I dissent. The United States Court of Appeals for the Ninth Circuit has asked us to answer two related certified questions of California law: (1) May attorney fees, as provided for in Code of Civil Procedure section 1021.5 and Government Code section 12965, subdivision (b), be awarded where the plaintiff has been the “catalyst” in bringing about the relief sought by the litigation? (2) If the catalyst theory is viable under California law, will that theory support an award of attorney fees where the plaintiff “activates” the defendant to modify its behavior; or does California law require a judicially recognized change in the legal relationship between the parties, such as a judgment on the merits, a consent decree, or a judicially ordered settlement?
The certification request (Tipton-Whittingham v. City of Los Angeles (9th Cir. 2003) 316 F.3d 1058, 1060-1061) informs us that plaintiffs filed this action as a class action seeking injunctive relief and damages against the City *611of Los Angeles (City) on behalf of women officers and women civil employees of the Los Angeles Police Department (LAPD). The suit alleges they have been subjected to racial or sexual discrimination, or both. The parties entered into settlement discussions leading to a consent decree that the district court later revoked. The plaintiffs then began new settlement negotiations with the newly appointed Los Angeles Police Chief, Bernard C. Parks. Those talks did not result in any contractual or court-ordered agreement. Instead, the LAPD voluntarily instituted several changes directed toward antidiscrimination. Noting that the changes were similar to the original consent decree, the plaintiffs told the district court that their injunctive relief claims were moot as they had been “resolved informally through negotiations that have not resulted in a formal agreement between the parties, but have resulted in comprehensive change sufficient to moot plaintiffs’ claims.” On the parties’ joint motion, the court dismissed the claims for injunctive relief.
Plaintiffs moved for attorney fees. They claimed they had prevailed on their injunctive relief claims due to the City’s policy changes, and that their efforts had brought about those changes. The court granted the motion and awarded plaintiffs costs and more than $1,703,383 in attorney fees. Later the City moved for reconsideration of the award in light of Buckhannon Board & Care Home, Inc. v. West Virginia Dept. of Health and Human Resources (2001) 532 U.S. 598 [149 L.Ed.2d 855, 121 S.Ct. 1835], The district court granted the motion for reconsideration. It denied attorney fees and costs under federal law but upheld the entire award under California law. The City appealed.
For the reasons explained in my dissent in the companion case of Graham v. DaimlerChrysler Corporation (2004) 34 Cal.4th 553, 584 [21 Cal.Rptr.3d 331], we should reject the catalyst theory. I would answer the certification questions as follows: (1) The catalyst theory alone will not support an award of attorney fees under California law. (2) California law requires a judicially recognized change in the legal relationship between the parties to support an award of attorney fees.
This case differs from Graham in one significant respect that makes the catalyst theory especially pernicious here: The defendant is a governmental entity. Plaintiffs are seeking $1.7 million in attorney fees from the taxpayers of Los Angeles. Obviously, it is not for this court to decide whether a federal court should award attorney fees under a state statute that federal law does not permit and, if so, how to apply the majority’s catalyst jurisprudence to this case. The federal court will have to struggle with these questions. But the majority has certainly opened the door to attorney fee awards of this kind.
Thus, the taxpayers may ultimately bear the cost even though, judging from what the Ninth Circuit has informed us, no court has found that the City *612or any City official ever violated the law. A city can finance a lot of municipal services for $1.7 million. It can pay for law enforcement officers, fire trucks and firefighters, parks, libraries, mental health care, and much more, all of which is actually beneficial to society. A large award like this against a school district could consume tax resources needed for teachers, textbooks, computers, music, art, and sports programs. Such an award could devastate, even bankrupt, a smaller governmental entity. Taxpayers should not have to pay millions of dollars to attorneys who never established that any public official violated the law. We should not endorse this raid on the public fisc.
In this case, the LAPD, under a newly appointed police chief, instituted voluntary changes that triggered a $1.7 million award of attorney fees against the City of Los Angeles. The threat of such large awards of attorney fees will “discourageQ public officials from taking initiatives to revise outmoded ordinances or to improve institutional conditions, because [the catalyst] theory expressly recognizes ‘voluntary actions taken by a defendant’ as a proper basis for a fee award . . . whether or not the court could have ordered that change in conduct. In this way, catalyst theory serves to disable public officials, who may come to fear that worthwhile changes may be retroactively linked to a lawsuit and result in a hefty bill for attorneys’ fees.” (S-l By and Through P-1 v. State Bd. of Educ. (4th Cir. 1993) 6 F.3d 160, 172 (dis. opn. of Wilkinson, J.).)
We expressed a similar concern ourselves in Westside Community for Independent Living, Inc. v. Obledo (1983) 33 Cal.3d 348 [188 Cal.Rptr. 873, 657 P.2d 365], where we overturned an award of attorney fees against a governmental agency predicated on the catalyst theory. In a discussion that argues against the entire catalyst theory, and not just its application in that case, we rejected the argument that attorney fees were appropriate because the lawsuit had caused the public agency to accelerate the process of issuing certain regulations: “[A]warding attorney fees to plaintiffs on the basis of the expedited fiscal study would have detrimental consequences for the public in future lawsuits involving similar causes of action against public agencies. Once an agency was sued, it would refrain from taking any steps that it would normally take to accelerate the promulgation process, for fear that its actions would be perceived by the court as having been induced by the litigation.” (Id. at p. 354, fn. 6.) This concern can be expressed more broadly. Awarding attorney fees on the basis of a governmental agency’s voluntary actions would have detrimental consequences for the public in future lawsuits against governmental agencies. Once sued, an agency may hesitate to take steps that might be good policy, although not legally required, for fear that its actions will expose it to substantial attorney fee awards.
*613We should reject the entire catalyst theory, not extend it to governmental defendants.
Baxter, J., and Brown, J., concurred.