OPINION
{1} In this case, we examine the reasonableness of attorney fees that were awarded on the basis of the common fund doctrine, pursuant to terms in a class action settlement agreement. We conclude that under the circumstances of this case, the settlement provisions regarding the common fund doctrine are dispositive and that the district court did not abuse its discretion in awarding fees by using the percentage-of-recovery method or in its application of the method. Relying on our evidentiary review of the Rule 16-105 NMRA factors used by the district court to evaluate the reasonableness of the fee, we also conclude that the fee awarded in this case was reasonable. Accordingly, we affirm.
I. BACKGROUND
{2} This class action is one of many related suits filed against Appellant Microsoft Corporation (Microsoft) in federal and state courts throughout the country. Numerous complaints were filed after the United States District Court for the District of Columbia issued in July of 1999 findings of fact establishing that Microsoft had engaged in conduct indicating an improper use of market advantage to stifle innovation. See United States v. Microsoft Corp., 84 F.Supp.2d 9, 12 (D.D.C.1999); see also United States v. Microsoft Corp., 87 F.Supp.2d 30, 35 (D.D.C. 2000) (concluding that Microsoft violated the Sherman Antitrust Act and analogous state statutes), aff'd in part and rev’d in part, 253 F.3d 34, 46 (D.C.Cir.2001).
{3} In New Mexico, three class action complaints were filed before March 2000. The three cases were then consolidated by a writ of superintending control issued by the New Mexico Supreme Court. The consolidated amended complaint, alleging that Microsoft had violated New Mexico’s Antitrust Act and Unfair Practices Act, was filed in late 2000. Microsoft filed a motion to dismiss, and Class Plaintiffs filed a motion to certify the Class, both of which were fully briefed and argued. The motion to dismiss was granted in part and denied in part, and a *887second consolidated amended complaint was filed and answered. Extensive discovery began and continued through June 28, 2002.
{4} By late 2002, the district court had certified the Class, and the order granting certification was on appeal to this Court. That appeal was fully briefed, and trial preparation continued while a settlement was negotiated. The Settlement Agreement (Agreement) was finally approved in December 2004, and the pending appeal of the class certification was dismissed. The terms of the Agreement fall into three main categories. First, Microsoft agreed to provide vouchers that could be claimed by a member of the Class and redeemed for a cash reimbursement after purchase of any qualified computer hardware or software. Second, Microsoft agreed to provide a cy pres award of vouchers to eligible schools in New Mexico. The amount of the cy pres award is based on a percentage of vouchers unclaimed and a percentage of claimed vouchers unredeemed by members of the Class. Third, Microsoft promised to pay reasonable attorney fees and expenses, and the parties agreed that the attorney fees would be determined on the basis of the “common fund” doctrine.
{5} Subsequently, after hearing oral arguments regarding Class Counsel’s fee application, the district court awarded $6.1 million in attorney fees, plus appropriate gross receipts tax, and $525,179 in expenses. The court used a percentage-of-recovery calculation and then applied the lodestar method to double-check the reasonableness of the fee award. For additional guidance regarding reasonableness, the court below also considered the factors in Rule 16-105 of the New Mexico Rules of Professional Conduct. Microsoft appeals this award. Additional facts will be developed as relevant to our discussion of the issues.
II. DISCUSSION
A. Standard of Review
{6} An award of attorney fees is reviewed for an abuse of discretion. N.M. Right to Choose/NARAL v. Johnson, 1999-NMSC-028, ¶ 6, 127 N.M. 654, 986 P.2d 450; Gavin Maloof & Co. v. Sw. Distrib. Co., 106 N.M. 413, 415, 744 P.2d 541, 543 (1987) (“[T]he amount of an award of attorney fees lies within the sound discretion of the trial court.”); Hertz v. Hertz, 99 N.M. 320, 331, 657 P.2d 1169, 1180 (1983) (“It is well[ fettled that an award of attorney’s fees on the basis of reasonable compensation is a finding not to be disturbed unless patently erroneous as reflecting an abuse of discretion.” (emphasis, internal quotation marks, and citations omitted)). A discretionary decision based on a misapprehension of the law is an abuse of discretion that must be reviewed de novo. NARAL, 1999-NMSC-028, ¶ 7, 127 N.M. 654, 986 P.2d 450. Thus, the question of whether the correct law has been applied and the district court’s application of that law to the facts are reviewed de novo. Id. ¶¶ 7-8. After we determine whether the correct law has been applied, we review a discretionary decision for an abuse of discretion and reverse “only if it [is] contrary to logic and reason.” Id. ¶ 8 (internal quotation marks and citation omitted). “The test is not what we would have done had we heard the fee request, but whether the trial court’s decision was clearly against the logic and effect of the facts and circumstances before the court.” In re Estate of Greig, 107 N.M. 227, 230, 755 P.2d 71, 74 (Ct.App.1988).
B. Microsoft’s Arguments
{7} Microsoft makes three main arguments that the district court abused its discretion by awarding attorney fees in an amount that is vastly disproportionate to the benefit obtained for the Class. First, Microsoft asserts that the district court erred by applying the percentage-of-recovery method (percentage method) to calculate the fee award and that in these circumstances, the lodestar method should have been used to calculate the fee. Microsoft alternatively contends that even if the percentage method were permissible, the district court erred in its valuation of the recovery and that this error resulted in a fee that was disproportionate to the recovery.
{8} Second, Microsoft argues that the district court erred by automatically accepting Class Counsel’s proffered lodestar because out-of-state Class Counsel’s hourly rates were not defined by market rates in New *888Mexico and because Class Counsel improperly submitted hours spent both on fee litigation and on work performed for Microsoft eases in other states. Microsoft also asserts that Class Counsel failed to provide supporting documentation “for $81,201 worth of lodestar.”
{9} Third, Microsoft argues that the district court erred by concluding that a multiplier of three was an appropriate enhancement of the lodestar in these circumstances. Microsoft contends that this case was a tag-along action, “piggybacking” actions from the federal court and other state courts, and that “[n]either the efforts expended, nor the risks assumed, nor the results obtained ... were exceptional” enough to justify a multiplier of three. Microsoft asserts that a proper lodestar would be $1,509,023 and that this lodestar, enhanced by a multiplier of 1.5, would produce a reasonable fee of $2,263,534, a reduction of almost $4 million from the amount actually awarded.
{10} In addition to the arguments regarding the disproportionality of the fee, Microsoft also challenges the district court’s automatic acceptance of Class Counsel’s claimed expenses. Microsoft contends that the district court did not address Microsoft’s objections to the claimed expenses and “gave no reason at all for its finding that expenses ... were reasonable.” Moreover, Microsoft asserts that out-of-state Class Counsel claimed $225,833 in expenses without a requisite “showing that those expenses were reasonably and necessarily incurred in the New Mexico case .... [and thereby] failed to satisfy their burden of proving that the expenses ... provided any benefit” to the Class.
C. Preliminary Considerations
{11} The reasonableness of attorney fees awarded after a class action settlement is an issue of first impression for this Court. Where New Mexico law is not instructive, we find guidance in other jurisdictions. Law from other jurisdictions is not binding on us, however, even though it may be persuasive. See Breen v. Carlsbad Mun. Sch., 2005-NMSC-028, ¶ 14, 138 N.M. 331, 120 P.3d 413 (“Federal case law is certainly informative, but only to the extent it is persuasive.”); Sec. Ins. Co. of Hartford v. Chapman, 88 N.M. 292, 298-99, 540 P.2d 222, 228-29 (1975) (noting that the decisions of other states with similar statutes are persuasive but not binding).
{12} Generally, the determination of a fee award after settlement of a class action and the subsequent review of that determination involve consideration of a variety of competing interests and responsibilities. First, the judiciary has a duty, pursuant to Rule 1-023(E) NMRA, to review the reasonableness of any award of fees in a class action. See 4 Alba Conte & Herbert B. Newberg, Newberg on Class Actions § 14:1, at 507 (4th ed. 2002) (“To fully discharge its duty to review and approve class action settlement agreements, a district court must assess the reasonableness of the attorney’s fees.”). A court acts as a fiduciary for class members and, in doing so, weighs the interests of the class in light of class counsel’s efforts on their behalf. In re Copley Pharm., Inc., 1 F.Supp.2d 1407, 1409 (D.Wyo.1998) (citing Brown v. Phillips Petroleum Co., 838 F.2d 451, 456 (10th Cir. 1988)); see also In re FPI/Agretech Sec. Litig., 105 F.3d 469, 473 (9th Cir.1997) (discussing the lower court’s role to protect class interests). A court assumes this responsibility because class members often have low individual stakes in the outcome of a case and thus often do not object to the settlement or fee award. Swedish Hosp. Corp. v. Shalala, 1 F.3d 1261, 1265 (D.C.Cir.1993). Moreover, the defendant often has no interest in division of the fund between class members and class counsel. Id. Thus, a court has a duty to establish the reasonableness of a fee award, which arises out of its obligation to protect class interests by, inter alia, preventing conflicts of interest between class counsel and the class members and preventing collusion between class counsel and the defendant. See Staton v. Boeing Co., 327 F.3d 938, 959-60 (9th Cir.2003) (discussing the inherent dangers of class settlements, including the compromising of class members’ interests and collusion between negotiating parties); In re Gen. Motors Corp. Pickup Truck Fuel Tank Prods. Liab. Litig. (In re GM Trucks), 55 F.3d 768, 819-20 (3d Cir.1995) (discussing the necessity for oversight to prevent abuse that may arise from *889potential conflict between the class and its counsel, even in situations that do not meet the strict definitions of a common fund), holding limited on other grounds, Amchem Prods., Inc. v. Windsor, 521 U.S. 591, 619-20, 117 S.Ct. 2231, 138 L.Ed.2d 689 (1997); Weinberger v. Great N. Nekoosa Corp., 925 F.2d 518, 524 (1st Cir.1991) (discussing the potential for conflict between a class and its attorneys, both when fees are paid by the defendant and when fees are distributed from the common fund, because class counsel might urge a settlement for less “in exchange for red-carpet treatment on fees”).
{13} Second, a court must acknowledge the public perception of windfall fees in class actions and use judicial oversight to ensure that fees awarded to class counsel are proportional to the benefit obtained for the class. See Third Circuit Task Force, Third Circuit Task Force Report on Selection of Class Counsel, 74 Temp. L.Rev. 685, 692 (2001) [hereinafter Task Force Report II]. We are sensitive to the criticism of fees that have been awarded in some class actions. See In re Copley, 1 F.Supp.2d at 1418. We are also aware, however, that this perception can be unfounded. See Task Force Report II, 74 Temp. L.Rev. at 692 (“When there is a public reaction to an attorney fee award in a given case, the public is usually unaware of what the lawyers actually did, what risks they took, what investment they made, and how important their lawyering was to victory for the class.”).
{14} Third, this Court must be mindful that it does not substitute its judgment for that of the district court. Mayeux v. Winder, 2006-NMCA-028, ¶ 32, 139 N.M. 235, 131 P.3d 85; cf. Staton, 327 F.3d at 960 (noting that the abuse of discretion standard limits an appellate court’s review of settlement fairness). We do not reverse a decision under the abuse of discretion standard, unless the ruling below “exceeds the bounds of all reason” or is “arbitrary, fanciful, or unreasonable.” Mayeux, 2006-NMCA-028, ¶ 32, 139 N.M. 235, 131 P.3d 85 (internal quotation marks and citation omitted). “Indeed ‘abuse of discretion’ — already one of the most deferential standards of review — takes on special significance when [a court is] reviewing fee decisions. The district court, which is intimately familiar with the nuances of the case, is in a far better position to make [such] decisions than is an appellate court, which must work from a cold record.” Goldberger v. Integrated Res., Inc., 209 F.3d 43, 47-48 (2d Cir.2000) (internal quotation marks and citations omitted); accord Waters v. Int’l Precious Metals Corp., 190 F.3d 1291, 1298 (11th Cir.1999).
{15} Fourth, we must conserve judicial resources and promote judicial efficiency. See NARAL, 1999-NMSC-028, ¶¶ 12, 13, 127 N.M. 654, 986 P.2d 450. Generally, equitable and statutory exceptions to the American rule are not a burden on judicial resources. See id. ¶¶ 15, 25, 31 (declining to adopt a private attorney general doctrine because, inter alia, it would not further the goal of conserving judicial resources). The United States Supreme Court reminds us that a fee application “should not result in a second major litigation.” Hensley v. Eckerhart, 461 U.S. 424, 437, 103 S.Ct. 1933, 76 L.Ed.2d 40 (1983).
{16} Finally, we note that the approval of the settlement is not before us; we review only the award of attorney fees, as provided by the parties’ Agreement. The provisions in the Agreement are controlling, unless the terms produce a result that is unfair to the Class or contrary to the policies underlying class actions. See Montoya v. Villa Linda Mall, Ltd., 110 N.M. 128, 129, 793 P.2d 258, 259 (1990) (stating that the scope of contractual authority to award fees outside of the American rule is defined by the contract and that determining what fees are appropriate is a matter of contract interpretation). We review the district court’s determination of the fee award through the lens of each of the above considerations before we reach our conclusions.
D. The Settlement Agreement and the Common Fund Doctrine
{17} We begin with a review of the pertinent provisions in the Agreement. Microsoft agreed “to pay reasonable attorneys’ fees and expenses in an amount to be determined by the [district c]ourt.” Microsoft reserved the right to oppose the fee application, and both parties reserved the right to appeal the district court’s award. The Agreement pro*890vided that litigation of the fee issues would be subject to the following stipulations relevant to this appeal:
(1) the attorneys’ fees and costs will be paid by Microsoft in addition to the recovery to the New Mexico Settlement Class and the cy pres remedy ...; (2) the amount of the attorneys’ fees will be determined upon the basis of the “common fund” doctrine rather than as a “prevailing party” or statutory fee, with each party being free to argue for what it believe [sic] is a reasonable common fund fee; (3) plaintiffs are free to argue that the value of the “common fund” created by the litigation is any amount up to (but in no case exceeding) Face Value, plus notice and administrative expenses and costs; and Microsoft is free to argue that the “common fund” created by the litigation is any amount less than the Face Value or is the amount of the benefit conferred directly on class members by virtue of this litigation[.]
Since the common fund doctrine is essential to our analysis, we proceed with an overview of the principles and development of the common fund and fee awards.
{18} The common fund doctrine is an equitable exception to the American rule. NARAL, 1999-NMSC-028, ¶ 19, 127 N.M. 654, 986 P.2d 450. Under the American rule, litigants are responsible for payment of their own attorney fees, unless otherwise provided by statute, court rule, or contract. Id. ¶ 9. Under the common fund doctrine, a litigant or a lawyer who recovers, preserves, or increases the value of a common fund, thereby benefitting other persons, may be reimbursed for reasonable fees and expenses from the fund as a whole. Boeing Co. v. Van Gemert, 444 U.S. 472, 478, 100 S.Ct. 745, 62 L.Ed.2d 676 (1980); Swedish Hosp. Corp., 1 F.3d at 1265. The common fund doctrine generally rests on the principles of quantum meruit and unjust enrichment. 4 Conte & Newberg, supra, § 14:6, at 547. Attorneys are entitled to compensation in proportion to the benefit obtained for the class, see id., and unnamed class members who share in a common recovery are unjustly enriched if named class members must bear the burden of litigation. Van Gemert, 444 U.S. at 480, 100 S.Ct. 745. The common fund doctrine is consistent with the American rule; fees are spread among all of the class members of the prevailing party by the allocation of proportional shares of the recovery. Id. at 479, 481, 100 S.Ct. 745.
{19} In this case, Microsoft has contractually agreed to pay attorney fees. A contract providing for attorney fees is enforceable. See NARAL, 1999-NMSC-028, ¶ 9, 127 N.M. 654, 986 P.2d 450 (stating that the American rule does not bar enforcement of a contractual provision for attorney fees). The Agreement specifically provides that Microsoft pay reasonable attorney fees to “be determined upon the basis of the ‘common fund’ doctrine.” Thus, the common fund doctrine is applied in this case because it is required under the Agreement; the doctrine is not applied as an equitable exception to the American rule. The Agreement created a fee arrangement that falls within the American rule — the parties provided by contract that the district court award fees on the basis of the common fund doctrine. See id.
{20} Historically, attorney fee awards made on the basis of the common fund doctrine are reviewed for reasonableness. See, e.g., Edwards v. Alaska Pulp Corp., 920 P.2d 751, 756 (Alaska 1996) (“Although courts may differ and the federal circuits are divided over how best to determine the amount of attorney’s fees under the common fund doctrine, all agree that a ‘reasonable’ attorney’s fee is the proper standard.”); 1 Alba Conte, Attorney Fee Awards § 2:3, at 61 (3d ed.2004); 4 Conte & New-berg, supra, § 14:8, at 603 (“Reasonableness is the touehstone[.]”). Benefit to the class is the talisman for fees and costs in a common fund case, and the common fund is the benchmark used to measure reasonableness. In re The Prudential Ins. Co. of Am. Sales Practices Litig., 148 F.3d 283, 338 (3d Cir.1998) (“[T]he amount of the benefit conferred logically is the appropriate benchmark against which a reasonable common fund fee charge should be assessed.” (internal quotation marks and citation omitted)); In re Horizon/CMS Healthcare Corp. Sec. Litig., 3 F.Supp.2d 1208, 1213 (D.N.M.1998); 1 Conte, supra, § 2:7, at 85 (“[T]he common fund represents the benchmark on which a reasonable fee will be awarded.”). Thus, we ask *891if the district court properly determined what benefit was obtained by Class Counsel on behalf of the Class. For the following reasons, we conclude that the district court did not abuse its discretion in determining the value of the benefit to the Class.
E. Benefit to the Class and Proportionality of the Fee Award
{21} Microsoft argues that the district court abused its discretion because the fee award is vastly disproportionate to the benefit obtained for the Class. Microsoft likens the fee award in the instant case to a windfall fee that contributes to the adverse public perception of the legal profession, as vividly described in City of Detroit v. Grinnell Corp., 495 F.2d 448 (2d Cir.1974), abrogated on other grounds, Goldberger, 209 F.3d at 49-50. “[T]he bitterest complaints [about the legal profession] from laymen [are directed at] the windfall fees and featherbedding that lawyers have managed to perpetuate through ... their influence with the judiciary.” Id. at 469 (second, third, and fourth alterations in original) (internal quotation marks and citation omitted). This argument rests on Microsoft’s assertion that the measure of recovery is “the actual amount expected to be received by the [C]lass,” which was a maximum of $3.1 million in vouchers (10% of the face value of the settlement). Ultimately, the value of claimed vouchers reached only $725,366. Microsoft argues that if either figure is used, a fee award of $6.1 million is disproportionate to the amount that the Class members will receive. The district court valued the benefit to the Class at $24,475,179.
1. Value of the Settlement to the Class
{22} When determining a fee award based on a common fund, the district court must make a reasonable estimate of the value of the settlement. In re Prudential, 148 F.3d at 333-34. The parties contemplated litigation over the reasonableness of the fee and stipulated to a minimum and maximum valuation of the common fund for purposes of argument: “[P]laintiffs are free to argue that the value of the ‘common fund’ created by the litigation is any amount up to (but in no case exceeding) Face Value, plus notice and administrative expenses and costs; and Microsoft is free to argue that the ‘common fund’ created by the litigation is any amount less than the Face Value or is the amount of the benefit conferred directly on class members by virtue of this litigation[.]” Class Counsel did rely on the face value of $31.5 million and argued that it was entitled to 25% of $32 million ($31.5 million, plus the costs of settlement administration), for a total of $8 million in attorney fees. The district court listened to arguments regarding the value of the common fund and, in its discretion, declined to award fees as a percentage of the face value, as argued by Class Counsel. Instead, the court relied on the approximate figures put forth by Microsoft.
{23} Microsoft estimated the value of the total settlement in pleadings below: “The total value of the settlement is likely to be $17 million, and the net value to the [C]lass around $3.1 million.” Counsel for Microsoft repeated these estimations at the fee hearing: “The total amount at issue is about $17 million. That’s how much the schools will receive in vouchers, and about 3.1 million will go in vouchers to the [C]lass---- So that’s the total value of the settlement when you really value what’s going to the New Mexico schools and to the [C]lass.” The district court did not abuse its discretion when it relied on these representations made by Microsoft’s counsel.
{24} On appeal, Microsoft compares the fee award of $6.1 million to the value of the vouchers actually claimed by the Class members — $725,366. The district court did not have this amount before it because the claims period had not expired at the time of the fee hearing and the information was not yet available. Microsoft did not object that the court did not have the actual figures before it. Although the parties stipulated that the “subsequently available claims rate be included in the appellate record,” we decline to consider evidence that was not before the court below. An appellate court does not review a district court decision on the basis of facts that are ostensibly in the record but were not before the court below when it made its ruling. See Rule 12-209(A), (C) NMRA (defining the record proper as papers and pleadings filed in the district *892court and stating that modification of the record may be made by stipulation “[i]f anything material to either party is omitted from the record proper by error or accident”); Martinez v. N.M. State Eng’r Office, 2000-NMCA-074, ¶¶ 47-49, 129 N.M. 413, 9 P.3d 657 (stating that appellate courts do not consider matters that are made part of the record but were not presented to the initial decision maker); In re Estate of Keeney, 121 N.M. 58, 60, 908 P.2d 751, 753 (Ct.App.1995) (indicating that an appellate court would consider only matters that were considered by the trial court at the time it made its decision); John Hancock Mut. Life Ins. Co. v. Weisman, 27 F.3d 500, 506 (10th Cir.1994); cf. State v. Vincent, 2005-NMCA-064, ¶ 9, 137 N.M. 462, 112 P.3d 1119 (“[U]nless the facts necessary to consider a contention are in the record on appeal, we cannot consider the claim.”), cert. granted, 2005-NMCERT-005, 137 N.M. 523, 113 P.3d 346. The value of the claimed vouchers was not “omitted” from the documents filed in the district court; this value had not yet been ascertained. Thus, we review the lower court’s decision by using only the facts before the court at that time.
{25} On appeal, Microsoft cites two New Mexico cases, one saying that the fee award must be “based upon the benefits actually provided to the client” and the other saying that in determining fees, the court is to consider “the amount involved and the results obtained.” These cases establish some general principles regarding fee awards and illustrate fee awards that New Mexico courts have considered proportional or reasonable. However, these cases are distinguishable; they do not address class actions, common funds, or ey pres recoveries. See Calderon v. Navarette, 111 N.M. 1, 2, 800 P.2d 1058, 1059 (1990) (addressing an award of attorney fees for quantum meruit after concluding that the attorney’s contract with his individual client was unenforceable); Thompson Drilling, Inc. v. Romig, 105 N.M. 701, 702, 705, 736 P.2d 979, 980, 983 (1987) (addressing attorney fees available pursuant to a contract in a cause of action to recover money due and to enforce a mechanic’s lien). Microsoft also cites two other New Mexico cases for the proposition that fee awards greater than the client’s recovery are “routinely struck down as excessive.” See Lenz v. Chalamidas, 113 N.M. 17, 18-19, 821 P.2d 355, 356-57 (1991) (reducing fees from $17,616, based on time spent, to $8,000, based on time reasonably required, where the judgment was $13,365); Ulibarri v. Gee, 106 N.M. 637, 638, 640, 748 P.2d 10, 11, 13 (1987) (reducing attorney fees from $30,101 to “a near mid-point” of $10,000, where the judgment amounted to $28,820); see also Thompson Drilling, 105 N.M. at 702, 705-06, 736 P.2d at 980, 983-84 (ruling that a fee award of $15,945, almost two times greater than the jury’s verdict of $8,309, was unreasonable and reducing the award to $5,211). Because these cases do not address class actions, common funds, or cy pres awards, see Lenz, 113 N.M. at 18, 821 P.2d at 356 (enforcing a lien); Ulibarri, 106 N.M. at 638, 748 P.2d at 11 (same), the cases are not controlling. They do provide general guidance, however. The fee awards in Lenz, Ulibarri, and Thompson Drilling were respectively 59.8%, 34.7%, and 62.7% of a client’s recovery. Cf. Nava v. City of Santa Fe, 2004-NMSC-039, ¶ 25, 136 N.M. 647, 103 P.3d 571 (concluding that the district court did not abuse its discretion by awarding attorney fees equivalent to approximately 34% of the total judgment); Lucero v. Aladdin Beauty Colls., Inc., 117 N.M. 269, 272, 871 P.2d 365, 368 (1994) (taking “judicial notice of the standard 33^% of the total recovery (50% of the client’s share of the recovery) that attorneys typically receive when retained on a contingent fee contract” and holding that the district court did not abuse its discretion by awarding about 27% of the total judgment, including attorney fees).
{26} Microsoft also relies on an Alabama case, Union Fidelity Life Insurance Co. v. McCurdy, 781 So.2d 186 (Ala.2000), to support the assertion that the actual amount to be received by the class is the appropriate measure. Union Fidelity’s limited holding rested on the court’s conclusion that notice was inadequate and that class members could therefore not make an informed decision regarding whether or not to file a claim made available by the settlement. Id. at 190-91 (noting generally “that an informed decision not to participate should not dimmish the award to class counsel, because the benefit in such circumstances has actually been provid*893ed, and the class member has simply chosen not to reap the rewards due him or her”); see also Van Gemert, 444 U.S. at 480, 100 S.Ct. 745 (holding that the fee award could be based on the entire recovery, even though some class members did not elect to pursue their individual recoveries, because the “right to share the harvest of the lawsuit upon proof of their identity, whether or not they exercise it, is a benefit in the fund created by the efforts of the class ... counsel”). Microsoft does not contend that Class members in the instant case could not make an informed decision because of inadequate notice.
{27} Moreover, the lower court in Union Fidelity estimated that total claims “might reach $10,000 by the end of the class period,” only 0.2% of the $4.5 million face value of the common fund. 781 So.2d at 191 n. 7. In ornease, Microsoft represented to the court below that estimated claims could reach 10% of the face value of the settlement and that $3.1 million was an acceptable estimate of “the value of vouchers likely to be given to the [CJlass.” Microsoft also repeatedly represented to the court that the total value of the settlement was $17 million, which included the cy pres award of “another $14 million or so.” The value of the settlement represented to the lower court in Union Fidelity and the value of the settlement represented to the court below in our case are markedly different. Thus, Union Fidelity is not helpful to our analysis.
{28} Similarly, the federal cases and the unreported state case cited by Microsoft do not support Microsoft’s contention that the fee award in our case is vastly disproportionate to the benefit received by the Class. With these cases, Microsoft asserts that the benefit to the Class must be measured by the “actual amount expected to be received by the class.” Each case clearly supports the proposition that the benefit to the class must be considered in determining the reasonableness of a fee award. See In re Prudential, 148 F.3d at 336-37 n. 116 (noting that “the final award must depend on a full assessment of the extent of the benefits received by [the] plaintiffs” (internal quotation marks and citation omitted)); In re FPI/Agretech, 105 F.3d at 473 (stating that In re Agent Orange Product Liability Litigation, 818 F.2d 216 (2d Cir.1987), “supports the proposition that a court may reject a fee allocation agreement where it finds that the agreement awards an attorney in disproportion to the benefits that [a specific] attorney conferred upon the class”); Dotson v. Bell Atlantic-Maryland, Inc., Nos. CAL 99-21004, CAL 00-09962, 2003 WL 23508428, at *13 (Md.Cir.Ct. Nov.13, 2003) (not reported in A.2d) (“[U]n-less the [c]ourt requires ... a rational connection between the fee award and the amount of the actual distribution of [sic] the [c]lass, the underlying purpose of class actions will be undermined.”). However, to the extent Microsoft may be contending that the fee award in this case must be compared with an in-the-hand recovery by Class members, none of these cases expresses such a mandate. See In re Prudential, 148 F.3d at 333-34 (agreeing with the district court that though there was no common fund per se, the case was more like a “common fund paradigm than ... a statutory fee-shifting case” and that “[c]onsequently, the district court was required to make a ‘reasonable estimate’ of the settlement’s value in order to calculate attorneys’ fees using the percentage of recovery method”); In re FPI/Agretech, 105 F.3d at 475 (ruling that one of the attorneys representing the plaintiffs was not entitled to share the fees in a second settlement fund because she was not involved in litigating that case); Dotson, 2003 WL 23508428, at *2 (“[T]he benefit received by the affected individuals and by society should be weighed against the costs of obtaining it before the legal system[.]” (emphasis added)); see also Waters, 190 F.3d at 1295 (“[N]o case has held that a district court must consider only the actual payout in determining attorneys’ fees.”); id. at 1296 n. 8 (noting that the cases cited by the defendant “illustrate that whether a district court judge considers the total fund or the actual payment will vary according to the circumstances of each ease” and concluding that “[h]ere, where the district court considered both the total fund and the possibility that the actual payment would be substantially lower, there is no abuse of discretion”). Thus, these cases do not support Microsoft’s position.
{29} In addition, Microsoft discusses the federal Class Action Fairness Act of 2005(Act), § 3, Pub.L. No. 109-2, 119 Stat. *8944 (enacted and effective Feb. 18, 2005) (adding 28 U.S.C. § 1712(a), (e) (Supp.2006)), in support of the position that the fee award must be measured against the actual benefit to the Class. The Act limits attorney fees that are based on a recovery of coupons to the value of the coupons that are actually redeemed. See 28 U.S.C. § 1712(a). As noted earlier, we are not bound by federal law when we interpret state law. Moreover, the original actions in this case were filed almost five years before enactment of the Act. Cf. Waters, 190 F.3d at 1297 n. 11 (declining to consider the policy behind legislation that did not apply to actions filed before its effective date). Thus, we decline to consider the provisions of the Act under these circumstances.
{30} Finally, we note the conflict that can exist in a district court between a defendant’s representations regarding benefit to the class in a settlement approval hearing and the defendant’s representations regarding benefit to the class in a fee hearing. See Union Fidelity, 781 So.2d at 197 (John-stone, J., dissenting) (criticizing the majority opinion for rewarding the defendant-appellant’s “likely duplieitousness and lack of candor” regarding benefit to the class). The judiciary’s duty to protect the interests of the class includes the obligation to protect the class from a settling defendant’s conflict of interest when the defendant extols the benefits to the class at the fairness hearing (in order to gain settlement approval) and then denigrates the benefits to the class at the fee hearing (in order to minimize fees). See id. at 196-97 (Johnstone, J., dissenting) (noting the conflict that arises when the defendant extols the benefits to the class at the settlement hearing and disparages the benefits to the class at the fee hearing).
2. Error in the Choice of the Percentage Method
{31} Microsoft relies on United Nuclear Corp. v. Allendale Mutual Insurance Co., 103 N.M. 480, 709 P.2d 649 (1985), to argue that “application of the percent-of-recovery method ... was legal error.” Our Supreme Court, in dicta, observed that “[t]he ‘lodestar’ and multiplier have generally been applied in civil rights cases and class action suits.” Id. at 485-86, 709 P.2d at 654-55 (addressing the district court’s use of the lodestar method to award statutory attorney fees). United Nuclear is not controlling in this case for several reasons. First, as noted by the two Supreme Court justices who wrote on this issue, the issue was not necessary to the ruling in United Nuclear. Id. at 486, 709 P.2d at 655. Second, the circumstances are inapposite; United Nuclear involved attorney fees available pursuant to statutory authority. Id. at 485-86, 709 P.2d at 654-55; see Sloan v. State Farm Mut. Auto. Ins. Co., 2004-NMSC-004, ¶ 12, 135 N.M. 106, 85 P.3d 230 (“[C]ases are not authority for propositions not considered.” (alteration in original) (internal quotation marks and citation omitted)). Third, the two-justice opinion is not considered precedential. See Gracia v. Bittner, 120 N.M. 191, 195, 900 P.2d 351, 355 (Ct.App.1995). Finally, stating that the lodestar method has been applied in class action suits does not necessarily preclude the existence of another appropriate method to calculate attorney fees in class action suits, particularly where the parties have agreed that the district court would determine a reasonable fee “calculated on a common fund basis.” The phrase “calculated on a common fund basis” would be a strange choice of words if the lodestar method were the only method available.
a. Discretionary Decision
{32} In the majority of jurisdictions, the district court has discretion to determine which method should be used to award fees under the common fund doctrine, depending on the circumstances of each particular case. See, e.g., Goldberger, 209 F.3d at 49-50 (joining the majority of circuit courts, which afford the district court discretion to choose the method); In re FPI/Agretech, 105 F.3d at 472 (stating that deference under the abuse of discretion standard is afforded the court’s choice to use either method for calculating an award); Gottlieb v. Barry, 43 F.3d 474, 487 (10th Cir.1994); Edwards, 920 P.2d at 757-58; Gen. Motors Corp. v. Bloyed, 916 S.W.2d 949, 960-61 (Tex.1996). Historically, the percentage method was most commonly used to award fees from a common fund. Swedish Hosp. Corp., 1 F.3d at 1265; see also Third Circuit Task Force, Court Awarded Attorney Fees, 108 F.R.D. 237, 242 *895(1985) [hereinafter Task Force Report I] (stating that the general standard was “reasonableness under the circumstances of the particular case” and that reasonableness was measured by “what the court believed was a ‘reasonable percentage’ of the amount recovered, with the percentages varying considerably from ease to case”).
{33} Utilization of the percentage method is like calculation of a contingent fee; this method awards to counsel a variable percentage of the amount recovered for the class. In re GM Trucks, 55 F.3d at 819 n. 38. The percentage method allows the judge to “focus on a showing that the fund conferring a benefit on the class resulted from the lawyers’ efforts” and to award thereby the fee in a way “that rewards counsel for success and penalizes it for failure.” In re Cendant Corp. Prides Litig., 243 F.3d 722, 732 (3rd Cir.2001) (quoting In re Prudential, 148 F.3d at 333); In re Copley, 1 F.Supp.2d at 1411. Courts often prefer the percentage method because it is less burdensome than the lodestar method. In re Copley, 1 F.Supp.2d at 1411 (“Simply put, it is much easier and far less demanding of scarce judicial resources to calculate a percentage of the fund fee than to review hourly billing practices over a long, complex litigation.”). In addition, the percentage method rewards efficiency, helps to align the attorneys’ interests with their clients’ interests, and more closely approximates attorney fees found in the marketplace. Id. The rationale for the percentage method can also be applied in situations where the fee and the settlement are claimed to be independent but the funds actually come from the same source. In re GM Trucks, 55 F.3d at 821. Use of the percentage method, however, has led to criticism of the legal profession because the method has sometimes resulted in large fee awards that were disproportionate to the actual efforts of counsel. Task Force Report I, 108 F.R.D. at 242.
{34} Alternatively, the lodestar method can be used to calculate attorney fees in a common fund case. The lodestar is ordinarily used in statutory fee-shifting cases because it provides adequate fees to attorneys who undertake litigation that is socially beneficial, irrespective of the pecuniary value to the classes. In re GM Trucks, 55 F.3d at 821; see also Brown, 838 F.2d at 453-54 (“[T]he lodestar ... provides the starting point for appellate court review of statutory fee awardsf.]”). A lodestar is determined by multiplying counsel’s total hours reasonably spent on the case by a reasonable hourly rate. Kennedy v. Dexter Consol. Sch., 2000-NMSC-025, ¶¶ 34, 35, 129 N.M. 436, 10 P.3d 115 (discussing the application of the requisite lodestar method to a Section 1988 case). An award based on a lodestar may be increased by a multiplier if the lower court finds that a greater fee is more reasonable after the court considers the risk factor and the results obtained. Goldberger, 209 F.3d at 47. The lodestar provides an objective basis for valuing the attorney’s services, a helpful prelude to calculating a fee that serves the purpose of the class action fee award: “to compensate the attorney for the reasonable value of services benefiting [sic] the unrepresented claimant.” Lindy Bros. Builders, Inc. of Phila. v. Am. Radiator & Standard Sanitary Corp., 487 F.2d 161, 167 (3d Cir.1973), limited on other grounds as recognized by Brytus v. Spang, 203 F.3d 238, 242-43 (3d Cir.2000).
{35} The lodestar method has been extensively criticized, however, for deficiencies that outweigh its benefits. Task Force Report I, 108 F.R.D. at 246-49; Fed. Judicial Ctr., Moore’s Federal Practice: Manual for Complex Litigation § 14.121, at 188 (4th ed.2004) [hereinafter Manual for Complex Litigation] (“In practice, the lodestar method is difficult to apply, time-consuming to administer, inconsistent in result, and capable of manipulation.”). Among these problems are the burden the lodestar method imposes on the judicial system and the lack of flexibility in a court’s rewarding counsel for the risk of going to trial. In re GM Trucks, 55 F.3d at 821; Task Force Report I, 108 F.R.D. at 246, 248. Moreover, use of the lodestar has the potential to create conflict between a class and its counsel, In re GM Trucks, 55 F.3d at 820-21, because this use leads to inefficiency and resistance to expeditious settlement by attorneys who want to raise their fees by billing more hours. Ramah Navajo Chapter v. Babbitt, 50 F.Supp.2d 1091, 1096 (D.N.M.1999) [here*896inafter Ramah I]; Task Force Report I, 108 F.R.D. at 248.
{36} The Third Circuit led the movement toward use of the lodestar method for common fund cases in Lindy Brothers. Task Force Report I, 108 F.R.D. at 242-43; see Lindy Bros., 487 F.2d at 167. Some courts will choose to use the lodestar method in common fund cases when it is difficult to determine the value of the common fund, particularly where it is easier to calculate the hours expended and hourly rate than it is to determine an appropriate percentage to award. In re GM Trucks, 55 F.3d at 821. The lodestar method is also used when the percentage method will result in a fee award that is unusually large or unusually small. See Manual for Complex Litigation § 14.121, at 189-90. See generally In re Copley, 1 F.Supp.2d at 1410-12 (comparing the advantages and disadvantages of each method).
{37} The percentage method is preferred in some jurisdictions, including the Tenth Circuit, because this method rewards efficient and prompt resolutions of class actions and because it reflects more accurately the realities of litigation economics in the marketplace. See, e.g., In re Cendant, 243 F.3d at 734; Rosenbaum v. MacAllister, 64 F.3d 1439, 1445 (10th Cir.1995); Uselton v. Commercial Lovelace Motor Freight, Inc., 9 F.3d 849, 853 (10th Cir.1993); Ramah I, 50 F.Supp.2d at 1107-08; Union Fidelity, 781 So.2d at 189 (stating that “the decision whether to apply [the percentage of] the common-fund approach or the lodestar approach is within the trial court’s discretion” even though the preferred method is the percentage of the common fund method); see Gottlieb, 43 F.3d at 483 (discussing Uselton’s implied preference for the percentage method). In using the percentage method, the Tenth Circuit also takes a “hybrid” approach by combining the percentage method with the reasonableness factors derived from a statutory fee case, Johnson v. Georgia Highway Express, Inc., 488 F.2d 714, 717-19 (5th Cir.1974), abrogated on other grounds, Blanchard v. Bergeron, 489 U.S. 87, 92-93, 109 S.Ct. 939, 103 L.Ed.2d 67 (1989), which used the lodestar method. Gottlieb, 43 F.3d at 482-83. The Eleventh Circuit uses only the percentage method to award fees from a common fund. Camden I Condo. Ass’n v. Dunkle, 946 F.2d 768, 774 (11th Cir.1991). In contrast, the Fifth Circuit uses only the lodestar method. Strong v. BellSouth Telecomms. Inc., 137 F.3d 844, 850 (5th Cir.1998).
{38} The United States Supreme Court has not directly addressed the discretion of the lower courts in choosing a method to award fees based on the common fund doctrine. The only guidance provided by the Court can be found in Blum v. Stenson, 465 U.S. 886, 900 n. 16, 104 S.Ct. 1541, 79 L.Ed.2d 891 (1984). There, the Court recognized the difference between the lodestar method applied in statutory fee cases and the percentage method as applied to the common fund doctrine: “Unlike the calculation of attorney’s fees under the ‘common fund doctrine,’ where a reasonable fee is based on a percentage of the fund bestowed on the class, a reasonable fee under [Section] 1988 reflects the amount of attorney time reasonably expended on the litigation.” Id.
{39} Our extensive review of the case law addressing each method leads us to believe that there are advantages and disadvantages to each method, depending on the circumstances of each case. Because the district court is in a better position to assess the circumstances of each case, we join the majority of jurisdictions and hold that the choice of method is within the district court’s discretion.
{40} Microsoft relies on Burke v. Arizona State Retirement System, 206 Ariz. 269, 77 P.3d 444 (Ct.App.2003), to contend that “[a] trial court’s decision to apply the pereent-ofrecovery method over the lodestar method is commonly held to be a question of law.” We disagree. Burke clearly states that “whether the common fund doctrine applies is ... a question of law.” 77 P.3d at 447 (emphasis added); see Edwards, 920 P.2d at 756 (concluding as a matter of law that the common fund doctrine applied because the facts gave rise to unjust enrichment concerns). We do not equate this proposition with Microsoft’s assertion that the district court’s choice of method is a matter of law. As discussed earlier, the Agreement provided that the fee would be awarded on the basis of the com*897mon fund doctrine. In light of our ruling that the district court has discretion to choose the method used in applying the common fund doctrine, we are not persuaded by the cases Microsoft cites. See Brytus, 203 F.3d at 244 (addressing counsel’s argument that the lower court erred as a matter of law because it allegedly proceeded on the legal misunderstanding that a statutory fee provision precluded a common fund fee award); Edwards, 920 P.2d at 758 (“We therefore hold that a trial court applying the common fund doctrine has the discretion to determine whether to apply the percentage of the fund method or the modified lodestar method in order to calculate attorney’s fees.”); Kuhnlein v. Dep’t of Revenue, 662 So.2d 309, 312 n. 4, 315 (Fla.1995) (rejecting, in a 4-3 decision, the percentage approach as a matter of law and noting that “[t]he abuse-of-discretion standard would have applied if we had accepted the percentage approach and we were deciding whether the percentage picked by the trial judge was within the boundaries of reasonableness”). Microsoft also cites to Ketchum v. Moses, 24 Cal.4th 1122, 104 Cal.Rptr.2d 377, 17 P.3d 735, 738-39 (2001) (addressing mandatory fees available by statute), and Serrano v. Priest, 20 Cal.3d 25, 141 Cal.Rptr. 315, 569 P.2d 1303, 1316 (1977) (addressing fees available under the private attorney general doctrine). Because neither ease applies the common fund doctrine, these cases do not affect our holding that the choice of method is within the district court’s discretion.
{41} Microsoft also argues that the percentage method should not have been used because the fees have been shifted: they will be paid by Microsoft and not paid proportionally by the Class members. Therefore, Microsoft reasons that there is no “true” common fund. See Burke, 77 P.3d at 450 (declining to apply the common fund doctrine because the fee spreading element was not present); cf. NARAL, 1999-NMSC-028, ¶ 20, 127 N.M. 654, 986 P.2d 450 (noting that the common fund exception is consistent with the American rule because the fees are deducted from the recovery and “a losing litigant is no better or worse off as a result of the common fund doctrine’s application” (alteration, internal quotation marks, and citation omitted)). Whether a true common fund exists could be relevant when determining if the common fund doctrine applies; however, this was not the question before the district court. As we noted earlier, the common fund doctrine is applied in this case pursuant to a contract and not as an equitable exception. See Montoya, 110 N.M. at 129, 793 P.2d at 259. Microsoft agreed to pay attorney fees to be determined by the district court on the basis of the common fund doctrine. The court was obligated to enforce the terms of the contract by applying the common fund doctrine. See Friedman v. Microsoft Corp., 213 Ariz. 344, 141 P.3d 824, 831 (Ct.App.2006) (concluding that the court properly applied fee-spreading principles when the parties agreed to calculate attorney fees using the common fund doctrine, though recognizing that the case, like ours, was a fee-shifting, not a fee-spreading, case). “[We] will not rewrite a contract to create an agreement for the benefit of one of the parties that, in hindsight, would have been wiser.” Watson Truck & Supply Co. v. Males, 111 N.M. 57, 60, 801 P.2d 639, 642 (1990); see Waters, 190 F.3d at 1300 (“Defense counsel, having reaped the benefits of their bargain in settling the class action suit, cannot expect the court to renegotiate on their behalf the terms of an agreement concluded after arms-length negotiations.”). Once the parties agreed to use the common fund doctrine, cases relating to whether the doctrine applies become immaterial. See Weinberger, 925 F.2d at 520, 526 (concluding that the district court should not be divested of authority to use the lodestar when there is no common fund); Burke, 77 P.3d at 449 (noting that the defendant had not conceded “that fees either would be or should be awarded under the common fund doctrine”); see also In re Cendant, 243 F.3d at 734 (concluding that the percentage method was appropriate even though it was not a “traditional common-fund case”).
{42} In further support of its contention that the district court should have used the lodestar method, Microsoft cites to three state district court cases and asserts that “every other court that has addressed contested fee applications in virtually identical Microsoft settlements has used the lodestar method.” We are not persuaded by this *898argument for three reasons. First, the standard for review of a fee application is the same regardless of whether the application is contested or not. Cf. In re S.D. Microsoft Antitrust Litig., 2005 SD 113, ¶¶ 69-70, 707 N.W.2d 85, 109 (stating that the value of the benefit to the class is measured by the face value of the voucher settlement when the fee application is unopposed and by the value of the vouchers that are actually redeemed or claimed by class members when the fee application is opposed). As discussed earlier, the court has a duty to protect class interests and ensure that the fee is reasonable, even when the fee application is unopposed. See Strong, 137 F.3d at 849-50 (discussing the court’s obligation to examine the fee application when the defendant is paying fees from its own funds and does not oppose the fee application); In re GM Trucks, 55 F.3d at 819-20; Weinberger, 925 F.2d at 520, 524. Second, under South Dakota law, the district court was required to use the lodestar method to calculate the fee. See In re S.D. Microsoft Litig., 2005 SD 113, ¶ 30, 707 N.W.2d 85. Moreover, a comparison of the resulting percentages of the face value in Microsoft settlements from other states demonstrates that the fee awarded is about the same regardless of the method used. See 1 Conte, supra, § 2:6, at 77 (discussing the consistency of fee awards, regardless of the method used to determine the amount).
{43} When comparing resulting percentages, the relative size of the settlement is important. Generally, a reasonable percentage of a common fund decreases as the size of the fund increases. See In re Prudential, 148 F.3d at 339 (stating that the rationale for this inverse relationship rests on the belief that an increase in recovery is often “merely a factor of the size of the class and has no direct relationship to the efforts of counsel” (internal quotation marks and citation omitted)); 1 Conte, supra, § 2:9, at 128 (noting that “[deviations from the normal range of percentage fees are appropriate when ... the fund recovered is extraordinarily large”). In common funds up to about $50 million, fees in complex class actions are usually 20% to 30% of the recovery. 1 Conte, supra, § 2:8, at 106-07. Cases with recoveries ranging from $50 million to $75 million have resulted in fee awards of about 11% to 19%. Goldberger, 209 F.3d at 52. In recoveries of $75 million to $200 million and more, fees range from 6% to 10%. 1 Conte, supra, § 2:9, at 134-35; see id. at 130 n. 3 (listing cases with recoveries over $100 million resulting in fee awards ranging from 5.5% to 9%). With this principle in mind, we compare the fee awards in other Microsoft state cases.
{44} Although the district court in Arizona used the lodestar method, its use of a 3.42 multiplier brought the fee award to 18% of the settlement’s face value of $104.6 million. Friedman, 141 P.3d at 828, 834-35 (concluding that 18% of the face value or 33-35% of the estimated actual value to the class, including the cy pres award, was not unreasonable). The fee award in California was $101 million, 9.17% of a $1.1 billion settlement. In re S.D. Microsoft Litig., 2005 SD 113, ¶ 67, 707 N.W.2d 85; cf. In re Cendant, 243 F.3d at 737 (comparing, inter alia, three cases with recoveries of $1.8 billion to $3.16 billion, with fees ranging from 5% to 8.275%).
{45} In state cases that were uncontested, Microsoft agreed to pay fees in amounts between 8.32% and 20%. See In re S.D. Microsoft Litig., 2005 SD 113, ¶ 67, 707 N.W.2d 85 (comparing the percentage of face value in different states); id. ¶ 93 (Meierhenry, J., dissenting) (quoting the lower court’s comparison of cases in other states). In comparison, the district court in our ease awarded 19.365% of the face value. In light of the relative size of the settlement in our case and the percentages of face value awarded in “virtually identical” cases, we cannot conclude that the district court abused its discretion by awarding 19.365% of the face value of the settlement. See generally 1 Conte, supra, § 2:32, at 240-50 (chart comparison of fees).
{46} In its reply brief, Microsoft relies heavily on In re S.D. Microsoft Litigation. After more than four years of proceedings, an evidentiary hearing, and review of the reasonableness factors, see In re S.D. Microsoft Litig., 2005 SD 113, ¶¶ 4, 20, 93, 707 N.W.2d 85, the South Dakota district court applied a multiplier of two to award a fee of $2,064,000, which is 22.12% of the face value. Id. ¶¶23, 67. The South Dakota Supreme Court, in a decision split 3-2, reversed the *899award, remanded for a recalculation of the lodestar, and instructed the district court to forego the use of a multiplier. Id. ¶¶ 43, 51, 74, 80. We are not persuaded by the three-judge majority opinion. First, the parties in the South Dakota case did not agree that the district court would determine the fee on the basis of the common fund doctrine. See id. ¶ 22; see also id. ¶¶ 10, 32 (regarding the provisions of the settlement agreement). Second, South Dakota law requires the district court to use the lodestar method. Id. ¶¶ 30, 34. Third, as noted earlier, we do not agree that a different standard should be applied to fee applications that are opposed by the defendant. See id. ¶¶ 69, 70. We are more persuaded by the two-judge dissenting opinion — “the majority is merely substituting its ruling for that of the trial court.” Id. ¶ 88; see also id. ¶ 92 (“The trial court’s decision is not clearly against reason and evidence.”). We agree with the dissent, “[A] judicial mind, in view of the law and the circumstances in this case, could reasonably have reached the conclusion arrived at by the trial court.” Id. ¶ 93 (alteration, internal quotation marks, and citation omitted); see Task Force Report I, 108 F.R.D. at 273 (stating that the Task Force believed the abuse of discretion standard “simply has not been employed,” that “reversible error occurs when a district court errs as a matter of law by utilizing improper standards or procedures in determining fees,” and that “the court of appeals should honor the discretion its standard accords to district judges and exercise restraint in reviewing fee awards”).
b. Propriety of Estimating the Value of the Settlement
{47} Microsoft further contends that the percentage method is not appropriate because “the parties could only estimate the value of the settlement.” In support of its argument, Microsoft relies on Strong: “[S]everal courts have advocated the use of the lodestar method in lieu of the percentage of fund method precisely in the situation where the value of the settlement is difficult to ascertain}.]” 137 F.3d at 852 n. 5. Microsoft further asserts that the benefit to the Class “was impossible to quantify ... with any degree of certainty until after the close of the period within which [C]lass members could apply for vouchers.” See In re GM Trucks, 55 F.3d at 822 (“The problem ... is not simple, for arguably, any settlement based on the award of certificates would provide too speculative a value on which to base a fee award.”). We are not convinced. We agree that whether the value of the settlement can be ascertained is a significant factor in the lower court’s decision to use one method or the other. See Van Gemert, 444 U.S. at 478-79, 100 S.Ct. 745 (spreading fees when “each member of a certified class has an undisputed and mathematically ascertainable claim to part of a lump-sum judgment recovered on his behalf’). However, “the ultimate choice of methodology will rest within the district court’s sound discretion.” In re GM Trucks, 55 F.3d at 821.
{48} Other jurisdictions have awarded fees on the basis of common funds that could vary depending on the number of class members who take advantage of a right to share in the common fund. See Van Gemert, 444 U.S. at 480, 100 S.Ct. 745. Although valuation of the common fund in our case requires computation, the value to the Class is not difficult to ascertain. Each member has an “undisputed and mathematically ascertainable claim.” See id. at 479, 100 S.Ct. 745; see also In re GM Trucks, 55 F.3d at 821-22 (concluding that the percentage method was appropriate when the settlement included vouchers because “the settlement, though difficult to value, did not award the even more hard-to-value intangible rights that could in some limited circumstances justify using the lodestar method”); cf. Edwards, 920 P.2d at 756 n. 9 (noting that the common fund is not applicable where “sophisticated economic analysis would be required” (internal quotation marks and citation omitted)). Vouchers for fixed amounts were made available to each qualifying Class member, pursuant to the Agreement; thus, the value to the Class could be determined. The face value of the settlement was specifically set at $31.5 million, based on Microsoft’s best estimate of the number of qualifying operating systems and software licenses indirectly purchased by members of the Class, a total of 2,769,617, and on the value of the vouchers to each Class member.
*900{49} Other courts have conducted reasonable estimates of benefit to the class, even when vouchers or coupons were used, and these courts have recommended use of the percentage method based on the estimate. The settlement in In re GM Trucks consisted of $1,000 coupons redeemable for purchase of any new GM truck or Chevrolet light-duty truck built by the defendant. 55 F.3d at 780. Although approval of the settlement was reversed, thus obviating the need for review of the fee award, the court reviewed the primary issues surrounding a fee award in those circumstances. Id. at 819-22. In dicta, the court concluded that the coupon settlement and the separate fee agreement were “more closely aligned with the common fund paradigm than the statutory fee paradigm,” and thus the percentage method was more appropriate, despite the fact that “the nature of the settlement evade[d] the precise evaluation needed for the percentage of recovery method.” Id. at 821. It advised the court below to “make some reasonable assessment of the settlement’s value and determine the precise percentage represented by the attorneys’ fees.” Id. at 822.
{50} The court in O’Keefe v. Mercedes-Benz USA LLC, 214 F.R.D. 266, 304-07 (E.D.Pa.2003), valued the common fund created by a settlement agreement that provided vouchers and limited extended warranties to class members. Id. at 272. The court accepted the face value of the vouchers, id. at 305, and considered four different valuations of the extended warranty before the court determined a reasonable estimate of the settlement’s value. Id. at 305-07. Thus, though the fund “involved a difficult to value non-monetary common fund,” id. at 310, the court based its fee award on a percentage of the common fund’s total estimated value. Id. at 304; see also Wing v. Asarco Inc., 114 F.3d 986, 990 (9th Cir.1997) (stating that the district court must estimate the total value of the settlement when it is impossible to determine the actual value of the recovery); In re Copley, 1 F.Supp.2d at 1412 (“The first step in a percentage of the fund analysis is a determination of the value of the fund.”).
{51} Microsoft cites to several cases in support of its assertion that the recovery in this case was difficult to value. None is persuasive because the cases do not involve the question of estimating value; rather, they merely illustrate the discretion afforded the district court in determining which method, percentage or lodestar, is most appropriate. Three cases are district court opinions in which the courts made a discretionary determination that the lodestar was most appropriate under the particular circumstances of each case. See Charles v. Goodyear Tire & Rubber Co., 976 F.Supp. 321, 324, 325 (D.N.J.1997) (choosing to use the lodestar method where the benefit to the class included equitable relief and vouchers that could be used for only the defendant’s products and services); Osher v. SCA Realty I, Inc., 945 F.Supp. 298, 302, 307 (D.D.C.1996) (choosing to use the lodestar method when the benefit to the class consisted of a restructuring of the business entities and the option to select specific types of shares for particular class members); Cooperstock v. Pennwalt Corp., 820 F.Supp. 921, 926 (E.D.Pa.1993) (choosing to use the lodestar method because the class plaintiffs were only partly responsible for the benefit conferred and had acted only in a monitoring role in the litigation). Each appellate decision cited by Microsoft affirms the lower court’s determination that the percentage of recovery method was not appropriate under the facts of its case. See Strong, 137 F.3d at 846, 848, 850 (affirming the lower court’s use of the requisite lodestar to determine that class counsel was not entitled to the entire fee specifically provided by the settlement agreement); Municipality of Anchorage v. Gallion, 944 P.2d 436, 438, 445, 447 (Alaska 1997) (concluding that the lower court did not err by declining to award fees on the basis of the benefit to the class where the recovery was a restoration of surplus funds in a defined benefit retirement plan). While we agree that the district courts in these cases used the lodestar method, these cases do not stand for the proposition that the lodestar method is the only method that can be used in this type of case. Accordingly, we do not find these cases helpful under the facts of our case.
{52} Although Microsoft argues that the values are difficult to quantify, it did quantify the value of the settlement. Microsoft represented to the court below that the recovery was about $17 million. It posited that 10% of *901the face value, $3.1 million, was an acceptable estimate of the vouchers that would be claimed and that $14 million was an acceptable estimate of the benefit through the cy pres award. As discussed earlier, the district court did not err by relying on Microsoft’s representations below.
{53} Finally, the Agreement specifically provided that the parties could argue that the benefit to the Class was face value plus notice and administrative expenses and costs, with Microsoft being able to argue any lesser amount. Nothing was mentioned about attorney hours or hourly rates. The fact that each party under the Agreement was expressly allowed to argue its view of the amount of the common fund provides strong support for the view that the parties committed the decisions on method and value to the district court’s discretion, as well as believed that the percentage method could be central to the court’s decision. Otherwise, the Agreement would have been written in different terms.
3. Error in Application of the Percentage Method — Gross Versus Net Recovery
{54} Microsoft next argues that the district court erred in applying the percentage method because the court included the cy pres award, the fee award, and expenses in its calculation of the recovery. Without citation to authority, Microsoft asserts that benefit to class members derives “not from the transaction costs that it took to deliver that benefit and not from the amount that will be paid to the lawyers’ preferred charitable cause.” They contend that “lawyers should not be rewarded in proportion to the extent that they managed to increase the transaction costs of the litigation.”
{55} Our own research has revealed a wide variety of calculations of common funds used to measure the benefit to a class, ranging from net benefit distributed to class members to gross benefit to the class. The gross settlement fund is used by most courts applying the percentage method. 4 Conte & Newberg, supra, § 14:6, at 567; of. id. n. 49 (citing district courts that use net recovery). Moreover, the benefit conferred through a cy pres award is ordinarily considered a part of the benefit to the class. See Nat’l Ass’n of Consumer Advocates, Standards and Guidelines for Litigating and Settling Consumer Class Actions, 176 F.R.D. 375, 399 (1997) (including cy pres distributions in the value of a settlement).
{56} The Ninth Circuit has expressly held that the percentage method is not limited to calculations based on net recovery. In a securities class action, an unnamed class member challenged a fee award that was based on gross recovery. Powers v. Eichen, 229 F.3d 1249, 1251 (9th Cir.2000). The class member argued that under the Private Securities Litigation Reform Act of 1995, the district court erred by awarding attorney fees calculated on a percentage of the gross recovery rather than a percentage of the recovery minus expenses. Id. at 1258; see 15 U.S.C. § 78u-4(a)(6) (2000) (“Total attorneys’ fees and expenses awarded by the court to counsel for the plaintiff class shall not exceed a reasonable percentage of the amount of any damages and prejudgment interest actually paid to the class.”). The class member contended that the phrase “amount ... actually paid to the class” limited the fund value, for purposes of calculating the fee, to the net amount received after expert fees, litigation costs, and other expenses have been subtracted. Powers, 229 F.3d at 1258. Thus, the appellant argued that the fee should be calculated as a percentage of net recovery rather than as a percentage of gross recovery. Id. The court disagreed:
Although the new provision requires reasonable fees and expenses, it does not mandate a particular approach to determining fees. The legislation’s primary purpose was to prevent fee awards under the lodestar method from taking up too great a percentage of the total recovery. The new provision, however, does not eliminate the use of the lodestar approach, nor does it require that fees be based on a percentage of net recovery. It simply requires that the fees and expenses ultimately awarded be reasonable in relation to what the plaintiffs recovered.
Id. (citation omitted).
{57} The court further reasoned that “the choice of whether to base an attorneys’ fee award on either net or gross recovery should not make a difference so long as the end *902result is reasonable. Our case law teaches that the reasonableness of attorneys’ fees is not measured by the choice of the denominator.” Id. The court stated that “[i]f twenty-five percent of gross is reasonable, perhaps thirty-five percent of net would be reasonable,” and concluded that the district court was not prohibited from calculating the fee award using the gross settlement amount. Id.; see Waters, 190 F.3d at 1296 n. 8 (“[W]hether a district court judge considers the total fund or the actual payment will vary according to the circumstances of each case.”).
{58} In a later class action for race discrimination, the Ninth Circuit concluded that the lower court could include attorney fees in its valuation of a putative common fund created under a consent decree. Staton, 327 F.3d at 972-73 (“Under these requirements, the monetary relief for the plaintiff class (including attorneys’ fees) provided for in the consent decree could be converted as described above so as to qualify as a common fund from which class counsel could obtain an award of attorneys’ fees.”). The court disallowed the valuation of injunctive relief in calculating the hypothetical fund but included fees and costs in its recalculation of the common fund. Id.; see id. at 975 (discussing net versus gross recovery and relying on Powers to conclude that “[t]he post-settlement cost of providing notice to the class can reasonably be considered a benefit to the class”); see also Wing, 114 F.3d at 987-88, 990 (including a $60.5 million potential recovery on an insurance claim to calculate the value of a common fund totaling $67.5 million). We agree with the Ninth Circuit. An appropriate fee is determined by reasonableness, and reasonableness does not depend on whether the percentage is based on net recovery or gross recovery.
{59} Federal district courts have also included attorney fees as part of the estimated recovery when a court is approving uncontested attorney fees in class action settlements. Varacallo v. Mass. Mut. Life Ins. Co., 226 F.R.D. 207, 223, 250 & n. 28 (D.N.J.2005) (acknowledging the value to the class of defendants’ agreement to pay fees and expenses and calculating the total estimated recovery including the value of fees and expenses paid by defendants); Carnegie v. Mut. Sav. Life Ins. Co., No. Civ.A.CV-99S3292NE, 2004 WL 3715446, at *37 (N.D.Ala. Nov.23, 2004) (mem.) (concluding that the request for attorney fees was reasonable whether viewed as less than 19.6% of the aggregate benefits, including fees and expenses, or viewed as 25.2% of the net benefits); see Lucero, 117 N.M. at 272, 871 P.2d at 368 (including attorney fees in its calculation of the total judgment); see also In re Copley, 1 F.Supp.2d at 1415 (comparing the percentage awarded based on different valuations of the common fund); Chun v. Bd. of Trustees of Employees’ Ret. Sys., 106 Hawai'i 416, 106 P.3d 339, 349, 350 (2005) (“basing the ascertained common fund on the gross amount available to be paid” to the class members and calculating the fee using five different values for the common fund).
{60} As listed in Microsoft’s brief in chief, the district court determined the value of the common fund by adding the following amounts:
Expected benefit to the [C]lass, based on a 10% claims rate: $ 3,150,000
Expected cypres distribution: $14,200,000
Attorneys’ fees of approximately 25% of $24,475,179: $ 6,100,000
Plaintiffs’ counsel’s claimed expenses: $ 525,179
Settlement administration expenses: $ 500,000
Thus, the district court awarded 25% of a common fund valued at $24,475,179. We explain the court’s calculations in a footnote.1 *903In exercising its discretion, the district court valued the common fund by relying on representations made by Microsoft below, provisions of the Agreement, and principles of the common fund doctrine as discussed above. Even if we deleted the attorney fees from the denominator, the fees are approximately one-third of the recovery amount Microsoft used in argument below, plus expenses. This percentage is only slightly higher than the percentages considered usual in cases up to $50 million, and the value of this case was considerably less than $50 million. See paragraph 43. We conclude that the district court did not abuse its discretion by estimating the value of the benefit to the Class in this manner.
F. Errors in Application of the Lodestar Method as a Crosscheck
{61} The district court used the lodestar method to crosscheck the reasonableness of the award calculated as a percentage of recovery. The court used the lodestar provided by Class Counsel of approximately $2 million. Microsoft asserts that the district court erred by automatically accepting Class Counsel’s proffered lodestar because out-of-state Class Counsel’s hourly rates were not defined by market rates in New Mexico and because Class Counsel improperly submitted hours spent both on fee litigation and on work performed for Microsoft cases in other states. Microsoft also contends that Class Counsel failed to provide supporting documentation “for $81,201 worth of lodestar.”
{62} Given our conclusion that the district court did not abuse its discretion by using the percentage method,' our review of the district court’s use of Class Counsel’s lodestar is limited. When used to crosscheck a percentage award, the lodestar is estimated, using information provided by the fee applicant. Manual for Complex Litigation § 14.122, at 193; see Goldberger, 209 F.3d at 50 (“[W]here used as a mere cross-check, the hours documented by counsel need not be exhaustively scrutinized by the district court.”); O’Keefe, 214 F.R.D. at 304, 310 (noting that “[t]he Third Circuit also recommends using a less exacting version of the lodestar to cross-check the fee award’s reasonableness” and stating that the types of arguments made by the defendant in regard to the lodestar “lead to the demise of the lodestar method”); In re Copley, 1 F.Supp.2d at 1414 (“[t]aking all of counsel’s claimed hours ... as legitimate” to conduct a “rough lodestar analysis” and confirm that the award “bears some relation to the work performed by counsel”); Lachance v. Harrington, 965 F.Supp. 630, 650-51 (E.D.Pa.1997) (“It is not necessary to determine with precision what an appropriate lodestar would be. Indeed, one of the primary advantages of using the [percentage] method is to avoid reviewing lodestar figures with a fine tooth comb.”). The crosscheck is conducted by dividing the percentage award by Class Counsel’s lodestar. See Manual for Complex Litigation § 14.122, at 193. If the resulting multiplier is not unreasonable in comparison with multipliers in other cases, the percentage of recovery is considered reasonable. See id.
{63} Microsoft relies on Kennedy and Calderon to argue that the district court erred by accepting Class Counsel’s lodestar. As discussed earlier, the New Mexico cases are inapposite because they do not involve class actions. Moreover, we are not convinced that Kennedy’s discussion of a multiplier in a statutory fee case is helpful in using a lodestar to crosscheck a percentage of recovery award. Under federal law, a multiplier based on risk or contingency cannot be used in a statutory fee case. See City of Burlington v. Dague, 505 U.S. 557, 559, 567, 112 S.Ct. 2638, 120 L.Ed.2d 449 (1992) (adopting the position put forth by Justice White’s principal opinion in Pennsylvania v. Del. Valley Citizens’ Council for Clean Air, 483 U.S. 711, 727, 107 S.Ct. 3078, 97 L.Ed.2d 585 (1987), and concluding that an enhancement for contingency is not appropriate under the fee-shifting provisions of the Solid Waste Disposal Act and the Clean Water Act); see also Del. Valley Citizens’ Council, 483 U.S. at 727, 107 S.Ct. 3078 (“[I]n the absence of further legislative guid*904anee, we conclude that multipliers or other enhancement of a reasonable lodestar fee to compensate for assuming the risk of loss is impermissible under the usual fee-shifting statutes.”). Moreover, the purpose of the crosscheck by the lodestar method is to compare the resulting multiplier with multipliers from other common fund cases. We discuss multipliers in paragraph 74 of this opinion and conclude that the multiplier used in this case was comparable to other similar common fund cases.
{64} In light of our holding that the district court did not err by calculating the fee award using the percentage method, we conclude that the lower court did not err by using an estimated lodestar provided by Class Counsel. In this regard, we examine briefly Microsoft’s specific arguments regarding Class Counsel’s lodestar.
1. Excessive Out-of-State Hourly Rates
{65} The district court has discretion to determine a reasonable hourly rate that reflects the “ ‘prevailing market rates in the relevant community.’ ” Jane L. v. Bangerter, 61 F.3d 1505, 1510 (10th Cir.1995) (quoting Blum, 465 U.S. at 895, 104 S.Ct. 1541). The judge, familiar with the case and the normal rates in the area, may rely on his own knowledge to supplement the evidence regarding a reasonable hourly rate. Smith v. Freeman, 921 F.2d 1120, 1122 (10th Cir.1990); see Calderon, 111 N.M. at 3, 800 P.2d at 1060 (“[T]he court may also apply its own knowledge and expertise regarding the nature of the services rendered ... to calculate the value of the fee.”). Ordinarily, the relevant community is the local area in which the litigation takes place. See Thompson Drilling, 105 N.M. at 705, 736 P.2d at 983; see also Ramos v. Lamm, 713 F.2d 546, 555 (10th Cir.1983), overruled on other grounds, Del. Valley Citizens’ Council, 483 U.S. at 718 n. 4, 730-31, 107 S.Ct. 3078. However, if the litigation is unusual or requires special knowledge and skills, the relevant community ceases to be the local area. See Ramos, 713 F.2d at 555. We cannot conclude that the district court abused its discretion by accepting out-of-state counsel’s hourly rates for estimating a lodestar used to crosscheck a fee award calculated with the percentage method. See Gottlieb, 43 F.3d at 485 n. 8 (noting that the proper hourly rate “is less critical in a percentage fund analysis”). The record reflects that out-of-state counsel has an established national practice in class actions and had extensive experience in litigating similar actions throughout the nation. A nationally established firm brings the significant experience and expertise in large class actions necessary to properly litigate this type of case. See id. (approving the district court’s determination that out-of-state counsel’s higher rates were reasonable for the services rendered). Moreover, the reasonableness and necessity for obtaining out-of-state counsel is illustrated by Microsoft’s substantial use of out-of-state counsel. See Wales v. Jack M. Berry, Inc., 192 F.Supp.2d 1313, 1324 (M.D.Fla.2001) (stating that while local counsel could have handled the case, it was reasonable to retain out-of-state counsel where the defendant found it reasonable to obtain out-of-state counsel).
{66} In addition, we note that Microsoft, in support of its argument regarding hourly rates, cites to cases that do not address the lodestar computation in a class action common fund case. See Ellis v. Univ. of Kan. Med. Ctr., 163 F.3d 1186, 1192, 1203 (10th Cir.1998) (reviewing a fee that was awarded pursuant to 42 U.S.C. § 1988(b)); Bangerter, 61 F.3d at 1509 (same); Avalon Cinema Corp. v. Thompson, 689 F.2d 137, 138 (8th Cir.1982) (same); Homans v. City of Albuquerque, 264 F.Supp.2d 972, 976 (D.N.M. 2003) (same). Thus, these cases do not alter our analysis or conclusion that the district court did not abuse its discretion by using an estimated lodestar, which included hourly rates exceeding those found in New Mexico, to crosscheck its percentage award.
2. Supporting Billing Records
{67} Microsoft argues that the district court erred by accepting Class Counsel’s lodestar because the lodestar included time spent on fee litigation, time spent on work performed in Microsoft eases in other states, and time submitted without supporting billing records. Microsoft also challenges the district court’s automatic acceptance of Class Counsel’s claimed expenses. Microsoft contends that the district court did not address *905its objections to the claimed expenses and “gave no reason at all for its finding that expenses ... were reasonable.” Moreover, Microsoft asserts that out-of-state Class Counsel claimed $225,833 in expenses without a “showing that those expenses were reasonably and necessarily incurred in the New Mexico case____[and thereby] failed to satisfy their burden of proving that the expenses ... provided any benefit” to the Class.
{68} We are unpersuaded by Microsoft’s arguments regarding time and expenses for three reasons. First, as discussed earlier, once the percentage method is used in a common fund case, the lodestar is used as a crosscheck, and the lodestar itself is estimated. Second, Microsoft has neglected to cite to any documentary evidence in support of its assertions, other than its expert’s report. Our review of the record has revealed no additional evidence to support Microsoft’s contentions. With regard to time spent on fee litigation, Microsoft’s expert’s report concludes the following, without citation to the billing records:
[In-state Class Counsel’s] affidavit ... states that his firm’s lodestar does not include work on the attorneys’ fee application. It should not, but the fact is that [Class Counsel] charged $14,851 in fees on work relating to petitioning for fees and expenses, even though the firms intended to exclude such time from the lodestar. It is interesting, however, to contrast the time and fees devoted in a very concentrated effort from November 2 through November 28, 2004 on the attorneys’ fees and costs Application. In one month the ... attorneys devoted 591 hours or $151,852 ... to the fee application. That is the equivalent of 22% of the total lodestar for work the firm did on the case in almost five years from January 2000 through September 2004.
The report contains other similar conclusions regarding hours spent on petitioning for fees and expenses, work performed on cases in other states, and fees and expenses. See Case v. Unified Sch. Dist. No. 233, 157 F.3d 1243, 1250 (10th Cir.1998) (‘While the parties may submit affidavits from experts regarding reasonableness of the hours billed, the practice is not very helpful when the testimony varies greatly.”). Detailed billing records were made available to the court and defense below, but these records are not included in the appellate record. Moreover, Microsoft does not cite to the summary billing and expense records that can be found in the appellate record. To the extent that the billing records are necessary for our review, the record is deficient. See State v. Rojo, 1999-NMSC-001, ¶ 53, 126 N.M. 438, 971 P.2d 829 (“Where there is a doubtful or deficient record, every presumption must be indulged by the reviewing court in favor of the correctness and regularity of the [district] court’s judgment.” (internal quotation marks and citation omitted)); In re Estate of Heeter, 113 N.M. 691, 694, 831 P.2d 990, 993 (Ct.App.1992) (“This court will not search the record to find evidence to support an appellant’s claims.”).
{69} Class Counsel submitted their own expert report and affidavits that countered Microsoft’s contentions. Although Microsoft does not present its argument as one of substantial evidence to support the facts (e.g., whether the time allocated was time spent preparing the fee application, whether out-of-state counsel’s multi-state time was beneficial to the New Mexico Class, or whether claimed expenses were reasonably incurred), it is clear that substantial evidence exists in Class Counsel’s affidavits to support the lower court’s decision. See Mayeux, 2006-NMCA-028, ¶ 33, 139 N.M. 235, 131 P.3d 85; Las Cruces Prof'l Fire Fighters v. City of Las Cruces, 1997-NMCA-044, ¶ 12, 123 N.M. 329, 940 P.2d 177 (“The question is not whether substantial evidence exists to support the opposite result, but rather whether such evidence supports the result reached.”); cf. Kennedy, 2000-NMSC-025, ¶¶ 34, 36, 129 N.M. 436, 10 P.3d 115 (affirming this Court’s reversal of attorney fees based on a lodestar because counsel failed to supply the district court with any time records).
{70} Third, and most importantly, after the court and Microsoft had the opportunity to view the detailed billing records, both parties presented argument regarding these issues, and the court below resolved the issue in favor of Class Counsel. See Case, 157 F.3d at 1250 (“More important is the discre*906tionary determination by the district court of how many hours, in its experience, should have been expended on the specific case, given the maneuverings of each side and the complexity of the facts, law, and litigation.”); see also Reeves v. Wimberly, 107 N.M. 231, 236, 755 P.2d 75, 80 (Ct.App.1988) (“Upon a doubtful or deficient record, every presumption is indulged in favor of the correctness and regularity of the trial court’s decision, and the appellate court will indulge in reasonable presumptions in support of the order entered.”); Sanchez v. Molycorp, Inc., 103 N.M. 148, 153, 703 P.2d 925, 930 (Ct.App.1985) (“[T]he opinions of an expert even where uneontradicted, are not conclusive on facts in issue and the fact[-]finder may reject such opinion in whole or in part.”); Gottlieb, 43 F.3d at 487 (stating that a decision-maker’s determination about the particular circumstances of a ease is like a finding of credibility).
{71} In addition, Microsoft argues unconvincingly, in reliance on New Mexico cases, that the time “supposedly spent by the Lerach firm ‘on the overall Microsoft litigation’ ” did not confer a specific benefit on the New Mexico Class. As noted earlier, New Mexico law regarding attorney fees paid by individual clients is not helpful in analyzing attorney fee awards in large class actions that are national in nature. Microsoft ignores the apparently common practice in large class actions of allocating to each state a percentage of overall time spent on “virtually identical” cases. See In re S.D. Microsoft Litig., 2005 SD 113, ¶¶ 37-43, 707 N.W.2d 85 (remanding for recalculation of fees with instructions to allocate to the South Dakota litigation one-eighth of the total hours worked in eight jurisdictions). Moreover, Class Counsel submitted affidavits detailing their justifications for allocating these expenses to New Mexico. The district court, as the fact-finder, is entitled to resolve any conflicts in the evidence. State v. Casteneda, 97 N.M. 670, 678, 642 P.2d 1129, 1137 (Ct. App.1982).
{72} Finally, the parties agree that the district court erred in granting expenses first claimed by Class Counsel in the fee application, in the amount of $525,179. In the reply brief below, Class Counsel revised their request for reimbursement to $521,601 and acknowledged that they had been reimbursed certain expenses after the fee application was filed. Thus, expenses awarded by the court below shall be offset by any payments made by Microsoft after the fee application was filed.
3. Multiplier of Three
{73} Microsoft asserts that the district court erred by concluding that a multiplier of three was an appropriate enhancement of the lodestar in these circumstances. It contends that this case was a tag-along action, “piggybacking” actions from the federal court and other state courts and, as such, “[njeither the efforts expended, nor the risks assumed, nor the results obtained ... were exceptional” enough to justify a multiplier of three. Microsoft asserts that a proper lodestar would be $1,509,023 and that this lodestar, enhanced by a multiplier of 1.5, would produce a reasonable fee of $2,263,534.
{74} As discussed earlier, use of the lodestar to crosscheck a percentage award involves comparing the resulting multiplier with multipliers from similar cases. The district court divided its fee award of about $6 million by Class Counsel’s lodestar of about $2 million, resulting in a multiplier of three. Although this multiplier approaches the high end of multipliers awarded in recent years, see In re Cendant, 243 F.3d at 737-38, 742, we conclude that the district court did not err when it determined the multiplier of three was reasonable. Other comparable cases have had similar multipliers. See, e.g., In re Infospace, Inc., 330 F.Supp.2d 1203, 1215-16 (W.D.Wash.2004) (using a multiplier of 3.5 in a “garden variety securities case that did not present novel issues of law”); O’Keefe, 214 F.R.D. at 304, 311 (noting that a $4,896,783 award, 15% of $32,645,220, yields a multiplier of “2.95 using class counsel’s estimated reasonable hours and 6.08 using [the defendant’s] estimated reasonable hours”; concluding that neither multiplier seems unreasonable); Kuhnlein, 662 So.2d at 315 (concluding that a multiplier of five is reasonable “to alleviate the contingency risk factor involved and attract high level counsel to common fund cases”); see also In re Cendant, 243 F.3d at 742-43 (concluding that a *907multiplier of three would be acceptable even though the “case was neither legally nor factually eomplex[,] ... did not require significant motion practice or discovery^] ... and the entire duration of the case from the filing of the Amended Complaint to the submission of a Settlement Agreement ... was only four months”); In re Prudential, 148 F.3d at 341 (recognizing that multipliers ranging from one to four are often awarded in common fund cases). In the related Arizona Microsoft case, the court awarded fees to plaintiffs’ counsel by using a multiplier of 3.42. Friedman, 141 P.3d at 835. We accordingly disagree with the majority in South Dakota’s three-two decision. In re S.D. Microsoft Litig., 2005 SD 113, ¶ 76, 707 N.W.2d 85 (concluding that a multiplier should not be applied, “given the lack of ‘exceptional success’ ”). To our mind, the majority’s miserly view of the benefits of the litigation as well as its seemingly de novo review of the issue of attorney fees is entirely inappropriate. See generally id. ¶¶ 85-94 (Meierhenry, J., dissenting).
{75} We are not persuaded by Microsoft’s assertion that a multiplier of three is not justified by Class Counsel’s risk, efforts, or success because it piggybacked a prior government action. First, Microsoft relies on cases that calculate the fee award using the lodestar method, rather than cases that use the lodestar method as a crosscheck, to support its assertion. See In re Auction Houses Antitrust Litig., No. 00 CIV.0648, 2001 WL 210697, at *1, 3 (S.D.N.Y. Feb.26, 2001) (amended order) (concluding that a multiplier was not appropriate where there was .no risk of non-recovery because an interested party had admitted publicly to the alleged wrongdoing); Goldberger, 209 F.3d at 45, 56 (affirming the district court’s decision not to use a multiplier to calculate a lodestar fee award where corporations related to the defendant pleaded guilty prior to the filing of the class action at issue); City of Detroit v. Grinnell Corp., 560 F.2d 1093, 1095-96, 1103 (2d Cir.1977) (concluding that a multiplier was not appropriate when the complaints were filed after the government action had been appealed to the United States Supreme Court, remanded for hearings on relief, and a final decree had been entered), abrogated by Goldberger, 209 F.3d at 49-50; In re Bausch & Lomb, Inc. Sec. Litig., 183 F.R.D. 78, 82, 87, 88 (W.D.N.Y.1998) (using a multiplier of two where the “case did not present any especially novel or difficult issues,” litigation coincided with government action, and news investigations exposed “a serious risk of liability”). Second, the factual circumstances in these cases are not similar enough to ornease for us to conclude that the district court abused its discretion. We review now the reasonableness factors used by the district court in awarding fees.
G. Reasonableness of the Award and Rule 16-105
{76} As we discussed earlier, reasonableness is the ultimate question regarding an award of attorney fees. Under either the lodestar method or the percentage method, “the fee awarded must be reasonable.” Gottlieb, 43 F.3d at 482. The attorney bears the burden of proving that his legal services provided a benefit to the client. Calderon, 111 N.M. at 3, 800 P.2d at 1060. In determining the value of an attorney’s legal services, the district court can use its own knowledge and expertise. Id.; Gavin Maloof, 106 N.M. at 415, 744 P.2d at 543. Historically, New Mexico courts have also used the factors now found in Rule 16-105 of the Rules of Professional Conduct to examine the reasonableness of attorney fees. See, e.g., Calderon, 111 N.M. at 2, 800 P.2d at 1059 (addressing the reasonableness of fees awarded on a theory of quantum meruit after a contingency fee agreement was voided); Fryar v. Johnsen, 93 N.M. 485, 486-87, 601 P.2d 718, 719-20 (1979) (addressing the issue of reasonable attorney fees in the context of workers’ compensation cases). These factors are effectively identical to the “Johnson factors” used by the Tenth Circuit, see Ramah Navajo Chapter v. Norton, 250 F.Supp.2d 1303, 1306 (D.N.M.2002) [hereinafter Ramah IT], and are commonly referred to as the “Fryar factors” in New Mexico. See Econ. Rentals, Inc. v. Garcia, 112 N.M. 748, 764, 819 P.2d 1306, 1322 (1991).
1. Rule 16-105
{77} Rule 16-105(A) establishes the following factors to be considered:
*908(1) the time and labor required, the novelty and difficulty of the questions involved, and the skill requisite to perform the legal service properly;
f2) the likelihood ... that the acceptance of the particular employment will preclude other employment by the lawyer;
(3) the fee customarily charged in the locality for similar legal services;
(4) the amount involved and the results obtained;
(5) the time limitations imposed by the client or by the circumstances;
(6) the nature and length of the professional relationship with the client;
(7) the experience, reputation, and ability of the lawyer performing the services; and
(8) whether the fee is fixed or contingent.
{78} The factors are not of equal weight, and all of the factors need not be considered. Ramah I, 50 F.Supp.2d at 1097 (declining to consider “the time and labor involved factor ... when, in the judgment of the district court, a reasonable fee is derived by giving greater weight to other factors, the basis of which is clearly reflected in the record” (internal quotation marks and citation omitted)). The court below expressly stated that it considered the Rule 16-105 factors, and ample evidence exists in the record to support the court’s decision. See Landavazo v. Sanchez, 111 N.M. 137, 138, 802 P.2d 1283, 1284 (1990) (“Substantial evidence is such relevant evidence that a reasonable mind would find adequate to support a conclusion.”); Brown, 838 F.2d at 455 (“There is ample evidence in the record to support each of the reasons relied upon by the trial judge.”); see also id. at 456 (finding no abuse of discretion where the relevant factors were considered and the evidence in the record supported the court’s determination). We review the relevant factors that were considered by the district court after evidence and argument were presented.
a.Time and Labor, Novelty and Difficulty, and Requisite Skill
{79} This case began more than five years ago and has been vigorously litigated by both parties. Multiple motions and memoranda have been filed. The Class was certified, the certification was appealed, and that appeal was briefed to this Court. In addition, Class Plaintiffs in New Mexico were required to coordinate with attorneys representing similar class actions in other states and federal courts. Further, an indirect purchaser antitrust class action presents novel and complex issues regarding certification, causation, and damages. Moreover, an action against a corporation like Microsoft requires a high level of skill and expertise.
b. Preclusion of Other Employment
{80} The record demonstrates that at specific times during the five-year course of proceedings, Class Counsel would have been precluded from other employment, particularly while preparing for coordination with other jurisdictions, while preparing for class certification and appeal, and while challenging the proposed global settlement.
c. Ordinary Fee for Similar Actions in the Locale
{81} The record contains support for the conclusion that fees awarded in cases recently resolved within New Mexico were based on percentages of 25%, 30%, and 33.33%. In addition, the Federal District Court of New Mexico has noted that percentages usually range from 20% to 30%. Ramah I, 50 F.Supp.2d at 1096.
{82} Moreover, in a class action of this size, relatively infrequent in New Mexico, comparison to similar cases throughout the country is appropriate. The record references studies revealing that fee awards nationwide are often based on 25% to 30%. See Manual for Complex Litigation, § 14.121, at 188; see also, e.g., Edwards, 920 P.2d at 758 & n. 14 (adopting 25% as a baseline percentage and permitting modification according to circumstances); Nat’l Ass’n of Consumer Advocates, 176 F.R.D. at 397 (“In the absence of special circumstances, including either an unusually large monetary recovery or a relatively small monetary recovery coupled with very beneficial but difficult to value equitable relief, the courts have recognized percentage benchmarks ranging from 19 percent to 45 percent of the common fund.”).
*909d. Amount Involved and Results Obtained
{88} On a per capita basis, the New Mexico settlement is the fourth highest nationwide. The face value of the settlement, $31.5 million, represents approximately 25% of the total potential damages quantified by the Class’s experts. Microsoft estimated the value of the settlement, excluding attorney fees and costs, at $17.5 million. The actual value of the settlement is less than face value, due to the nature of the voucher system and the two-step process required to take advantage of the direct benefit to the Class. Members of the Class who purchased an operating system may claim a voucher valued at $13; those who purchased application software may claim a voucher for $6.75. A voucher may be redeemed only after the purchase of computer hardware or a non-custom software application. Products that may be purchased to redeem vouchers are not limited to Microsoft-related products; vouchers may be redeemed for the purchase of any type or brand of hardware or non-custom software. The vouchers are transferable and can be aggregated, within limits. Large-volume users, including New Mexico schools as beneficiaries of the cy pres award, will receive substantial and tangible benefits because the vouchers are transferable and can be aggregated.
e. Time Limitations
{84} As noted by the district court, the risk for Class Counsel was significant when it committed to a case that would be vigorously litigated for years. The proceedings have lasted more than five years. It is not the limit placed on counsel’s time that is significant in a ease like the one at hand. Rather, it is Class Counsel’s commitment to represent the Class, despite the probability of years of litigation. If the Agreement had not been reached, the proceedings could have continued for an indeterminate time.
f. Experience, Reputation, and Ability of Class Counsel
{85} In its expert’s report submitted to the court below, Microsoft acknowledged that Class Counsel includes “one of the premier trial attorneys in New Mexico[, who] has extensive experience and recognized expertise in class action litigation ... on behalf of plaintiffs.” The report also recognized that Class Counsel includes a firm with an established national practice in class actions. A nationally established firm brings the significant experience and expertise in large class actions necessary to properly litigate this type of case. See In re Global Crossing Sec. & ERISA Litig., 225 F.R.D. 436, 467 (S.D.N.Y.2004) (“[T]he quality of opposing counsel is also important in evaluating the quality of plaintiffs’ counsels’ work.” (internal quotation marks and citation omitted)).
g.Type of Fee Arrangement
{86} Class Counsel accepted the ease on a contingent basis. They advanced expenses and worked without payment for five years.
2. District Court’s Consideration of the Factors
{87} After careful consideration, the district court concluded that the majority of the factors weigh in favor of the reasonableness of the fee award. The district court’s application of these factors to our case revealed that experienced and highly skilled attorneys made a significant investment of time and expense in a case with a substantial risk of no payment in order to successfully resolve a complex class action. We cannot say that “the trial court’s decision was clearly against the logic and effect of the facts and circumstances before the court.” In re Estate of Greig, 107 N.M. at 230, 755 P.2d at 74; see Calderon, 111 N.M. at 3, 800 P.2d at 1060 (noting that a district court can rely on its own knowledge and expertise in assessing a reasonable fee); Gavin Maloof 106 N.M. at 415, 744 P.2d at 543 (declaring that a court may rely on its knowledge of the case and the pleadings filed in assessing a reasonable fee). Thus, we conclude that the district court’s determination of reasonableness based on the Rule 16-105 factors was not an abuse of discretion.
{88} In summary, the district court reached its decision after careful consideration of all pleadings, evidence, and oral argument. The judge presided over the entire course of the consolidated litigation and settlement approval. Our review of the record *910reveals that the district court’s award was reasonable based on the facts before it; we will not second-guess the court below, nor can we redo the reasonableness analysis based on facts different from those before the court below. Although Microsoft argues that it was Class Counsel’s burden to prove that the expenses they incurred benefitted the Class, Class Counsel adequately rebutted those contentions in their pleadings and at oral argument below, thereby carrying their burden. On appeal, Microsoft carries the burden of showing that the district court erred in concluding that Class Counsel met its burden below. See Martinez v. Sw. Landfills, Inc., 115 N.M. 181, 184-86, 848 P.2d 1108, 1111-13 (Ct.App.1993) (discussing the appellant’s burden on appeal). Microsoft has not met this burden.
III. CONCLUSION
{89} We hold that the district court did not abuse its discretion by choosing to use the percentage method to calculate the fee award on the basis of the common fund. We also conclude that the court did not abuse its discretion in awarding expenses, with the proviso that the expenses awarded in the order are offset by any expenses reimbursed after the filing of the fee application. Accordingly, we affirm. In light of our opinion, we do not deem it necessary to rule on the motion to supplement the record with material that we have not referenced.
{90} IT IS SO ORDERED.
WE CONCUR: LYNN PICKARD, Judge and JONATHAN B. SUTIN, Judge (specially concurring).