationale is that it helps to perpetuate the wrongful discrimination. The panel refers to congressional hearings in which it was reported that the KGB was interested in a person's being a homosexual " 'since the homosexual frequently is shunted [sic] by society and made to feel like a social outcast. Such a personality may seek to retaliate against a society that has placed him in this unenviable position.' " Id. at 575 (quoting Senate hearings). In other words, if our society treats a group unfairly, then our government is justified in treating that group even more unfairly because the KGB will seek to exploit the "outcast" feelings of that group. Who is going to break the discriminatory cycle if we don't?

The irrationality of the panel's reasoning is pounded home by its conclusion that it is irrelevant that the KGB's policy itself may be irrational. Id. at 578. Thus, even if the KGB's policy never works, gays and lesbians may be subjected to stricter standards of examination for security clearances than other groups, just because the KGB (perhaps indulging in some stereo-typical prejudices of its own) incorrectly assumes homosexuals to be less loyal than other citizens. The Supreme Court has made it clear that bias on the part of our own citizenry cannot justify governmental discrimination. See Cleburne, 473 U.S. at 448, 105 S.Ct. at 3258-59 (in zoning against a home for the mentally retarded, a "city may not avoid the strictures of [the Equal Protection] Clause by deferring to the wishes or objections of some fraction of the body politic"); Palmore v. Sidoti, 466 U.S. 429, 104 S. Ct. 1879, 80 L. Ed. 2d 421 (1984) (existence of prejudice among public not sufficient to justify removing child from custody of white mother who is living with black man). As one of the amici bluntly puts it: "If the government may not adopt the prejudices of its own citizens as a rationale for discrimination, certainly it cannot adopt those of the Soviet Union." ACLU Foundation brief in support of rehearing, p. 7.

There is every reason to believe that any "targeting" of homosexuals by the KGB is irrational, at least insofar as this record reveals. Over 30 years ago, the Navy's own Crittendon Report concluded that "[n]o factual data exist to support the contention that homosexuals are a greater [security] risk than heterosexuals." Crittendon Report, 1957 (Gibson, 1978). There is no more factual support for the contention now than there was then. The panel opinion points to testimony that, of 40 espionage cases examined by a Senate subcommittee, 2 involved homosexuals. 895 F.2d at 575. In one of those cases, it is hard to see the relevance of homosexuality; the individual sold classified information because he needed cash to buy a new car. Federal Government Security Clearance Program; Hearings before the Permanent Subcommittee on Investigations of the Committee on Governmental Affairs, 99th Cong., 1st Sess. 920 (1985). In any event, 2 out of 40 does not justify treating homosexuals differently from other groups. For purposes of equal protection, it is not enough to show that a burden imposed on a group serves some governmental purpose. It must be shown that there is some rational basis for treating that group differently from others. See Cleburne, 473 U.S. at 449-50, 105 S.Ct. at 3259.

In truth, the government's argument based on KGB targeting was an afterthought; the first evidence of such targeting was submitted in connection with the government's Rule 59 motion for reconsideration of the district court's decision. The rationale for distinctive treatment of homosexuals that was offered by the government at the summary judgment stage focused on three professed concerns: (1) homosexuals may be less willing than others to abide by the law; (2) they may be especially subject to coercion or blackmail; and (3) they may be emotionally or mentally unstable. There is little in the record to support any of these concerns as constituting a rational basis for the classification.

The government subjects all homosexuals, wherever located, to expanded security procedures. In half the states, gay and lesbian sexual conduct, including sodomy, is legal. There is certainly no rational basis for assuming that homosexuals in those states, who break no law, are more likely than other citizens to break laws.

Even in the states where sodomy is proscribed by law, those laws are ignored by enforcement authorities. Even Hardwick was never prosecuted. It is irrational to say that frequent violation of a law that has fallen into disuse is reason to believe that such violators will commit the serious crimes of treason or espionage. And again, the issue is whether it is rational to treat homosexuals differently from others. Adultery is illegal in many states; the Department of Defense does not subject the class of adulterers automatically to expanded security procedures.

As for susceptibility to blackmail, the case is not made by the evidence. The government's own evidence indicated that the KGB was not primarily interested in homosexuals because of their susceptibility to blackmail. J. Barron, KGB: The Secret work of Soviet Secret Agents, 207. It is true that evidence was offered that the KGB may consider sexuality a potentially exploitable vulnerability, see 895 F.2d at 576, but that fact does not render homosexuals more susceptible than heterosexuals.

Finally, there is no evidence that homosexuals have any greater degree of problems with their sexuality, causing emotional or mental instability, than have heterosexuals. The American Psychiatric Association and the American Psychological Association have stated that there is no reason to believe that homosexuality impairs judgment, stability, or reliability. See High Tech Gays, 668 F. Supp. at 1374. Of course, individual homosexuals, like individual heterosexuals, may have problems that cause emotional or mental instability. The plaintiffs do not argue that the Department is precluded from identifying those individuals and denying them clearances. Nothing in the district court's injunction would have precluded the Department from so acting. But the Department subjects all homosexuals to expanded security procedures and adjudications when it has not shown that they are any more likely to have emotional or mental problems than members of other groups not treated in the same wholesale manner.

Accordingly, under the active rational basis standard of Cleburne, the Department has failed to show a rational basis for discriminating against homosexuals as a class. It is true, as the government has emphasized, that protecting the nation's secrets is a governmental duty of the highest order. The government, however, finds it unnecessary to subject other classes automatically to expanded security procedures in order to accomplish that goal. It has shown no reason to treat homosexuals differently. As amicus ACLU Foundation points out, the last time we upheld class discrimination in the interest of national security, it took 42 years to remedy that wrong in the law of the circuit. See Hirabayashi v. United States, 828 F.2d 591 (9th Cir.1987).

The failure of the government to show a rational basis for its classification is therefore the final reason why this case should have gone en banc. But even if the government had succeeded in showing a rational basis, the case would still have required en banc rehearing to correct the panel's standard of scrutiny. To leave on the books the rule that the government can discriminate against homosexuals whenever it has a rational basis to do so, is an invitation to tragedy. Homosexuals are hated, quite irrationally, for what they are, what they did not choose to be, and what they cannot easily change. Mainstream society has mistreated them for centuries. If the equal protection clause means anything, it should mean that the government cannot, on the slightest of justifications, join in the discrimination.