dissenting in part.
There is for me considerable appeal in the majority’s resolution of plaintiffs’ equal protection claims. The governmental conduct these claims challenge involves a patently good faith and facially reasonable effort to accommodate the competing interests of two historie losers in Boston’s housing wars: the racial and ethnic minorities entitled to invoke the protections of the consent decree in NAACP v. Kemp and federal fair housing laws, and the former West Enders, an ethnically diverse, lower middle class group which, in the name of “urban renewal,” was forced from its neighborhood and could not afford to return.
*19Nonetheless, I cannot join the portion of the majority opinion that affirms the district court’s pleadings-based dismissal of the equal protection claims. While I agree with the majority that reverse discrimination claims like the present one are “on the edge of developing law,” ante at 18, I do think it settled that, when the government withdraws benefits from a class of citizens because of the race or ethnicity of the class, courts are to scrutinize strictly the government’s conduct so as to ensure that it furthers a compelling governmental interest and is narrowly tailored to advance that interest. See Adarand Constructors, Inc. v. Pena, 515 U.S. 200, 226-227, 115 S.Ct. 2097, 2112-13, 132 L.Ed.2d 158 (1995). For the reasons that follow, I believe plaintiffs’ complaint fairly alleges such a withdrawal of benefits. And I do not see how we can, at this stage of the litigation, conclude that such a withdrawal of benefits passes strict scrutiny.
The complaint alleges that the governmental defendants curtailed plaintiffs’ statutory preference in order to comply with the consent decree in NAACP v. Kemp. See ante at 14. Because the purpose of that consent decree is “the achievement of a racial composition, in HUD-assisted housing located in neighborhoods which are predominantly white, which reflects the racial composition of the City of Boston as a whole,” id. at 14, a reasonable inference to be drawn from plaintiffs’ complaint, see Aybar v. Crispin-Reyes, 118 F.3d 10, 13 (1st Cir.1997) (reasonable inferences are to be drawn in favor of the party opposing a Fed.R.Civ.P. 12(b)(6) motion), is that defendants acted as they did because the putative plaintiff class was predominantly white. The fact that defendants “do not dispute” this accusation, ante at 16, only underscores our obligation to subject defendants’ conduct to strict scrutiny, see Adarand, 515 U.S. at 224, 115 S.Ct. at 2111 (“[A]ny person, of whatever race, has the right to demand that any governmental actor subject to the Constitution justify any racial classification subjecting that person to unequal treatment under the strictest judicial scrutiny.”).
The majority reaches a contrary conclusion — that strict scrutiny does not apply— because it regards the facts plaintiffs have pleaded in support of their equal protection claims insufficient to describe a “racial classification.” See ante at 16-17 (holding that this ease is outside the principle of Adarand). In reaching its conclusion, the majority emphasizes the effect of curtailing the preference on non-parties to this litigation, see ante at 16 (“[T]he apartments freed from the statutory preference aré made available to all applicants regardless of race.”), and the defendants’ intent in enacting the curtailment, see id. at 16 (“It might not seem remarkable that the government should insist ... that a fair number of the apartments should be effectively open to application by tenants of all races.”). The majority also reads the complaint to allege only that defendants acted as they did because plaintiffs are racially identifiable; it does not read the complaint to allege that defendants acted as they did because plaintiffs are white. Id. at 17.
Taking this last point first, I simply disagree with the majority’s reading of the complaint. The complaint does not allege that the preference was curtailed because plaintiffs are racially monolithic; it alleges that the preference was curtailed because of the consent decree. And, as I have stated, because the consent decree operates only in favor of racial and ethnic minorities, it could not be read to require curtailment of the preference if the former West Enders were predominantly black. Thus, for purposes of evaluating defendants’ Rule 12(b)(6) motion, I believe we must read into the complaint the allegation the majority believes necessary to trigger strict scrutiny, see ante at 17: that defendants would not have acted as they did had the plaintiff class been predominantly of color. See Aybar, 118 F.3d at 13; see also Conley v. Gibson, 355 U.S. 41, 47-48, 78 S.Ct. 99, 102-03, 2 L.Ed.2d 80 (1957) (Fed.R.Civ.P. 8(a)(2) does not require a complaint to set forth specific facts in support of a general allegation of discrimination).
Even if the majority has properly construed the complaint, I believe plaintiffs’ equal protection claims are sufficient to withstand a Rule 12(b)(6) motion and to trigger strict scrutiny. In the redistricting cases, *20the Supreme Court has emphasized that government action which subordinates race neutral considerations to an overriding racial purpose is constitutionally suspect: ‘We recognized in Shaw [v. Reno, 509 U.S. 630, 113 S.Ct. 2816, 125 L.Ed.2d 511 (1993)] that, outside the districting context, statutes are subject to strict scrutiny under the Equal Protection Clause not just when they contain express racial classifications, but also when, though race neutral on their face, they are motivated by a racial purpose or object.” Miller v. Johnson, 515 U.S. 900, 913, 115 S.Ct. 2475, 2486-87, 132 L.Ed.2d 762 (1995) (affirming the invalidation, under equal protection principles, of a Georgia congressional redistricting plan designed to increase the number of majority black districts in Georgia) (citation omitted). It remains to be seen whether the Court will press this principle to its outer limit and strictly scrutinize even governmental conduct which, though predominantly motivated by a racial purpose, would not appear to burden any person because of his or her race — e.g., a public university’s efforts at recruiting fully qualified applicants of color for its first year law school class. But it seems apparent that defendants’ lack of hostility towards whites in particular, cf. ante at 16, does not shield their conduct — which has burdened plaintiffs because they are, as a group, racially identifiable — from the most searching judicial inquiry. See Miller, 515 U.S. at 913, 115 S.Ct. at 2486-87.
This leads to a final point. I think the majority runs afoul of Adarand in concentrating its focus so heavily on both defendants’ intent and the effect of defendants’ actions on non-parties to this case. The Supreme Court has squarely rejected the argument that classifications motivated by “benign” considerations should not be scrutinized strictly. See Adarand, 515 U.S. at 226, 115 S.Ct. at 2112. And though this case does appear unique in that the government conduct at issue is more a withdrawal of a special benefit from whites than a giving of special benefits to members of minority groups,9 the clear import of Adarand is that it is the plaintiffs personal right to equal protection of the laws,” 515 U.S. at 227, 115 S.Ct. at 2112-13, and not some non-party’s interest in competing for that which would be the plaintiffs but for his or her race, that is constitutionally safeguarded. Thus, in evaluating the constitutionality of defendants’ conduct, we must not look to its effect and motivation with respect to others; we must look at its effect and motivation with respect to plaintiffs. And here, quite clearly, defendants’ conduct has had the effect of depriving plaintiffs of a benefit and was prompted by the fact that plaintiffs are mostly white.
Because defendants’ conduct should have been strictly scrutinized, their motion to dismiss plaintiffs’ equal protection claim should have been denied and they should have been required to produce evidence that their conduct was narrowly tailored to advance a compelling governmental interest. See Adarand, 515 U.S. at 227,115 S.Ct. at 2112-13 (reciting the components of the strict scrutiny inquiry); see also Aiken v. City of Memphis, 37 F.3d 1155, 1163 (6th cir.1994) (“When, as here, a race-based affirmative action plan is subjected to strict scrutiny, the party defending the plan bears the burden of producing evidence that the plan is constitutional.”). Plaintiffs should then have been put to the burden of proving the unconstitutionality of defendants’ conduct. See Aiken, 37 F.3d at 1162 (“The party challenging [a racially-preferential] plan retains the ultimate burden of proving its unconstitutionality.”). To the extent that the majority has reached a different conclusion, I most respectfully dissent.
. It is important to note that there is no indication that the benefit originally given (i.e., the full preference) was given to plaintiffs because they were predominantly white; rather, the preference was given because plaintiffs were ousted from their homes.