Opinion by Judge FRIEDMAN; Dissent by Judge B. FLETCHER.
FRIEDMAN, Circuit Judge:We decide each of these three appeals, which were argued together, in a single opinion. In these cases, lawyers who successfully represented social security claimants under contingent-fee contracts challenge the sufficiency of the fees the United States District Court for the Central District of California awarded for the services they performed in the judicial phase of the cases. The lawyers contend that in setting their fees the district court failed to follow the methodology the Supreme Court prescribed in Gisbrecht v. Barnhart, 535 U.S. 789, 122 S.Ct. 1817, 152 L.Ed.2d 996 (2002), for determining attorney fees in social security cases.
We hold, however, that in awarding the fees in these eases the district court did not deviate from the standards announced in Gisbrecht and did not abuse its discretion. We, therefore, affirm in all three cases.
I. Each of these cases followed the same pattern: (1) after the Social Security Administration (“Administration”) denied a claim for benefits, the claimant retained an *857attorney to institute a judicial challenge to the administrative action and signed a written agreement under which the attorney would be paid twenty-five percent of “the back pay awarded”; (2) after suit was filed, the case was remanded to the Administration; (3) the Administration awarded the claimant benefits, including substantial past benefits; (4) the attorney requested from the district court a fee of less than twenty-five percent of the back benefits; and (5) the district court (acting through a magistrate judge) awarded the attorney less than the amount sought.
Each of the three claimants was represented by a different attorney: Brian C. Shapiro (“Shapiro”) represented Clara Crawford, Young Cho (“Cho”) represented Ruby Washington, and Denise Bourgeois Haley (“Haley”) represented Daphne M. Trejo. The three attorneys were affiliated with the Lawrence D. Rohlfing (“Rohlf-ing”) law firm, which specializes in social security matters. Since the firm handled all its work on a contingent-fee basis, it had no regular hourly billing rates. In each case the district court calculated what would be a reasonable hourly rate for the work performed. The differences among the three cases relate to the work the attorney performed in the particular case, the time expended, the amount of back benefits awarded, the attorney fee requested, and the fee awarded.
In the Craivford case (No. 06-55822), Shapiro expended 19.5 hours of his own time and 4.5 hours of paralegal time. The past-due benefits awarded were $123,891.20, twenty-five percent of which was $30,972.80. Shapiro sought a fee of $21,000.00, which was 16.95 percent of the award. The court determined that a reasonable fee would be $8,270.00, which it reduced by $3,150.00 to reflect the fee already paid under the Equal Access to Justice Act, producing a net fee award of $5,120.00.
In the Washington case (No. 06-55954), Cho devoted 17.45 hours of his time and 4.7 hours of paralegal time. The past-due benefits awarded were $76,041.00, twenty-five percent of which was $19,010.25. Cho sought a fee of $11,500.00, which was fifteen percent of those benefits. The court determined that a fee of $8,825.53 would be reasonable, which it reduced by the $2,800.00 that had already been paid under the Equal Access to Justice Act, resulting in a net fee award of $6,025.33.
In the third case, Trejo (No. 06-56284), Haley spent 25.5 hours of her time and 1.1 hours of paralegal time. The past-due benefits awarded were $172,223.25, twenty-five percent of which was $43,055.75. Haley sought a fee of $24,000.00, which was fourteen percent of the benefits. The district court determined that a reasonable fee would be $12,650.40, which it reduced by the $3,200.00 prior payment made under the Equal Access to Justice Act, resulting in a net fee of $9,450.00.
We discuss in detail, in Parts IIB and IID, below, the court’s determinations and reasoning in setting the fee in each case.
II. A. Before Gisbrecht, courts ordinarily used a “lodestar” calculation for determining a “reasonable” attorney fee under fee-shifting statutes that required the losing party to pay the prevailing party’s legal fees. See Gisbrecht, 535 U.S. at 801-02, 122 S.Ct. 1817. Under that calculation, to produce a reasonable fee a court multiplied the “number of hours reasonably devoted to each case ... by a reasonable hourly fee.” Id. at 797-98, 122 S.Ct. 1817.
The lodestar calculation also was used in determining reasonable attorney fees in social security cases, although the Social Security Act does not shift payment of attorney fees from the prevailing to the *858losing party. Instead, it provides for payment of a successful claimant’s attorney fee out of the benefits the claimant recovers. A court rendering a judgment favorable to a social security claimant represented by an attorney may “allow as part of its judgment a reasonable fee for such representation, not in excess of 25 percent of the total of the past-due benefits to which the claimant is entitled by reason of such judgment.” 42 U.S.C. § 406(b)(1)(A). Such fee will be paid to the attorney “out of, and not in addition to, the amount of such past-due benefits.” Id. When Gisbrecht was decided, such contingent-fee contracts were “the most common fee arrangement between attorneys and Social Security claimants.” Gisbrecht, 535 U.S. at 800, 122 S.Ct. 1817.
Gisbrecht, in which the Supreme Court reviewed one of our decisions, was such a lodestar social security attorney fee case. There, the district court used the lodestar calculation to determine attorney fees in four social security cases, but “set hourly lodestar rates lower than those that Plaintiffs had requested.” Gisbrecht v. Apfel, 238 F.3d 1196, 1198 (9th Cir.2000). We affirmed the attorney fee awards, stating that “[t]his court follows the ‘lodestar’ method of calculating fees under 42 U.S.C. § 406(b)(1)(A),” and rejected the contention that “the district courts abused their discretion by refusing to increase the lodestar fees based on the contingent nature of their fee agreements.” Id. at 1197, 1198-99.
The Supreme Court reversed. The Court rejected this court’s view that the lodestar calculation was the sole basis for determining attorney fees under the Social Security Act and that a contingent-fee agreement was not to be considered. It “concludefd]” that “ § 406(b) does not displace contingent-fee agreements as the primary means by which fees are set for successfully representing Social Security benefits claimants in court. Rather, § 406(b) calls for court review of such arrangements as an independent check, to assure that they yield reasonable results in particular cases.” Gisbrecht, 535 U.S. at 807, 122 S.Ct. 1817 (footnote omitted). It added that “[wjithin the 25 percent boundary ... the attorney for the successful claimant must show that the fee sought is reasonable for the services rendered.” Id. The Court additionally stated:
Judges of our district courts are accustomed to making reasonableness determinations in a wide variety of contexts, and their assessments in such matters, in the event of an appeal, ordinarily qualify for highly respectful review.
Id. at 808, 122 S.Ct. 1817. The Court concluded the opinion by stating that:
The courts below erroneously read § 406(b) to override customary attorney-client contingent-fee agreements. We hold that § 406(b) does not displace contingent-fee agreements within the statutory ceiling; instead, § 406(b) instructs courts to review for reasonableness fees yielded by those agreements. Accordingly, we reverse the judgment of the Court of Appeals for the Ninth Circuit and remand the case for further proceedings consistent with this opinion.
Id. at 808-09, 122 S.Ct. 1817.
The Court’s holding was narrow: that, in setting reasonable attorney fees under contingent-fee agreements in social security cases, courts cannot base their calculations solely on lodestar calculations. Instead, courts must look primarily to the contingent-fee specified and then, if necessary, adjust the fee to reflect the circumstances of the particular case. The Court stated that the attorney had the burden of showing that “the fee sought is reasonable for the services rendered.” Id. at 807, 122 S.Ct. 1817. The Court then discussed *859five examples of ways in which the contingent-fee arrangement could be tested for reasonableness: (1) the character of the representation; (2) the results the representative achieved; (3) if an attorney were responsible for delay; (4) if benefits were large in comparison to the amount of time spent on the case; and (5) a record of the hours spent representing the claimant and a statement of the lawyer’s normal hourly billing charge. Id. at 808, 122 S.Ct. 1817.
The Court did not decide, or indicate any views on, the following issues:
1. Whether “contingent-fee agreement[s] between claimant and counsel, if not in excess of 25 percent of past-due benefits, [were] presumptively reasonable.” Gisbrecht, 535 U.S. at 792, 122 S.Ct. 1817.
2. The extent to which, if any, courts could utilize a lodestar calculation based upon an appropriate hourly rate and the hours properly spent on the case, in determining the reasonability of the fee.
The Court did not indicate how or on what basis district courts were to determine that a fee was reasonable, but apparently left it to the court’s discretion. See id. at 808,122 S.Ct. 1817. In each of these cases, the district court followed the precepts of Gisbrecht in setting a reasonable fee.
B. In its orders, the district court noted that Gisbrecht controls. In Crawford, the court stated that its “determination ... is governed” by Gisbrecht, in which “the Supreme Court resolved a division among the Circuits on the appropriate method of calculating fees under § 406(b). Rejecting the ‘lodestar method’ which several of the Circuits (including the Ninth Circuit) had been applying, the Supreme Court held: .... ” The order then quoted most of the language from Gisbrecht referred to in Part IIA of this opinion. In Washington, the same magistrate judge used the same language. In Trejo, a different magistrate judge used, and quoted, almost identical language in discussing Gisbrecht.
The court in Crawford set forth seven items it had “considered” “[i]n determining whether the $21,000.00 award sought by Rohlfing is reasonable for the services rendered in the proceedings before this Court.” Those items included the following:
1. There was “no basis for finding that there was any fraud or overreaching by Rohlfing in the making of the contingent fee agreement with plaintiff.”
2. The fee sought “is within the 25% ’boundary’ and less than the total amount of fees that plaintiff agreed to pay Rohlfing under the terms of the contingent fee agreement.”
3. “Rohlfing was able to persuade the Commissioner to stipulate to a remand which ultimately resulted in a fully favorable decision awarding back benefits in the amount of $123,891.20 to Plaintiff.”
4. “This is not an instance where, due to excessive delay attributable to plaintiffs counsel, the back benefits accumulated during the pendency of the case in court.”
5. “[I]t would not be unreasonable for a law firm having the same degree of experience, expertise and reputation in the legal community as Rohlfing to have normal hourly billing charges” under which “adjusted for inflation the 19.5 hours of attorney time plus 4.5 hours of paralegal time correspond to total hourly fees of $5,907.14.” These are the same hours and inflation adjustments Rohlfing presented to the Court.
6. Using these figures, the $21,000.00 sought
includes an enhancement of $15,092.86, which represents 256% of the $5,907.14 *860figure for the risk of nonpayment (i.e., the “contingency factor”). The Court has duly considered Rohlfing’s rationale for the enhancement sought and does not find it to be persuasive. The Court notes that Rohlfing has not provided any data regarding his firm’s success rate that would enable the Court to assess the risk assumed by his firm in representing social security benefits claimants in the Cent[r]al District of California. Moreover, to the extent that the success rate of plaintiffs counsel is attributable to his special skills and expertise in this area (as distinguished from his selectiveness in the cases he agrees to take on), the Court already has taken into account his skills and expertise in arriving at the figure of $5,907.14 as the pre-enhancement value of the services rendered. The Court therefore finds that plaintiffs counsel has not done an adequate job of convincing the Court that the enhancement sought here is reasonable under the circumstances presented.... The Court finds that a reasonable enhancement here would be no more than 40% of the $5,907.14 figure (i.e., $2,362.86), which would result in a total fee award of $8,270.00.
“Put another way, it appears to the Court that the enhancement sought here by plaintiffs counsel is very excessive, and would constitute an unreasonable “windfall,” when the amount of back pay benefits awarded is compared to the amount of time counsel spent on the case. The Court finds that a reasonable enhancement here would be no more than 40% of the $5,907.14 figure (i.e., $2,362.86), which would result in a total fee award of $8,270.00.
“Based upon the foregoing considerations, the Court finds and concludes that the $21,000.00 in fees yielded by the contingent fee agreement and sought by plaintiffs counsel is completely unreasonable under the circumstances, and that a substantial reduction is warranted.”
The court’s order in the Washington case, authored by the same magistrate judge, discusses similar factors the court “considered” “[i]n determining whether the $11,500.00 award sought by Plaintiffs counsel is reasonable for the services rendered in the proceedings before this Court.” In addition, the court included the factor that “Plaintiffs counsel did not have to do much work to persuade the Commissioner to stipulate to a remand because it was very clear that the ALJ erred and a remand was warranted. Accordingly, Plaintiffs counsel incurred very little risk in this matter.” The court also rejected the lawyer’s argument that this case could be compared to class action securities litigation, in which much larger attorney fees were awarded. The court stated:
The complexity of the procedural and substantive issues in class action securities cases, and the risks of prosecuting those types of cases, eclipse those in social security cases. In social security cases, the costs of prosecuting the cases are relatively nominal because most of the plaintiffs are allowed to file their complaints without the prepayment of the filing fees, the issues are narrow, and the briefing in this Court is generally limited to preparing a joint stipulation. The same is not true in class action securities cases where it is not unusual for plaintiffs’ counsel to invest millions of dollars in a case, and engage in a substantial amount of discovery, pretrial briefing and hearings, just to resolve issues relating to class certification in addition to doing the same and more with respect to prosecuting cases on the merits. The complexity and costs of providing notice to the members of the class is another reason for the *861large enhancements in class action securities cases.
The court concluded that “the enhancement sought here by Plaintiffs counsel is excessive and would constitute an unreasonable ‘windfall.’ Therefore, the Motion is denied to the extent it seeks § 406(b) fees in the total amount of $11,500.00. On the other hand, the Court finds a reasonable enhancement here to be 40% of the $6,303.95 figure (i.e., $2,521.58), which results in a total fee award of $8,825.53.”
Finally, in Trejo, the court discussed eight factors it had “considered” “[i]n determining whether the $24,000.00 award sought by plaintiffs counsel is reasonable for the services rendered.” Several of these factors were phrased in language similar to that in Crawford. The court also stated: “The high quality of the representation provided in this case by plaintiffs counsel is evidenced by the fact that, in the face of an adverse ALJ decision which the Appeal Council already had adopted, plaintiffs counsel was able to persuade the Commissioner to stipulate to a remand which ultimately resulted in a fully favorable decision awarding over twelve years of back benefits.”
The court concluded that, based upon the time properly spent on the case and the appropriate hourly charges for the lawyer and paralegal, an appropriate fee would be $6,325.20. The court found that the attorney had
done a wholly inadequate job convincing the Court that the 279% enhancement sought here is reasonable under the circumstances presented. Put another way, it appears to the Court that the enhancement sought here by plaintiff s counsel is grossly excessive and would constitute an unreasonable ‘windfall,’ when the amount of back pay benefits awarded is compared to the amount of time counsel spent on the ease. Even after taking into account the 7-year delay in the receipt of § 406(b) fees, the Court finds that a reasonable enhancement here would be no more than 100% of the $6,325.20 figured, which would result in a total fee award of $12,650.40.
The court then stated:
Based upon the foregoing considerations, the Court finds and concludes that the $24,000.00 in fees sought by plaintiffs counsel is not reasonable, and that a further reduction to $12,650.40 is warranted.
C. The district court’s orders in these three cases show that, in determining a reasonable fee, the court in each case complied with the Gisbrecht principles. In each case the court stated that it was following Gisbrecht and quoted the pertinent portions of that opinion. The court recognized the primacy of the contingent-fee agreements by first determining that they met the § 406(b)(1) guidelines, and then testing them for reasonableness. Id. at 808.
In Crawford and Trejo, the court explicitly stated that, considering all the circumstances, a substantial reduction in the fee sought, which was within but less than the twenty-five percent contingent fee, was required to produce a reasonable fee. Although the Washington court did not explicitly state that it was reducing the fee sought in light of the factors it had considered, such ruling was implicit in the Trejo decision, and this was what the court did (as distinguished from what it said). In each case, the court also stated that the attorney had not met his or her burden of showing that the fee sought was reasonable. The district court’s reduction of the contingent fees in these three cases accorded with Gisbrecht’s recognition that “[i]f the benefits are large in comparison to the amount of time counsel spent on the *862case, a downward adjustment is similarly in order.” 535 U.S. at 808, 122 S.Ct. 1817.
The district court’s analysis of the relevant factors in these cases, including the twenty-five percent contingent-fee agreement, starkly contrasts with the way the district court determined the fees in Gisbrecht. There, the district court based its determination of a reasonable fee solely on “lodestar rates” and refused to increase those rates to reflect the contingent-fee agreements. See Gisbrecht, 238 F.3d at 1198-99. Here, on the other hand, the lodestar calculation was but one of several factors the district court considered when testing the contingent-fee agreement for reasonableness after “looking first to the contingent-fee agreement.” Gisbrecht, 535 U.S. at 808, 122 S.Ct. 1817.
We read Gisbrecht not to prohibit a district court from making lodestar-type calculations, but only from relying exclusively on such calculations and refusing to consider the contingent-fee agreement. Here, the district court noted that Gisbrecht controls, and considered the contingent-fee agreements. The district court, however, concluded that substantial reductions in the fees under those agreements were necessary for the fees to meet the statutory standard of reasonableness. Those rulings complied with the requirements of Gisbrecht.
Unlike the dissent, we do not read the Supreme Court’s Gisbrecht opinion as mandating any particular procedure or format that the district courts must follow in determining a reasonable attorney fee in social security cases. The Court did not, as the dissent apparently concludes, prescribe that in every case the district court mechanically must begin its analysis with the twenty-five percent contingent fee and then make any reduction in that amount that appears appropriate in the particular case.
As we have noted above, what Gisbrecht held was that “§ 406(b) does not displace contingent-fee agreements within the statutory ceiling; instead, § 406(b) instructs courts to review for reasonableness fees yielded by those agreements.” 535 U.S. at 808, 122 S.Ct. 1817. The methodology by which a district makes such reasonableness determinations is for that court to select in the exercise of its sound discretion. In making these determinations, it would be preferable for a district court to begin with the contingency-fee agreement and decrease from there, rather than increase from a lodestar calculation. It does not seem to be an abuse of discretion, however, to use the latter approach as long as the court takes the necessary factors into consideration.
The district courts’ determinations of a reasonable fee in each of these cases satisfied the Gisbrecht standards.
D. Once we conclude that the district court’s fee determinations satisfied Gisbrecht, the remaining question is whether the district court abused its discretion in setting the fees. This court uses the “abuse of discretion” standard in reviewing social security attorney fee awards. See Allen v. Shalala, 48 F.3d 456, 457 (9th Cir.1995) abrogated on other grounds by Gisbrecht v. Barnhart, 535 U.S. 789, 122 S.Ct. 1817, 152 L.Ed.2d 996 (2002).
As the district court noted, it considered a number of factors. While those factors have been previously outlined in this opinion, we again provide these factors to evidence the amount of explanation the district courts provided. In Crawford, the court considered the experience, expertise, and reputation of the firm in determining an hourly fee and accepted counsel’s estimate of the amount of time spent on the *863case. The court then applied the inflation rate counsel presented to it and calculated a reasonable fee. Then the court increased that fee by 40 percent because (a) there was not excessive delay in getting the result and (b) plaintiffs counsel was able to persuade the Commissioner to stipulate to a remand which ultimately resulted in a fully favorable decision. However, the court would not increase it by 256 percent (the amount counsel requested), because plaintiffs counsel did not provide him the success rate data. After arriving at that figure, the court then reduced the $21,000 claimed by plaintiffs counsel to that figure, saying “a substantial reduction is warranted.”
In Washington, the court again considered the experience, expertise, and reputation of the firm in calculating an hourly fee and applied it to the amount of time that counsel had requested. The court then applied the inflation rate counsel had presented to it to calculate a reasonable fee. Then the court increased the fee by 40 percent, because (a) the Commissioner stipulated to the remand, thereby sparing plaintiff and his counsel of the need to research and prepare an exhaustive argument in support of the remand or award of benefits and (b) there was no excessive delay in getting the result. However, the court would not increase it by 82 percent (the amount counsel requested), because (1) “counsel did not have to do much work to persuade the Commissioner to stipulate to a remand, and (2) it was very clear that the ALJ erred and a remand was warranted,” therefore counsel “incurred very little risk in this matter.” The court also rejected a comparison between this case and securities litigation and wrote that counsel “had not provided specific data regarding his particular success rate that enables the Court to assess the true risk assumed by him.”
In Trejo, the court did not accept counsel’s request for time spent by the attorney or paralegal, instead adjusting the appropriate hours spent on the case. The court also adjusted the inflation rate to be applied to an hourly fee. However, the court calculated the hourly fee based on the experience, expertise, and reputation of the firm. The court then increased the fee by 100 percent, because (a) the high quality of representation provided in the case, evidenced by the fact counsel had convinced the Commissioner to stipulate to a remand (which ultimately resulted in a fully favorable decision) in the face of an adverse ALJ decision and (b) although there was excessive delay, it was not due to Plaintiffs counsel. However, the court would not increase the fee by 279 percent, because (1) counsel provided no data regarding the law firm’s success rate enabling the court to assess the risk assumed by counsel; (2) there was no evidence her firm was precluded from any other employment due to their acceptance of this case; and (3) the case did not entail unduly short time limitations. After arriving at that figure, the court then reduced the $24,000 (claimed by plaintiffs counsel) to that figure, saying “a further reduction is warranted.”
The selection and evaluation of the appropriate factors, both pro and con, for determining a reasonable attorney fee in a particular case involves the essence of discretionary action. As the Supreme Court indicated, the district court’s determination of a reasonable fee ordinarily is entitled to “highly respectful review.” Gisbrecht, 535 U.S. at 808, 122 S.Ct. 1817. In setting reasonable fees in these cases, the district court did not abuse its discretion. Unlike the dissent, we conclude that the opinions of the magistrate judges who decided these cases adequately explained the basis and reasons for their decisions.
*864The orders of the district court awarding attorney fees in these three cases are AFFIRMED.