Opinion
MORENO, J.In this case defendant offered to repurchase a truck that had been marketed with false statements about its towing capacity. This offer came after a lawsuit plaintiffs filed against defendant seeking this repurchase remedy, but before any kind of court judgment was rendered. Plaintiffs were awarded substantial attorney fees under Code of Civil Procedure section 1021.5.1 Defendant raises several issues regarding those fees. The first is whether we should reconsider the catalyst theory, recognized by this court in Westside Community for Independent Living, Inc. v. Obledo (1983) 33 Cal.3d 348 [188 Cal.Rptr. 873, 657 P.2d 365] (Westside Community). Under the catalyst theory, attorney fees may be awarded even when litigation does not result in a judicial resolution if the defendant changes its behavior substantially because of, and in the manner sought by, the litigation. We conclude the catalyst theory should not be abolished but clarified. In order to be eligible *561for attorney fees under section 1021.5, a plaintiff must not only be a catalyst to defendant’s changed behavior, but the lawsuit must have some merit, as discussed below, and the plaintiff must have engaged in a reasonable attempt to settle its dispute with the defendant prior to litigation. Because these limitations on the catalyst theory are to some degree new and were not addressed by the parties or the trial court, we remand for reconsideration of the trial court’s award of attorney fees in this case.
Defendant also contends the trial court erred in concluding that the present lawsuit substantially benefited a large group of people or the general public, as required by section 1021.5. We conclude the trial court did not abuse its discretion in making that conclusion. Finally, defendant, while conceding that a plaintiff could be awarded attorney fees for attorney fee litigation, contends that these fees should not be enhanced beyond the “lodestar” amount. We do not endorse such a categorical rule, but we explain below that fees for fee litigation usually should be enhanced at a significantly lower rate than fees for the underlying litigation, if they are enhanced at all. We therefore will remand the cause to the trial court to recalculate the amount of the fee in light of the principles discussed below, assuming it finds on remand that plaintiffs are eligible for some attorney fees.
I. Statement of Facts
The facts, taken largely from the Court of Appeal’s opinion, are as follows:
DaimlerChrysler incorrectly marketed its 1998 and 1999 Dakota R/T trucks as having a 6,400-pound towing capacity when they could actually tow only 2,000 pounds. The error occurred because the Dakota R/T was a sporty version of an existing truck model, which could tow 6,400 pounds. However, to obtain a sporty design, DaimlerChrysler lowered the suspension on the Dakota R/T, thus reducing its towing capacity.
The reduced towing capacity was a potential risk factor. The lowered suspension meant that towing more than 2,000 pounds would cause the suspension to bottom out, stressing the frame and increasing fatigue and wear. The DaimlerChrysler response team considered this a potential safety issue.
Buyers who wanted to tow more than 2,000 pounds were told they could do so only if their Dakota R/T was modified with a trailer hitch costing $300. The factory installed some of these hitches, while other buyers who wanted to tow had dealer-installed or after-market hitches attached.
Nationwide, DaimlerChrysler sold or leased fewer than 7,000 of the Dakota R/T’s in the two relevant years. Fewer than 1,000 affected R/T’s were sold in California during the two years.
*562By February 1999, DaimlerChrysler set up a response team to address the problem. By June 1999, DaimlerChrysler had taken steps to replace the incorrect marketing materials, owners manuals, and engine and door labels for not-yet-sold Dakota R/T’s, although public agency investigation revealed that brochures misrepresenting the trucks’ towing capacity were still being distributed as of August 1999. DaimlerChrysler also had notified existing buyers of the error, told them not to attempt to tow more than 2,000 pounds, and provided them with the same modified materials. Simultaneously, DaimlerChrysler began to address remedial measures for customers who had bought or leased their Dakota R/T’s under the incorrect marketing program.
Many Dakota R/T buyers never intended to tow more than 2,000 pounds. When informed by DaimlerChrysler of the error, most of those customers were satisfied with DaimlerChrysler’s offers of cash and merchandise.
Initially, DaimlerChrysler offered $300 refunds to buyers who had purchased hitches of that amount. By the summer, DaimlerChrysler authorized dealers to repurchase or replace Dakota R/T’s on a case-by-case basis, but only for customers who demanded such a remedy.
On July 29, 1999, the Santa Cruz County District Attorney contacted DaimlerChrysler about the problem, threatened legal action, and requested DaimlerChrysler’s input before acting. On August 10, 1999, the California Attorney General notified DaimlerChrysler it had joined the Santa Cruz County District Attorney. The public agencies requested a response by the end of August 1999.
Plaintiffs filed their case on August 23, 1999, in Los Angeles County Superior Court. Plaintiffs alleged they all bought 1999 Dakota R/T’s from various DaimlerChrysler dealers. Only Graham lived and bought his truck in California. Plaintiffs alleged DaimlerChrysler marketed, sold, and warranted their 1998 and 1999 Dakota R/T’s as capable of towing 6,400 pounds when the trucks actually could tow only 2,000 pounds. Plaintiffs alleged DaimlerChrysler acknowledged the error by letter to all purchasers dated June 16, 1999. Plaintiffs alleged they notified DaimlerChrysler of their (1) trucks’ failure to comply with the warranted towing capacity, and (2) revocation of their acceptance of their trucks on July 19, 1999. Plaintiffs sought (but never obtained) class certification for all those who bought Dakota R/T’s nationwide. Plaintiffs alleged a single breach of express warranty cause of action. Plaintiffs sought return of their purchase or lease payments, compensatory damages, and attorney fees. Also on August 23, 1999, the Detroit News contacted DaimlerChrysler’s legal counsel about plaintiffs’ case. DaimlerChrysler’s counsel claimed DaimlerChrysler had responded appropriately to the marketing error, including offering buybacks to customers who *563requested it. Plaintiffs faxed their complaint to DaimlerChrysler the same day. The next day, August 24, 1999, DaimlerChrysler’s employee newsletter ran an article on plaintiffs’ case.
DaimlerChrysler’s response team met throughout August 1999. The team knew about both public agency inquiries and the response deadline. Indeed, DaimlerChrysler wrote the public agencies that its internal approval process prohibited a response by August 31, but promised a response by September 8, 1999. On September 10, 1999, DaimlerChrysler issued its offer to all previous Dakota R/T buyers of repurchase or replacement. In response to later inquiries, response team members conceded they were aware of the class action lawsuit filed in California before DaimlerChrysler’s September 10, 1999, letter offering repurchase or replacement to all Dakota R/T buyers.
DaimlerChrysler demurred to the complaint. Plaintiffs filed an amended complaint, acknowledging DaimlerChrysler’s offer of, among other remedies, repurchase or replacement of the trucks for all previous buyers. The trial court sustained the demurrer without leave to amend and dismissed the case, finding it was moot because DaimlerChrysler already had offered all purchasers the relief plaintiffs sought. Meanwhile, the public agencies continued to pursue legal action against DaimlerChrysler, pointing to the fact that the erroneous marketing of the Dakota R/T continued as late as September 1999. In late 2000, DaimlerChrysler settled the public agency investigations by paying a $75,000 fine and agreeing to ensure that the marketing error did not reoccur. Nationwide, 2,549 Dakota R/T buyers opted for repurchase or replacement. Another 3,101 buyers opted for service contracts and parts coupons. The total value of these offers exceeded $15 million. Fewer than 1,000 of the R/T buyers were Californians.
Although plaintiffs’ case was dismissed, the parties continued to litigate plaintiffs’ entitlement to attorney fees. DaimlerChrysler insisted throughout that plaintiffs were not entitled to attorney fees, contending plaintiffs had no effect on DaimlerChrysler’s recognition of the problem and decision to offer all buyers repurchase or replacement. For over a year, there were hotly contested discovery and other motions to clarify the facts described above. The court held a lengthy evidentiary hearing on October 18, 2000. DaimlerChrysler contended that the Dakota R/T response team was not even aware of the litigation until after September 10, 1999, when its repurchase offer was made, a position that the trial court found to lack credibility.
The trial court filed its final order awarding attorney fees on July 6, 2001. The court concluded after its review of the declarations and documentary evidence presented that DaimlerChrysler’s “position that the lawsuit was not a catalyst was largely a transparent fabrication . . . .” It rejected *564DaimlerChrysler’s argument that plaintiffs’ action was unnecessary because of the enforcement action of the Santa Cruz County District Attorney and the California Attorney General. The trial court found that these agencies “had only made an inquiry and had not commenced any proceeding when plaintiffs filed this action. Further [those agencies] were only concerned with DaimlerChrysler’s false advertising materials and never sought any remedies on behalf of the consumers who acquired these vehicles while they were being misrepresented.”
In addition to finding that plaintiffs were the successful party, the trial court found the other requirements of section 1021.5 had been met. It found that the lawsuit “resulted in the enforcement of an important right affecting the public interest, ... the protection and enforcement of consumer rights, including highway safety,” and that “as a result of the lawsuit, thousands of consumers received pecuniary benefits and enhanced safety. Thousands more are likely to benefit from it if DaimlerChrysler and/or other manufacturers are deterred from similar conduct in the future.”
The court also concluded that “DaimlerChrysler should pay plaintiffs attorneys fees in the interest of justice. Plaintiffs’ attorney fees will otherwise go unpaid. Fees cannot be paid out of the benefits conferred upon the consumers because DaimlerChrysler . . . distributed the benefits of [its] offer to the consumers without any discussion with plaintiffs or their attorneys. Justice is served by encouraging lawyers to bring meritorious consumer cases, of which this action is an example.”
The trial court found the lodestar fee amount was $329,620 through the October 18, 2000, hearing, with a multiplier of 2.25 for the fees incurred until the October 18, 2000, hearing, including fees for litigating attorney fees, and applied no multiplier for time thereafter. The court awarded no fees for work after April 23, 2001. The total award was $762,830.
The Court of Appeal affirmed. It observed that the United States Supreme Court had recently rejected the catalyst theory as a basis for attorney fee awards under various federal statutes in 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] (Buckhannon). But the court declined to follow the United States Supreme Court’s lead, noting that the catalyst theory has been long recognized in California. The court also rejected arguments that the litigation was not in the public interest and that it (fid not benefit a substantial number of people. Further, the court concluded that the trial court did not abuse its discretion in awarding fees for seeking fees, and in permitting those fees to be enhanced over the basic lodestar amount. We granted review.
*565II. Discussion
A. Whether the Catalyst Theory Should Be Abolished
An important exception to the American rule that litigants are to bear their own attorney fees is found in section 1021.5.2 As we have stated: “The Legislature adopted section 1021.5 as a codification of the private attorney general doctrine of attorney fees developed in prior judicial decisions. [Citation.] Under this section, the court may award attorney fees to a ‘successful party’ in any action that ‘has resulted in the enforcement of an important right affecting the public interest if: (a) a significant benefit, whether pecuniary or nonpecuniary, has been conferred on the general public or a large class of persons, (b) the necessity and financial burden of private enforcement are such as to make the award appropriate, and (c) such fees should not in the interest of justice be paid out of the recovery, if any.’ . . . [T]he private attorney general doctrine ‘rests upon the recognition that privately initiated lawsuits are often essential to the effectuation of the fundamental public policies embodied in constitutional or statutory provisions, and that, without some mechanism authorizing the award of attorney fees, private actions to enforce such important public policies will as a practical matter frequently be infeasible.’ Thus, the fundamental objective of the doctrine is to encourage suits enforcing important public policies by providing substantial attorney fees to successful litigants in such cases.” (Maria P. v. Riles (1987) 43 Cal.3d 1281, 1288-1289 [240 Cal.Rptr. 872, 743 P.2d 932] (Maria P.).)
In order to effectuate that policy, we have taken a broad, pragmatic view of what constitutes a “successful party.” “Our prior cases uniformly explain that an attorney fee award may be justified even when plaintiff’s legal action does not result in a favorable final judgment. (Westside Community, [supra,] 33 Cal.3d 348, 352 [188 Cal.Rptr. 873, 657 P.2d 365]; *566see also Press v. Lucky Stores, Inc. (1983) 34 Cal.3d 311 [193 Cal.Rptr. 900, 667 P.2d 704] [although their action had become moot, plaintiffs were awarded fees under § 1021.5 because they had achieved the relief they sought through preliminary injunction].) It is also clear that the procedural device by which a plaintiff seeks to enforce an important right is not determinative of his or her entitlement to attorney fees under section 1021.5. (In re Head (1986) 42 Cal.3d 223, 228-229 [228 Cal.Rptr. 184, 721 P.2d 65].) Similarly, a section 1021.5 award is not necessarily barred merely because the plaintiff won the case on a preliminary issue. (Woodland Hills Residents] Assn., Inc. [v. City Council (1979)] 23 Cal.3d [917,] 938 [154 Cal.Rptr. 503, 593 P.2d 200].) In determining whether a plaintiff is a successful party for purposes of section 1021.5, ‘[t]he critical fact is the impact of the action, not the manner of its resolution.’ (Folsom [v. Butte County Assn, of Governments (1982)] 32 Cal.3d [668], 685 [186 Cal.Rptr. 589, 652 P.2d 437] (Folsom).) [f] The trial court in its discretion ‘must realistically assess the litigation and determine, from a practical perspective, whether or not the action served to vindicate an important right so as to justify an attorney fee award’ under section 1021.5. (Woodland Hills Residents Assn., supra, 23 Cal.3d at p. 938.)” (Maria P, supra, 43 Cal.3d at pp. 1290-1291.)
The catalyst theory is an application of the above stated principle that courts look to the practical impact of the public interest litigation in order to determine whether the party was successful, and therefore potentially eligible for attorney fees. We specifically endorsed that theory in Westside Community, supra, 33 Cal.3d 348. The plaintiffs in that case sued to compel the Secretary of the Health and Welfare Agency to issue regulations implementing Government Code section 11135, which bars state-funded programs from engaging in various forms of discrimination. Shortly thereafter, the defendant issued proposed regulations, thereby mooting the case. The plaintiffs filed for private attorney general fees pursuant to section 1021.5. (Westside Community, supra, 33 Cal.3d at pp. 351-352.) The defendants argued that there was no evidence the plaintiffs were successful parties because the litigation had not reached a “final judgment.” (Id. at p. 352.) We rejected that argument and cited with approval Fletcher v. A.J. Industries, Inc. (1968) 266 Cal.App.2d 313, 325 [72 Cal.Rptr. 146] (Fletcher), in which the Court of Appeal upheld an attorney fee award in a shareholder derivative action under the theory that the plaintiffs were successful in conferring a substantial benefit to the corporation, even though the litigation “was resolved through a settlement. The court held that ‘[i]t was not significant that the “benefits” found were achieved by settlement of plaintiffs’ action rather than by final judgment.’ ” (Westside Community, supra, 33 Cal.3d at p. 352, quoting Fletcher, supra, 266 Cal.App.2d at p. 325.)
We further observed that “[n]umerous federal decisions have reached the same conclusion, holding that attorney fees may be proper whenever an *567action results in relief for the plaintiff, whether the relief is obtained through a ‘voluntary’ change in the defendant’s conduct, through a settlement, or otherwise. (See, e.g., Sullivan v. Com. of Pa. Dept. of Labor, etc. (3d Cir. 1981) 663 F.2d 443, 447-450; Robinson v. Kimbrough (5th Cir. 1981) 652 F.2d 458, 465-466; American Constitutional Party v. Munro (9th Cir. 1981) 650 F.2d 184, 187-188.) [f] Thus, an award of attorney fees may be appropriate where ‘plaintiffs’ lawsuit was a catalyst motivating defendants to provide the primary relief sought. . . .’ (Robinson, supra, 652 F.2d at p. 465, italics added.) A plaintiff will be considered a ‘successful party’ where an important right is vindicated ‘by activating defendants to modify their behavior.’ ” (Westside Community, supra, 33 Cal.3d at pp. 352-353.)
Robinson v. Kimbrough, supra, 652 F.2d 458 (Robinson), cited with approval in Westside Community, provides a useful example of the catalyst theory’s application. The plaintiffs filed suit in federal district court alleging that Harris County, Georgia, jury commissioners had compiled jury lists, from which grand and petit juries were summoned, with disproportionately low percentages of Blacks and women, resulting in their underrepresentation on the county’s juries. One month after the complaint was filed, the jury commissioners, in a nonadversarial proceeding that did not involve the plaintiffs, requested a county judge to order them to recompile jury lists to obtain a more representative cross-section. The court order was required because the jury commissioners were not authorized to revise jury lists other than biennially, and the regular revision had been recently completed. The court granted the order and the jury lists were revised. The district court then dismissed the plaintiffs’ suit as failing to raise a substantial constitutional question. The plaintiffs appealed and the court of appeals held that, in light of the revised jury lists and certain statutory changes during the pendency of the appeal, the plaintiffs’ challenge was moot. The plaintiffs therefore never received judicial relief. (652 F.2d at pp. 460-462.)
Nonetheless, the court of appeals in subsequent proceedings affirmed that plaintiffs may be entitled to an award of attorney fees pursuant to the Civil Rights Attorney’s Fees Award Act of 1976 (42 U.S.C. § 1988). Rejecting the argument that the plaintiffs were not a “prevailing party,” the court agreed with other federal appellate courts that recovery of attorney fees under the act “is not dependent upon plaintiffs’ ability to secure formal judicial relief by way of injunction or otherwise. Rather, these opinions have focused upon the type of relief obtained from the defendants as a result of the lawsuit. [Citations.] Common to these decisions is the recognition that plaintiffs may recover attorneys’ fees if their lawsuit is a substantial factor or a significant catalyst in motivating the defendants to end their unconstitutional behavior.” (Robinson, supra, 652 F.2d at pp. 465—466.) The court therefore remanded the case for “the purpose of determining whether plaintiffs’ lawsuit was a *568substantial factor or significant catalyst in bringing about an end to the unconstitutional underrepresentation of blacks and women in the Harris County jury lists.” (Id. at p. 467.)3
The Westside Community court, although endorsing the catalyst theory found in Robinson and other federal cases, nonetheless went on to conclude that no attorney fees were owed in that case because there was no demonstrable causal connection between the lawsuit and the government’s action. (Westside Community, supra, 33 Cal.3d at pp. 353-354.)
We continue to conclude that the catalyst theory, in concept, is sound. The principle upon which the theory is based—that we look to the “impact of the action, not its manner of resolution” (Folsom, supra, 32 Cal.3d at p. 685)—is fully consistent with the purpose of section 1021.5: to financially reward attorneys who successfully prosecute cases in the public interest, and thereby “ ‘prevent worthy claimants from being silenced or stifled because of a lack of legal resources.’ ” (Folsom, supra, 32 Cal.3d at p. 683.) We therefore reaffirm our endorsement of the catalyst theory.
DaimlerChrysler argues that we should reevaluate that endorsement in light of the rejection of the catalyst theory by the United States Supreme Court in Buckhannon, supra, 532 U.S. 598. At the outset we state what hardly needs stating: that United States Supreme Court interpretation of federal statutes does not bind us to similarly interpret similar state statutes. Indeed, in the realm of attorney fees for private attorneys general, this court *569has markedly diverged from United States Supreme Court precedent. In Serrano v. Priest (1977) 20 Cal.3d 25 [141 Cal.Rptr. 315, 569 P.2d 1303] (Serrano III), an opinion that predated the effective date of section 1021.5 (see Serrano v. Unruh (1982) 32 Cal.3d 621, 624, fn. 1 [186 Cal.Rptr. 754, 652 P.2d 985] (Serrano IV)), this court rejected the holding of Alyeska Pipeline Co. v. Wilderness Society (1975) 421 U.S. 240 [44 L.Ed.2d 141, 95 S.Ct. 1612] that attorney fees cannot be awarded on a private attorney general theory absent express statutory authorization. (Serrano III, supra, 20 Cal.3d at pp. 46-47.) More recently, we unanimously declined to follow the United States Supreme Court’s rejection of the use of a contingency fee multiplier in calculating private attorney general fees. (Ketchum v. Moses (2001) 24 Cal.4th 1122, 1137-1139 [104 Cal.Rptr.2d 377, 17 P.3d 735] (Ketchum).) We reaffirmed that the “ ‘fashioning of equitable exceptions’ to the California rule that parties must bear their own costs ‘is a matter within the sole competence of this court.’ ” (Id. at p. 1137.) As explained below, we do not find the reasoning of the five-to-four majority in Buckhannon persuasive, and decline to apply its holding to section 1021.5.
In Buckhannon, the plaintiffs sued alleging that certain state law requirements imposed on their assisted living facility violated the Fair Housing Amendments Act of 1988 (42 U.S.C. § 3601 et seq.) and the Americans with Disabilities Act of 1990 (42 U.S.C. § 12101 et seq.). After the state legislature changed the requirements in a way that vindicated the plaintiffs’ position, the plaintiffs sought attorney fees under the relevant statutes, both of which permit attorney fees for the “prevailing party.” (See 42 U.S.C. § 3613(c)(2) [“[T]he court, in its discretion, may allow the prevailing party ... a reasonable attorney’s fee and costs”] and 42 U.S.C. § 12205 [“[T]he court. . . , in its discretion, may allow the prevailing party ... a reasonable attorney’s fee, including litigation expenses, and costs”].)
A good deal of the Buckhannon court’s reason for rejecting the catalyst theory turns on the definition of “prevailing party.” The Buckhannon majority found the term “prevailing party” to be “a legal term of art,” defined according to Black’s Law Dictionary (7th ed. 1999) at page 1145 as “ ‘[a] party in whose favor a judgment is rendered, regardless of the amount of damages awarded <in certain cases, the court will award attorney’s fees to the prevailing partyx—Also termed successful party.' ” (Buckhannon, supra, 532 U.S. at p. 603.) This definition, together with prior court decisions, led the Buckhannon majority to conclude that a “prevailing party” must be a party that has brought about a “ ‘material alteration of the legal relationship of the parties’ ” (id. at p. 604), which could include both a “judgment[] on the merits,” and a settlement agreement “enforced through a consent decree.” (Ibid.)
*570The Buckhannon majority concluded that “the ‘catalyst theory’ falls on the other side of the line from these examples. It allows an award where there is no judicially sanctioned change in the legal relationship of the parties. Even under a limited form of the ‘catalyst theory,’ a plaintiff could recover attorney’s fees if it established that the ‘complaint had sufficient merit to withstand a motion to dismiss for lack of jurisdiction or failure to state a claim on which relief may be granted.’ [Citation.] This is not the type of legal merit that our prior decisions, based upon plain language and congressional intent, have found necessary. ... A defendant’s voluntary change in conduct, although perhaps accomplishing what the plaintiff sought to achieve by the lawsuit, lacks the necessary judicial imprimatur on the change. Our precedents thus counsel against holding that the term ‘prevailing party’ authorizes an award of attorney’s fees without a corresponding alteration in the legal relationship of the parties.” (Buckhannon, supra, 532 U.S. at p. 605.)
We agree with DaimlerChrysler that the terms “prevailing party” and “successful party,” as used in section 1021.5, are synonymous. (Schmier v. Superior Court (2002) 96 Cal.App.4th 873, 877 [117 Cal.Rptr.2d 497]; Urbaniak v. Newton (1993) 19 Cal.App.4th 1837, 1843, fn. 4 [24 Cal.Rptr.2d 333]; see also Maria P., supra, 43 Cal.3d at pp. 1291-1292 [using the two words interchangeably].) We also agree that in the context of section 1021.5, the term “party” refers to a party to litigation, and therefore precludes an award of attorney fees when no lawsuit has been filed. (See Black’s Law Dict. (4th rev. ed. 1968) p. 1278 [“Party” is a technical term having a precise meaning in legal parlance; it refers to “those by or against whom a suit is brought . . . , the party plaintiff or defendant . . . .”]; see also Flannery v. Prentice (2001) 26 Cal.4th 572, 578 [110 Cal.Rptr.2d 809, 28 P.3d 860] [the word “party” as part of a prevailing party fee statute refers to litigant or litigant’s attorney].) But we are aware of no judicial construction or legislative usage in California that limits the terms “prevailing party” or “successful party” to the meaning found in the most recent edition of Black’s Law Dictionary to the exclusion of other meanings, as DaimlerChrysler, following the Buckhannon majority, argues. (Cf. Sullivan v. Delta Air Lines, Inc. (1997) 15 Cal.4th 288, 302 [63 Cal.Rptr.2d 74, 935 P.2d 781] [the Legislature uses the phrase “ ‘action or proceeding’ . . . virtually as a term of art” “whenever it wishes to refer comprehensively to all the judicial remedies available under our law”].)4
*571We therefore turn to the “usual and ordinary meaning” of the statutory language in order to discern legislative intent. (Lennane v. Franchise Tax Bd. (1994) 9 Cal.4th 263, 268 [36 Cal.Rptr.2d 563, 885 P.2d 976].) The term “successful party,” as ordinarily understood, means the party to litigation that achieves its objectives. We agree with the dissenting opinion in Buckhannon: “In everyday use, ‘prevail’ means ‘gain victory by virtue of strength or superiority: win mastery: triumph.’ Webster’s Third New International Dictionary 1797 (1976). There are undoubtedly situations in which an individual’s goal is to obtain approval of a judge, and in those situations, one cannot ‘prevail’ short of a judge’s formal declaration. In a piano competition or a figure skating contest, for example, the person who prevails is the person declared winner by the judges. However, where the ultimate goal is not an arbiter’s approval, but a favorable alteration of actual circumstances, a formal declaration is not essential. . . . [f] A lawsuit’s ultimate purpose is to achieve actual relief from an opponent. Favorable judgment may be instrumental in gaining that relief. Generally, however, ‘the judicial decree is not the end but the means. At the end of the rainbow lies not a judgment, but some action (or cessation of action) by the defendant. . . .’ [Citation.] On this common understanding, if a party reaches the ‘sought-after destination,’ then the party ‘prevails’ regardless of the ‘route taken.’ [Citation.]” (Buckhannon, supra, 532 U.S. at pp. 633-634 (dis. opn. of Ginsburg, J.).)
This practical definition of prevailing or successful party is consistent with our construction of the meaning of “prevailing party” within the context of Civil Code section 1717, which provides that when a contract specifically provides for attorney fees for one party, fees are to go to the prevailing party “whether he or she is the party specified in the contract or not.” In Santisas v. Goodin (1998) 17 Cal.4th 599 [71 Cal.Rptr.2d 830, 951 P.2d 399], we held that although a defendant who has received the benefit of a voluntary dismissal of an action against it is not necessarily a prevailing party, it may be under some circumstances. In discussing the meaning of the term “prevailing party” when it is undefined by contract, we stated that “a court may base its attorney fees decision on a pragmatic definition of the extent to which *572each party has realized its litigation objectives, whether by judgment, settlement, or otherwise.” (Id. at p. 622, italics added.) If, as is clearly the case, a defendant can be a prevailing or successful party after a plaintiff has voluntarily dismissed the case against it, it is difficult to fathom why a plaintiff cannot be considered a prevailing or successful party when it achieves its litigation objectives by means of defendant’s “voluntary” change in conduct in response to the litigation. When a creditor sues a debtor to collect a debt, and the debtor pays the debt before a judgment is entered against it, the creditor has been a “successful party” by any conventional understanding of that term.
DaimlerChrysler also contends that the catalyst theory must be rejected because section 1021.5 requires that the party achieve success in an “action which has resulted in the enforcement of an important right.” It points to Black’s Law Dictionary’s definition of “enforcement” as “[t]he act or process of compelling compliance with a law, mandate, or command” (Black’s Law Dict., 7th ed., supra, p. 549), and also Merriam-Webster’s Collegiate Dictionary (10th ed. 1998) page 383, which defines “enforce” as to “constrain, compel,” or “to carry out effectively.” But neither definition requires the compulsion or constraint inherent in the term “enforcement” to entail a judicial decision. For example, in Belth v. Garamendi (1991) 232 Cal.App.3d 896 [283 Cal.Rptr. 829], the government initially refused the plaintiff’s California Public Records Act request and complied only after the plaintiff filed a writ of mandate. The Court of Appeal held that attorney fees should be awarded under Government Code section 6259, subdivision (d), which mandates an award to plaintiffs who “prevail in litigation” under the Public Records Act. (Belth, supra, 232 Cal.App.3d at p. 902.) It appears plain that the plaintiff in that case had “enforced” its right of access to public records, compelling the public agency to do what it would not do short of litigation. It would be perverse, and contrary to the basic public interest objectives of section 1021.5, to hold that a plaintiff who obtains a final judgment has “enforced” a right, but not a plaintiff whose litigation position is so strong that it achieves the same result by compelling the defendant to change its conduct rather than face a probable judgment against it.5
*573DaimlerChrysler also makes a number of policy arguments. Like the Buckhannon majority, it argues that “[a] request for attorney’s fees should not result in a second major litigation” (Buckhannon, supra, 532 U.S. at p. 609), and that the catalyst theory would require a complex causal determination. “Among other things, a ‘catalyst theory’ hearing would require analysis of the defendant’s subjective motivations in changing its conduct, an analysis that ‘will likely depend on a highly factbound inquiry and may turn on reasonable inferences from the nature and timing of the defendant’s change in conduct.’ ” (Ibid..) We find persuasive the argument of the Buckhannon dissent that although some time may be expended in fact finding under the catalyst theory, it is at least as likely as not that the catalyst rule “ ‘saves judicial resources,’ [citation] by encouraging ‘plaintiffs to discontinue litigation after receiving through the defendant’s acquiescence the remedy initially sought.’ ” (Buckhannon, supra, 532 U.S. at p. 640 (dis. opn. of Ginsburg, J.).)
Nor are we persuaded that cases decided under a catalyst theory will inevitably give rise to complex and time-consuming litigation over the issue of causality. Case law, as well as our own judicial experience, suggests that catalyst theory cases may be resolved by relatively economical, straightforward inquiries by trial court judges close to and familiar with the litigation. (See, e.g., Southwest Center for Biological Diversity, et al. v. Carroll, supra, 182 F.Supp.2d at pp. 951-952 (opn. of Moreno, J.).) Moreover, the defendant in such cases knows better than anyone why it made the decision that granted the plaintiff the relief sought, and the defendant is in the best position to either concede that the plaintiff was a catalyst or to document why the plaintiff was not. We are unpersuaded that DaimlerChrysler’s inability or unwillingness to do either in the present case, thereby prolonging the litigation, is necessarily attributable to the inherent difficulty of catalyst theory cases.
DaimlerChrysler further argues that overall, the benefits that the catalyst rule is supposed to possess are dwarfed by the harms the rule will engender. It contends the evil to which the catalyst rule is addressed—that meritorious plaintiffs and plaintiffs’ attorneys will be deprived of attorney fees by a favorable settlement—will be a relatively rare occurrence. It quotes the Buckhannon majority that “ ‘[I]t is well settled that a defendant’s voluntary cessation of a challenged practice does not deprive a federal court of its power to determine the legality of the practice’ unless it is ‘absolutely clear *574that the allegedly wrongful behavior could not reasonably be expected to recur.’ ” (Buckhannon, supra, 532 U.S. at p. 609.) On the other hand, DaimlerChrylser argues the catalyst rule could encourage nuisance suits by unscrupulous attorneys hoping to obtain fees without having the merits of their suit adjudicated. It quotes with approval from Justice Scalia’s concurrence in Buckhannon, joined by Justice Thomas: “If the [catalyst theory] sometimes rewards the plaintiff with a phony claim (there is no way of knowing), [its absence] sometimes denies fees to the plaintiff with a solid case whose adversary slinks away on the eve of judgment. But it seems to me the evil of the former far outweighs the evil of the latter. There is all the difference in the world between a rule that denies the extraordinary boon of attorney’s fees to some plaintiffs who are no less ‘deserving’ of them than others who receive them, and a rule that causes the law to be the very instrument of wrong—exacting the payment of attorney’s fees to the extortionist.” (Buckhannon, supra, 532 U.S. at p. 618 (conc. opn. of Scalia, J.).)
We, of course, have no way of quantifying the magnitude of the potential and actual abuses by plaintiffs under a catalyst rule or by defendants under its absence. DaimlerChrysler and the Buckhannon majority’s prediction—that defendants’ change of behavior depriving worthy plaintiffs of attorney fees will be relatively rare—is one we cannot verify. But as plaintiffs argue, what is objectionable about elimination of the catalyst theory is not only that in a given case an attorney will be unjustly deprived of fees, but that attorneys will be deterred from accepting public interest litigation if there is the prospect they will be deprived of such fees after successful litigation. (See Chemerinsky, Closing the Courthouse Doors to Civil Rights Litigants (2003) 5 U.Pa. J.Const.L. 537, 547.) As matters stand now, public interest attorneys often take a considerable risk that they will not be paid at all because they will not prevail in the litigation or because they will be deemed ineligible for fees under section 1021.5, as when the suit is adjudged not to be sufficiently in the public interest. Abolition of the catalyst theory will increase an already considerable risk. As plaintiffs’ attorney succinctly states: “[I]t defies common sense to think attorneys who take meritorious public interest cases with the expectation that they will be compensated if they obtained favorable results for their clients will not be deterred from doing so if the defendant can litigate tenaciously, then avoid paying their fees by voluntarily providing relief before a court order is entered.”6
*575Nor do we believe that avoiding this increased risk of public interest litigation must inevitably come at the expense of rewarding a significant number of extortionate lawsuits. We can adopt sensible limitations on the catalyst theory that discourage the latter without putting a damper on lawsuits that genuinely provide a public benefit. Our starting point in this endeavor is the observation that the Legislature has assigned responsibility for awarding fees under section 1021.5 “not to automatons unable to recognize extortionists, but to judges expected and instructed to exercise ‘discretion.’ ” (Buckhannon, supra, 532 U.S. at p. 640 (dis. opn. of Ginsburg, J.).) These judges are in a good position to make the determination, as one court has expressed it, that the lawsuits have achieved their result “ ‘by threat of victory,’ not ‘by dint of nuisance and threat of expense.’ ” (Buckhannon, supra, 532 U.S. at p. 628 (dis. opn. of Ginsburg, J.), quoting Marbley v. Bane (2d Cir. 1995) 57 F.3d 224, 234—235.) In order to make this determination, the court is to inquire not into a defendant’s subjective belief about the suit but rather to gauge, objectively speaking, whether the lawsuit had merit. (See Tyler v. Corner Const. Corp., Inc. (8th Cir. 1999) 167 F.3d 1202, 1206.) A number of circuits of the United States Court of Appeals, prior to Buckhannon, adopted a version of the catalyst theory that required not only a causal connection between the lawsuit and the relief obtained but also a determination that defendant’s conduct was required by law. (Nadeau v. Helgemoe (1st Cir. 1978) 581 F.2d 275 (Nadeau); see also Powder River Basin Resource Council v. Babbitt (10th Cir. 1995) 54 F.3d 1477, 1486; Zinn v. Shalala, supra, 35 F.3d 273, 274; Sablan v. Dept. of Finance for the Commonwealth of the Northern Mariana Islands (9th Cir. 1988) 856 F.2d 1317 (Sablan); Premachandra v. Mitts (8th Cir. 1984) 727 F.2d 717, 721-722.) Generally speaking, the “required by law” prong was tantamount to a finding that the lawsuit was “not frivolous, unreasonable, or groundless.” (Stivers v. Pierce (9th Cir. 1995) 71 F.3d 732, 752, fn. 9.)
This court has not explicitly adopted the above two-pronged test. (See Wallace v. Consumers Cooperative of Berkeley, Inc. (1985) 170 Cal.App.3d 836, 843-844 [216 Cal.Rptr. 649] [noting that some federal circuits focus only on the causal prong and that this court had not considered the Nadeau test].) We now do so. The trial court must determine that the lawsuit is not “frivolous, unreasonable or groundless” (Stivers v. Pierce, supra, 71 F.3d at p. 752, fn. 9), in other words that its result was achieved “by threat of victory, not by dint of nuisance and threat of expense.” (Marbley v. Bane, supra, 57 F.3d at pp. 234—235.) The determination the trial court must make is not unlike the determination it makes when asked to issue a preliminary injunction, i.e., not a final decision on the merits but a determination at a minimum *576that “ ‘the questions of law or fact are grave and difficult.’ ” (Wilms v. Hand (1951) 101 Cal.App.2d 811, 815 [226 P.2d 728]; 6 Witkin Cal. Procedure (4th ed. 1997) Provisional Remedies, § 357, p. 288.)
Although the catalyst rule is sometimes formulated to permit an award of attorney fees as long as a lawsuit can survive a motion to dismiss or for judgment on the pleadings (see Buckhannon, supra, 532 U.S. at p. 605), we see no reason to limit a trial court’s inquiry regarding the merits of the case to an examination of whether the pleadings state a cause of action. When a lawsuit has been mooted by a defendant’s change in conduct, some development of the factual record is required in order to prevail on a catalyst theory. At the very least, a plaintiff must establish “ ‘the precise factual/legal condition that [it] sought to change or affect’ ” as a prerequisite for establishing the catalytic effect of its lawsuit. (Folsom, supra, 32 Cal.3d at p. 685.) Sometimes this factual background will have been developed in the course of litigation. (See, e.g., DeGidio v. Pung, supra, 920 F.2d 525; Pembroke v. Wood County, Texas, supra, 981 F.2d 225.) When the suit is mooted early in its prosecution (as occurred in the present case), it may generally be established during the attorney fee proceeding by declarations, or, at the discretion of the trial court, by an abbreviated evidentiary hearing. (See Sablan, supra, 856 F.2d at pp. 1322-1323; Pearl, Cal. Attorney Fee Awards (Cont.Ed.Bar 2d ed. 1998 & 2003 supp.) § 14.39, pp. 444-445.) The trial court may review this factual background not only to determine the lawsuit’s catalytic effect but also its merits. Attorney fees should not be awarded for a lawsuit that lacks merit, even if its pleadings would survive a demurrer. We believe that trial courts will be able to conduct an abbreviated but meaningful review of the merits of the litigation designed to screen out nuisance suits without significantly increasing attorney fee litigation costs.7 On the other *577hand, the abolition of the catalyst theory, thereby giving plaintiffs the incentive to prolong the litigation until a judicial determination is made, is not necessarily a recipe for judicial efficiency.
In addition to some scrutiny of the merits, we conclude that another limitation on the catalyst rule proposed by the Attorney General, appearing as amicus curiae, should be adopted by this court. The Attorney General proposes that a plaintiff seeking attorney fees under a catalyst theory must first reasonably attempt to settle the matter short of litigation. (See Grimsley v. Board of Supervisors (1985) 169 Cal.App.3d 960, 966-967 [213 Cal.Rptr. 108].) We believe this requirement is fully consistent with the basic objectives behind section 1021.5 and with one of its explicit requirements— the “necessity ... of private enforcement” of the public interest. Awarding attorney fees for litigation when those rights could have been vindicated by reasonable efforts short of litigation does not advance that objective and encourages lawsuits that are more opportunistic than authentically for the public good. Lengthy prelitigation negotiations are not required, nor is it necessary that the settlement demand be made by counsel, but a plaintiff must at least notify the defendant of its grievances and proposed remedies and give the defendant the opportunity to meet its demands within a reasonable time. (See, e.g., S.D. v. Faulkner, supra, 705 F.Supp. at p. 1363 [letter notifying defendants of plaintiffs’ grievances, plus discussions over two-month period]; see also Garrison v. Board of Directors (1995) 36 Cal.App.4th 1670, 1676 [43 Ca.Rptr.2d 214] [Pub. Resources Code, § 21177, subd. (b) requires California Environmental Quality Act litigants to inform agency of objections before litigation to give agency opportunity to respond].) What constitutes a “reasonable” time will depend on the context.
Applying the catalyst rule, as discussed above, to the present case, the trial court applied the first prong of the rule to conclude that the lawsuit was in fact a substantial causal factor in DaimlerChrysler’s change in policy with respect to its willingness to repurchase or replace the Dakota R/T or to offer consumers substantial discounts. DaimlerChrysler does not contend that the trial court’s ruling on that point is unsupported by substantial evidence. But it is unclear whether the trial court considered the merits of the suit, and the trial court did not consider whether plaintiffs attempted to reasonably settle the matter short of litigation. We therefore remand the matter for a determination of whether plaintiffs are eligible for attorney fees under the catalyst rule as articulated above.8
*578B. Trial Court Did Not Abuse Its Discretion in Finding the Substantial Benefit and Public Interest Prongs of Section 1021.5 Were Met
DaimlerChrysler also contends that the attorney fee award must be overturned in its entirety because it failed to confer “a significant benefit. . . on the general public or large class of persons” as required by section 1021.5. This contention need not detain us long. We will uphold the trial court’s decision to award attorney fees under section 1021.5, unless the court has abused its discretion. (Hewlett v. Squaw Valley Ski Corp. (1997) 54 Cal.App.4th 499, 544 [63 Cal.Rptr.2d 118].) It is well settled that attorney fees under section 1021.5 may be awarded for consumer class action suits benefiting a large number of people. (Beasley v. Wells Fargo Bank (1991) 235 Cal.App.3d 1407, 1417-1418 [1 Cal.Rptr.2d 459] [upholding an award of section 1021.5 attorney fees for class action against bank charging excess credit card fees].) As Beasley recognizes, section 1021.5 requires both a finding of a significant benefit conferred on a substantial number of people and a determination that the “subject matter of the action implicated the public interest.” (Beasley, supra, at p. 1418.)
In the present case, the trial court found that the problem addressed by the lawsuit implicated an issue of public safety, and that the lawsuit benefited thousands of consumers and potentially thousands more by acting as a deterrent to discourage lax responses to known safety hazards. In light of the facts reviewed in the first part of this opinion, we conclude the trial court did not abuse its discretion in finding that the lawsuit met the substantial benefit and public interest requirements of section 1021.5.9
C. Whether There Should Be a Multiplier for Attorney Fees for Litigating Attorney Fees
In the present case, a large percentage of the attorney fees were awarded for litigation to obtain fees under section 1021.5. As noted, the *579lodestar amount calculated by the trial court was $329,620, and that amount was multiplied by an enhancement of 2.25, for a total $762,830. The trial court based the enhancement on “the contingency nature [of the litigation], the delay in payment and the quality of the result.” DaimlerChrysler argues that there should be no enhancement for fees for fee-related litigation, or “fees on fees.” Assuming the trial court concludes on remand that plaintiffs are entitled to some attorney fees, we address for its benefit whether it appropriately awarded enhancements for fees on fees. We conclude that, while fees for attorney fee litigation under section 1021.5 may be enhanced under some circumstances, that enhancement should generally be lower than fees awarded in the underlying litigation.
We first review some general principles regarding the calculation of attorney fees in public interest litigation. As we recently explained, under our decision in Serrano III, “a court assessing attorney fees begins with a touchstone or lodestar figure, based on the ‘careful compilation of the time spent and reasonable hourly compensation of each attorney . . . involved in the presentation of the case.’ [Citation.] We expressly approved the use of prevailing hourly rates as a basis for the lodestar, noting that anchoring the calculation of attorney fees to the lodestar adjustment method ‘ “is the only way of approaching the problem that can claim objectivity, a claim which is obviously vital to the prestige of the bar and the courts.” ’ [Citation.] In referring to ‘reasonable’ compensation, we indicated that trial courts must carefully review attorney documentation of hours expended; ‘padding’ in the form of inefficient or duplicative efforts is not subject to compensation. [Citation.]
“Under Serrano III, the lodestar is the basic fee for comparable legal services in the community; it may be adjusted by the court based on factors including ... (1) the novelty and difficulty of the questions involved, (2) the skill displayed in presenting them, (3) the extent to which the nature of the litigation precluded other employment by the attorneys, (4) the contingent nature of the fee award. [Citation.] The purpose of such adjustment is to fix a fee at the fair market value for the particular action. In effect, the court determines, retrospectively, whether the litigation involved a contingent risk or required extraordinary legal skill justifying augmentation of the unadorned lodestar in order to approximate the fair market rate for such services. The ‘ “experienced trial judge is the best judge of the value of professional services rendered in his court, and while his judgment is of course subject to review, it will not be disturbed unless the appellate court is convinced that it is clearly wrong.” ’ ” (Ketchum, supra, 24 Cal.4th at pp. 1131-1132.)
One of the most common fee enhancers, and one used by the trial court in the present case, is for contingency risk. We reaffirmed the propriety *580of a contingency risk enhancement in Ketchum: “The economic rationale for fee enhancement in contingency cases has been explained as follows: ‘A contingent fee must be higher than a fee for the same legal services paid as they are performed. The contingent fee compensates the lawyer not only for the legal services he renders but for the loan of those services. The implicit interest rate on such a loan is higher because the risk of default (the loss of the case, which cancels the debt of the client to the lawyer) is much higher than that of conventional loans.’ (Posner, Economic Analysis of Law (4th ed. 1992) pp. 534, 567.) ‘A lawyer who both bears the risk of not being paid and provides legal services is not receiving the fair market value of his work if he is paid only for the second of these functions. If he is paid no more, competent counsel will be reluctant to accept fee award cases.’ ” (Ketchum, supra, 24 Cal.4th at pp. 1132-1133.)
Turning to the question of compensation for fee-related litigation, we first note it is well established that plaintiffs and their attorneys may recover attorney fees for fee-related matters. (Serrano IV, supra, 32 Cal.3d at pp. 632-633, 639.) As we stated: “the [private attorney general] doctrine will often be frustrated, sometimes nullified, if awards are diluted or dissipated by lengthy, uncompensated proceedings to fix or defend a rightful fee claim.” (Serrano IV, supra, 32 Cal.3d at p. 632; see also Ketchum, supra, 24 Cal.4th at p. 1141.)
While DaimlerChrysler does not dispute that fees for fee-related litigation may be awarded, it asks this court to hold that there should be no multiplier for fees on fees. It cites to several out-of-state cases that have disallowed such multipliers, principally because fee litigation is tangential to the primary litigation and of less social value. (See City of Birmingham v. Horn (Ala. 2001) 810 So.2d 667, 684 [“While the law clearly allows for a fee award with respect to [fee litigation], we do not consider this time to be vital to the true purpose of the litigation”]; Baksinski v. Northwestern University (1992) 231 Ill.App.3d 7 [595 N.E.2d 1106, 1114, 172 Ill.Dec. 436] [“the fee litigation in this case is not part of the class action litigation, and . . . confers no benefit on the class” and is therefore “not the type of litigation warranting the application of a multiplier”]; see also Indiana Hospital Licensing Council v. Women’s Pavilion of South Bend (Ind.Ct.App. 1985) 486 N.E.2d 1070, 1080.) DaimlerChrysler argues that as a policy matter, enhancements of fees will serve only to encourage a “satellite” litigation of dubious social utility. (See Hensley v. Eckerhart (1983) 461 U.S. 424, 437 [76 L.Ed.2d 40, 103 S.Ct. 1933]; see also id. at p. 442 (conc. opn. of Brennan, J.) [referring to attorney fee litigation as “one of the least socially productive types of litigation imaginable”].)
As plaintiffs point out, our Court of Appeal adopted a contrary position in Downey Cares v. Downey Community Development Com. (1987) 196 *581Cal.App.3d 983 [242 Cal.Rptr. 272] (Downey Cares). In that case, the plaintiffs overturned an amendment to the city’s general plan that had been brought about by a conflict of interest on the part of one of the city council members. The trial court awarded attorney fees under Government Code section 91003 and applied a 1.5 multiplier to the entire lodestar amount, including fees for fee litigation. The Court of Appeal upheld the judgment. After reviewing the justification stated above for awarding fees for fee litigation, it stated: “Considering the numerous factors a trial court might legitimately weigh in determining the multiplier [citation], it is certainly conceivable that some of these factors might apply to the main litigation but not to the fee litigation. For instance, the underlying suit might involve complex issues, lengthy proceedings, and unusual skill, while at the same time the fee related motions might be routine and short. Under such circumstances, a trial court would not abuse its discretion if it chose to distinguish the two categories and apply a different multiplier to each. [Citation.] On the other hand, a trial court would not necessarily abuse its discretion if it chose not to distinguish the two categories but to apply the multiplier to the whole lodestar. For instance, if the contingency of receiving any fee and the long delay in receiving the fee . . . were important to the trial judge’s calculation, they seem equally applicable to the award for fee-related services.” (Downey Cares, supra, 196 Cal.App.3d at pp. 997-998.) Echoing Downey Cares, plaintiffs argue that there is no principled basis for categorically precluding appropriate enhancements for fees for fee litigation.
We noted the holding in Downey Cares in Ketchum, supra, 24 Cal.4th at page 1141, footnote 6. But in Ketchum, we declined to apply the contingency fee enhancement to fees for fee litigation. We reasoned that under the statute authorizing attorney fees at issue in that case, section 425.16, subdivision (c), the fees were mandatory once a party prevailed on the underlying antiSLAPP motion, and there was at that point no contingent risk to the pursuit of attorney fees that would justify an enhancement for the fees on fees. (Ketchum, supra, 24 Cal.4th at pp. 1141-1142.) We had no occasion to decide whether fees for fee litigation should be enhanced under section 1021.5.
In light of the above discussion, we reject DaimlerChrysler’s argument that fees for fee litigation can never be enhanced. Such a rule does not appear in harmony with the principle that the awarding of attorney fees and the calculation of attorney fee enhancements are highly fact-specific matters best left to the discretion of the trial court. (See Ketchum, supra, 24 Cal.4th at pp. 1131-1132.) Although we agree with DaimlerChrysler that the reduction of attorney fee litigation is a desirable objective, it is not clear that a categorical rule barring enhancements for fee litigation will accomplish that objective. It is not clear that the unnecessary prolongation of fee litigation is a significant problem, given that trial courts have the capacity to distinguish between reasonable and unreasonable attorney fee charges and the discretion *582to disallow the latter. Nor is it clear that, if there is such a problem, it is caused mainly by avaricious plaintiffs rather than recalcitrant defendants.
Furthermore, “[w]hen the Legislature has determined that the lodestar adjustment approach is not appropriate, it has expressly so stated. Thus, in 1993, it amended Code of Civil Procedure section 1021.5 to provide that attorney fees awarded to a public entity under the section ‘shall not be increased or decreased by a multiplier based upon extrinsic circumstances, as discussed in [Serrano III, supra,] 20 Cal.3d 25, 49.’ (Stats. 1993, ch. 645, § 2, p. 3747.) Its express restriction on the use of fee enhancements therein ‘can be read as an implicit endorsement of their use in other contexts.’ ” (Ketchum, supra, 24 Cal.4th at p. 1135.) One of those “other contexts” is for fees for fee litigation, as recognized six years prior to the 1993 amendments in Downey Cares.
Nonetheless, building on the discussion quoted above in Ketchum and Downey Cares, we recognize that the enhancement justified for fees in the underlying litigation may differ from the enhancement warranted in the fee litigation, and that a lower enhancement, or no enhancement, may be appropriate in the latter litigation. In fact, a closer examination of the enhancement factors set forth in Serrano III leads to the conclusion that in most cases, the enhancement for the fee litigation should be lower than the enhancement for the underlying litigation, if one is applied at all.
This is especially true of the “results obtained” factor that the trial court relied on in part to justify its multiplier. “The ‘results obtained’ factor can properly be used to enhance a lodestar calculation where an exceptional effort produced an exceptional benefit.” (Thayer v. Wells Fargo Bank (2001) 92 Cal.App.4th 819, 838 [112 Cal.Rptr.2d 284].) While the trial court may have legitimately concluded that the underlying litigation had produced an exceptional benefit for consumers in the present case, the same cannot be said of the fee litigation itself, which simply produced fees to compensate plaintiffs’ attorneys for their efforts. We conclude fees for fee litigation should not be enhanced on that basis.
Moreover, while this factor often takes into account the exceptional skill exhibited by the attorney (Leaolo v. Beneficial Cal. Inc. (2000) 82 Cal.App.4th 19, 50 [97 Cal.Rptr.2d 797]; Edgerton v. State Personnel Bd. (2000) 83 Cal.App.4th 1350, 1363 [100 Cal.Rptr.2d 491]; City of Oakland v. Oakland Raiders (1988) 203 Cal.App.3d 78, 85 [249 Cal.Rptr. 606]), an enhancement on that basis is rarely justified for fee-related litigation. This litigation, as discussed above, is for the most part simpler than litigation on the merits. On the other hand, while attorney fees may not be used to punish defendants (Ketchum, supra, 24 Cal.4th at p. 1141), fees for fee litigation *583may be enhanced when a defendant’s opposition to the fee motion creates extraordinary difficulties. (See, e.g., Edgerton v. State Personnel Bd., supra, 83 Cal.App.4th at p. 1363; Crommie v. P.U.C. (N.D.Cal. 1994) 840 F.Supp. 719, 726, affd. sub nom. Mangold v. P.U.C. (9th Cir. 1995) 67 F.3d 1470 [lodestar enhanced in part by increased difficulty due to defendant’s “excessively vexatious and often unreasonable opposition”].)10
Courts awarding attorney fees under section 1021.5 also may generally differentiate between the contingency risk undertaken during the litigation on the merits and the risk undertaken for litigation on fees. The risk that an attorney takes in the underlying public interest litigation has two components: the risk of not being a “successful party,” i.e., of not prevailing on the merits, and the risk of not estabhshing eligibility for an attorney fee award. (Serrano III, supra, 20 Cal.3d at p. 49.) As discussed, in Ketchum we declined to award a contingency enhancement for fee litigation because under section 425.16, award of the fee was mandatory once a party had prevailed on the underlying motion, and therefore neither of the two risk components was implicated. Generally speaking, by the time of the commencement of fee litigation in section 1021.5 cases, the first and perhaps most substantial component of risk, that of not being a successful party, has been eliminated. What remains is the second component, that plaintiffs may not be able to establish eligibility for fees, i.e., to establish that the litigation confers “a ‘significant benefit’ ... ‘on the general public or a large class of persons’ ” (Beasley v. Wells Fargo Bank, supra, 235 Cal.App.3d at pp. 1417-1418) or that there was the “ ‘necessity and financial burden of private enforcement,’ ” making the award appropriate (Hammond v. Agran (2002) 99 Cal.App.4th 115, 121 [120 Cal.Rptr.2d 646]). Although in the present case, as in other catalyst theory cases, plaintiffs had not established themselves as the successful party at the beginning of the fee litigation, and some enhancement for that risk may be justified, the achievement of their litigation objective before fee litigation would reduce somewhat the uncertainty over their “successful party” status. The fact that the risk of fee litigation is generally less than the risk of litigation on the merits of the suit justifies a lower attorney-fee multiplier for the former, if one is given at all. We do not believe a lower multiplier on fees for less risky fee litigation will deter attorneys from accepting worthwhile public interest cases. (See Ketchum, supra, 24 Cal.4th at pp. 1132-1133, 1141-1142.)
One enhancement factor that would be as applicable for fees on fees as for fees on the merits is a significant delay in the payment of the fees. (See Serrano III, supra, 20 Cal.3d at p. 49.) “Court-awarded fees normally are received long after the legal services are rendered. That delay can present *584cash-flow problems for the attorneys. In any event, payment today for services rendered long in the past deprives the eventual recipient of the value of the use of the money in the meantime, which use, particularly in an inflationary era, is valuable. A percentage adjustment to reflect the delay in receipt of payment therefore may be appropriate.” (Copeland v. Marshall (D.C. Cir. 1980) 205 U.S. App. D.C. 390 [641 F.2d 880, 893].) But this enhancement, which is tantamount to an interest rate, is by itself quite small and may be reduced or eliminated if the lodestar rate is based on the present hourly rate rather than the lesser rate applicable when the services were rendered. (Id. at p. 893, fn. 23; see also Pearl, Cal. Attorney Fee Awards, supra, § 13.10, p. 390.)
In the present case, the trial court made its initial decision regarding the fee multiplier before our decision in Ketchum and then, after further briefing, reduced the multiplier from 3.0 to 2.25, not differentiating between the fees in the underlying litigation and the fees on fees. It appears the court over-enhanced the fees on fees by inappropriately using the “results obtained” factor to arrive at the multiplier. On remand the court should also reexamine its use of the risk factor. While it was not required to explain how it calculated that factor, and we will generally presume the attorney fee award was correct “ ‘ “on matters as to which the record is silent” ’ ” (Ketchum, supra, 24 Cal.4th at p. 1140), it would be appropriate for the trial court to reassess its calculation of a risk enhancement for fees on fees in light of this opinion’s conclusion that the risk multiplier for those fees generally should be lower than for fees in the underlying litigation. The trial court is therefore directed on remand to recalculate the proper multiplier if it concludes that plaintiffs are eligible for some attorney fees.
III. Disposition
The judgment of the Court of Appeal affirming the award of attorney fees in the present case is reversed, and the cause is remanded for proceedings consistent with the views expressed in this opinion.
George, C. J., Kennard, J., and Werdegar, J., concurred.
All statutory references are to this code unless otherwise stated.
In its entirety, section 1021.5 provides: “Upon motion, a court may award attorneys’ fees to a successful party against one or more opposing parties in any action which has resulted in the enforcement of an important right affecting the public interest if: (a) a significant benefit, whether pecuniary or nonpecuniary, has been conferred on the general public or a large class of persons, (b) the necessity and financial burden of private enforcement, or of enforcement by one public entity against another public entity, are such as to make the award appropriate, and (c) such fees should not in the interest of justice be paid out of the recovery, if any. With respect to actions involving public entities, this section applies to allowances against, but not in favor of, public entities, and no claim shall be required to be filed therefor, unless one or more successful parties and one or more opposing parties are public entities, in which case no claim shall be required to be filed therefor under Part 3 (commencing with Section 900) of Division 3.6 of Title 1 of the Government Code.
“Attorneys’ fees awarded to a public entity pursuant to this section shall not be increased or decreased by a multiplier based upon extrinsic circumstances, as discussed in Serrano v. Priest [(1977)] 20 Cal.3d 25, 49 [141 Cal.Rptr. 315, 569 P.2d 1303].”
There have been a number of other federal cases in which attorney fees were granted or found potentially available, despite the absence of a judicial decree altering the legal relationship between the parties. (See, e.g., Baumgartner v. Harrisburg Housing Authority (3d Cir. 1994) 21 F.3d 541 [lawsuit caused housing authority to consult with plaintiffs-tenants regarding newly restrictive residence rules]; DeMier v. Gondles (4th Cir. 1982) 676 F.2d 92 [lawsuit was significant factor behind sheriff’s cessation of blanket strip-search policy]; Pembroke v. Wood County, Texas (5th Cir. 1993) 981 F.2d 225 [fees warranted where lawsuit triggered immediate and extensive improvements to substandard jail conditions]; Citizens Against Tax Waste v. Westerville City School District Board of Education (6th Cir. 1993) 985 F.2d 255 [lawsuit spurred revision of policy regulating speakers at school board meetings]; Zinn v. Shalala (7th Cir. 1994) 35 F.3d 273 [lawsuit spurred defendants to repeal contested Medicaid eligibility rules]; DeGidio v. Pung (8th Cir. 1990) 920 F.2d 525 [lawsuit prompted officials to take corrective action regarding prison tuberculosis epidemic]; Foremaster v. City of St. George (10th Cir. 1989) 882 F.2d 1485 [plaintiff’s suit contributed to City’s decision to discontinue 44-year-old practice of subsidizing Mormon Temple’s electrical bills]; Luethje v. Peavine School District of Adair County (10th Cir. 1989) 872 F.2d 352 [lawsuit prompted school board to amend rule impermissibly curtailing employees’ free speech rights]; Southwest Center for Biological Diversity, et al. v. Carroll (C.D.Cal. 2001) 182 F.Supp.2d 944 [plaintiff environmental groups caused Army Corps of Engineers to conduct biological assessment of dam project pursuant to the Endangered Species Act]; S.D. v. Faulkner (S.D.Ind. 1989) 705 F.Supp. 1361 [lawsuit prompted correctional school officials to remedy allegedly abusive treatment programs].)
Moreover, if DaimlerChrysler is arguing that Black’s Law Dictionary defines “successful party” as a term of art, identical to “prevailing party,” and that the Legislature was aware of the definition, then the definition that should be consulted is not from the most recent edition of the dictionary, but the one current when the Legislature adopted section 1021.5 in 1977. (Stats. 1977, ch. 1197, § 1, p. 3979.) Black’s Law Dictionary, 4th ed., supra, at page 1352, employs a number of alternative definitions of “prevailing party.” “That one of the parties to a suit who successfully prosecutes the action or successfully defends against it, prevailing on the main *571issue, even though not to the extent of his original contention. [Citation.] [f] The one in whose favor the decision or verdict is rendered and judgment entered. [Citations.] The party ultimately prevailing when the matter is finally set at rest. [Citation.] The party prevailing in interest, and not necessarily the prevailing person. [Citation.] To be such does not depend upon the degree of success at different stages of the suit, but whether, at the end of the suit, or other proceeding, the party who has made a claim against the other, has successfully maintained it. [Citation.] Thus, where the court grants defendant a new trial after verdict for plaintiff, defendant is the ‘prevailing party’ on that trial, and entitled to costs, although the plaintiff again gets verdict on retrial. [Citation.]” Only one of the alternate definitions set forth above specifies that a “prevailing party” is the one for whom the verdict or judgment is rendered. The above definitions do not exclude the possibility that a party may be considered to be prevailing “when the matter is finally set at rest” by means other than a judgment or verdict. (Ibid.)
Both sides cite legislative history in support of their position. The legislative history is inconclusive. DaimlerChrysler cites legislative testimony by some of Code of Civil Procedure section 1021.5’s proponents, but we generally will not consider such evidence in determining legislative intent. “Material showing the motive or understanding of an individual legislator, including the bill’s author, his or her staff, or other interested persons, is generally not considered. [Citations.] This is because such materials are generally not evidence of the Legislature’s collective intent.” (Metropolitan Water Dist. v. Imperial Irrigation Dist. (2000) 80 Cal.App.4th 1403, 1426 [96 Cal.Rptr.2d 314].) Plaintiffs cite an earlier version of section 1021.5 that provided for attorney fees only “if judgment is entered in” the plaintiff’s favor. Later the provision was completely rewritten. The new language made no reference to a “judgment” but instead referred to a “prevailing plaintiff.” (Compare Sen. Bill No. 664 *573(1975-1976 Reg. Sess.) as introduced Mar. 31, 1975, § 1, and as amended Sept. 11, 1975, § 2.) Plaintiffs argue that this change shows that no judgment is required for a plaintiff to be considered a prevailing party. The argument does not especially advance plaintiffs’ position. “Unpassed bills, as evidences of legislative intent, have little value.” (Dyna-Med, Inc. v. Fair Employment & Housing Com. (1987) 43 Cal.3d 1379, 1396 [241 Cal.Rptr. 67, 743 P.2d 1323].) This is particularly true when, as here, the bill was rewritten so extensively. Nor do we find other legislative materials cited by the parties, none of which focus on the question at issue, particularly helpful.
The dissent theorizes that under the private attorney general doctrine, plaintiffs already have a “great advantage” over defendants in settlement negotiations because plaintiffs face merely the risk they will not be compensated for attorney fees whereas defendants face the near certainty of incurring such fees. (Dis. opn., post, at p. 597.) This assertion substantially oversimplifies a complex matter. The prospect of attorney fees is only one factor .in determining settlement advantage, and other factors often weigh in defendants’ favor in public interest litigation, including the possession of superior information about their own conduct (see *575Farmer & Pecorino, Issues of Informational Asymmetry in Legal Bargaining, in Dispute Resolution: Bridging the Settlement Gap (Anderson ed., 1996), pp. 79-80), as well as greater material resources.
The dissent states that plaintiffs “have not shown that DaimlerChrysler was legally required to offer a full refund in addition to the steps it had already taken regarding plaintiffs, which included full disclosure, prospective correction, and offers to pay for a hitch that, so far as this lawsuit demonstrates, would have cured all harm.” (Dis. opn., post, at p. 601.) The trial court is obviously not bound by the dissent’s characterization of the facts. Moreover, the dissent appears to interpret our requirement that a lawsuit have merit as further requiring that the settlement or the action taken by defendant to moot the lawsuit must be legally required. But no such scrutiny of settlement terms has ever been required, not even under Buckhannon. For example, Buckhannon acknowledges that a consent decree is a valid basis for awarding private attorney general fees. (Buckhannon, supra, 532 U.S. at p. 604, citing Maher v. Gagne (1980) 448 U.S. 122 [65 L.Ed.2d 653, 100 S.Ct. 2570].) In a consent decree, “[t]he parties waive their right to litigate the issues involved in the case and thus save themselves the time, expense, and inevitable risk of litigation. Naturally, the agreement reached normally embodies a compromise; in exchange for the saving of cost and elimination of risk, the parties each give up something they might have won had they proceeded with the litigation.” (United States v. Armour & Co. (1971) 402 U.S. 673, 681 [29 L.Ed.2d 256, 91 S.Ct. 1752].) In other words, a consent decree gives effect to a compromise that is not necessarily required by law. So, too, with regard to a settlement that does not result in a consent decree, or unilateral action that *577moots pending litigation, it is not necessary to determine that the precise remedy chosen was required by law in order for a plaintiff to be eligible for attorney fees under section 1021.5. Rather, a plaintiff’s suit must have merit, as that term is defined above.
The dissent contends that plaintiffs have “failed to satisfy” these two latter prongs of the catalyst rule. (Dis. opn., post, at p. 600.) The dissent does not point to any trial court finding *578indicating that the court considered and ruled on either of these questions, probably because, as explained above, our previous iterations of the catalyst theory did not clearly establish that either was at issue. Remand is therefore appropriate so that the parties may litigate and the trial court may determine these two issues.
The dissent appears to question the rule stated in Beasley that a consumer class action suit conferring significant benefits on a large number of people vindicating a right of substantial societal importance can be the basis of an award of section 1021.5 attorney fees. It cites in support of its position Flannery v. California Highway Patrol (1998) 61 Cal.App.4th 629, 635-636 [71 Cal.Rptr.2d 632], a case involving a single plaintiff’s lawsuit under the Fair Employment and Housing Act. But Flannery merely held that a plaintiff who enforces a statutory right is not necessarily entitled to section 1021.5 fees when the primary effect of the suit is to vindicate an individual economic interest. Flannery does not contravene the rule in Beasley. Nor does the dissent’s reweighing and recharacterization of the evidence persuade us that the trial court’s conclusion—that the lawsuit itself furthered the public interest by resulting in extensive consumer remedies, which served as a deterrent to future conduct jeopardizing public safety—was unsupported by substantial evidence.
In the present case, the trial court expressly stated that it was not enhancing the fees because of the “novelty or difficulty of the issues.”