NELSON, Circuit Judge:

Pedro Quesada sued four Los Angeles County sheriffs under section 1983 for a brutal beating he received at their hands. He originally sought $50,000 in damages, though he later decreased this request to $25,000. Prior to trial, the County settled the case, giving Quesada $17,500. Quesa-da had previously agreed to pay his attorney a percentage of any monetary recovery. Quesada filed a motion for statutory attorneys’ fees under 42 U.S.C. § 1988, and submitted the usual documentation, including time records indicating that his attorney spent 147.3 hours preparing for trial, and a statement that this attorney usually charged $150.00 per hour. The County opposed any award of attorneys’ fees, and argued in the alternative that the case was worth no more than $100.00 per hour, and required at most 25-30 hours of preparation.

After an evidentiary hearing, the district judge awarded $7,500 in attorneys’ fees. Plaintiffs requested billing would have generated $22,095. The county’s calculation would have yielded an award of $3,000. The judge explained that although counsel’s hours did exceed the amount of the award, and the reasonable value of counsel’s services exceeded $50.00 per hour, the result was justified by the amount involved and the results obtained.

Quesada then moved for reconsideration, arguing that the award was insufficient, and that the district court had not considered the appropriate factors in reaching the result. The district judge refused to alter the award, explaining that he “considered all of the factors that are set forth in the Kerr case, and ... believe[d] there are a dozen of them, and ... [didn’t] believe the case was a particularly novel case legally ... but ... was influenced by the amount involved and the results obtained.” The court also noted “I’m sure that the reasonable value of your services exceeds $50.00 an hour, but under the circumstances of this case I do believe that the award of attorneys fees was reasonable and appropriate ... because of the amount involved and the fact that it was settled and the fact that you had a contingent-fee arrangement with your client and pursuant to that agreement you agreed to accept as attorneys’ fees for your representation one-third of any recovery.”

Quesada appeals this ruling on the grounds that the district court failed to consider all of the appropriate factors in setting a reasonable fee, and that the district court inappropriately considered his contingent-fee arrangement in setting the award. We review the amount of an attorneys’ fee award for abuse of discretion. See Transgo, Inc. v. Ajac Transmission Corp., 768 F.2d 1001, 1014 (9th Cir.1985), cert. denied, 474 U.S. 1059, 106 S.Ct. 802, 88 L.Ed.2d 778 (1986). Because we find that the district court offered no adequate justification for lowering the fee below a reasonable hourly rate multiplied by the number of hours reasonably spent in preparation, we reverse and remand for a new calculation.

I. Consideration of Appropriate Criteria

Quesada argues that the district court abused its discretion by not awarding fees based on the number of hours reasonably spent on the litigation multiplied by a reasonable rate (lodestar fee). Rather, recognizing that its award did not reflect the number of hours reasonably worked multiplied by a reasonable rate, the court none *539 theless reduced the presumptively reasonable lodestar fee. See City of Riverside v. Rivera, 477 U.S. 561, 106 S.Ct. 2686, 91 L.Ed.2d 466 (1986).

Deviations from the lodestar fee are not necessarily an abuse of discretion. This Circuit requires that courts reach attorneys’ fee decisions by considering some or all of twelve relevant criteria set forth in Kerr v. Screen Extras Guild, Inc., 526 F.2d 67 (9th Cir.1975), cert. denied, 425 U.S. 951, 96 S.Ct. 1726, 48 L.Ed.2d 195 (1976). 1 The court need not discuss each of the guidelines, so long as it discusses those most relevant to the particular case. See Kessler v. Associates Fin. Servs. Co., 639 F.2d 498 (9th Cir.1981). A mere statement that a court has considered the Kerr guidelines does not make a decision within the court’s discretion. The court must “articulate with sufficient clarity the manner in which it makes its determination.” Chalmers v. City of Los Angeles, 796 F.2d 1205, 1211 (9th Cir.1986), amended, 808 F.2d 1373 (9th Cir.1987). Although this articulation requirement does not mandate detailed calculations, “something more than a bald unsupported amount is necessary [to affirm an award of attorneys fees].... [A]t the very least, the district court must set forth the number of hours compensated and the hourly rate applied.” Id. at 1211 n. 3.

In this case, we would have preferred a more detailed explanation of how the district court arrived at its decision. The brief comments offered by the district court make evaluating the reasonableness of the award difficult. Nonetheless, the district court said enough for us to review its decision. At the original fee hearing, the court explained its reasons by stating that it had considered the 12 factors, and that it believed the case was not particularly novel or difficult. The court then noted that it was most influenced by the amount involved and the results obtained.

The district court’s initial reasons for setting the fee below the level that the district court itself found was reasonable for the hours reasonably worked are not sufficient. The simplicity of the issues may not be used to decrease a fee award below the amount calculated by the court as a reasonable lodestar fee. The Supreme Court has repeatedly emphasized that the lodestar fee should be presumed reasonable unless some exceptional circumstance justifies deviation. See, e.g. Pennsylvania v. Delaware Valley Citizens’ Council for Clean Air, — U.S. -, 107 S.Ct. 3078, 3088, 97 L.Ed.2d 585 (1987). It has also explained that “[t]he novelty and complexity of the issues [is] ... fully reflected in the number of billable hours recorded by counsel and thus do not warrant an ... adjustment in a fee based on the number of billable hours times reasonable hourly rates.” Blum v. Stenson, 465 U.S. 886, 898, 104 S.Ct. 1541, 1548, 79 L.Ed.2d 891 (1984). We therefore find that the district court abused its discretion by relying on the simplicity of the issues presented as a basis for reducing a fee below the lodestar.

Insofar as the district court relied on the results obtained to justify a fee below the lodestar, it was mistaken. Que-sada obtained a very favorable result. He sought only money damages from the suit, and he managed to settle for nearly half of his original request, and a full seventy percent of the amount he sought at trial. Furthermore, courts should not reduce lodestars based on relief obtained simply because the amount of damages recovered on a claim was less than the amount requested. Rather, the relief obtained justifies a lower fee if plaintiffs fail to obtain relief on all claims, and if hours spent on unsuccessful claims were not needed to pursue successful claims. See Hensley v. *540 Eckerhart, 461 U.S. 424, 434-35, 103 S.Ct. 1933, 1939-40, 76 L.Ed.2d 40 (1983). Failure to obtain all relief requested for a claim on which the plaintiff prevailed should not deprive plaintiff's attorney of a reasonable hourly fee for hours needed to obtain the relief. See id. at 435-36 n. 11, 103 S.Ct. at 1940-41 n. 11.

The district court should not have reduced the attorney’s fees simply because the damage award was small. Section 1988 was not enacted to encourage only those civil rights suits most likely to produce large damage awards. To the contrary, “[b]ecause damage awards do not reflect fully the public benefit advanced by civil rights litigation, Congress did not intend for fees in civil rights cases ... to depend on obtaining substantial monetary relief. Rather, Congress made clear that it ‘intended that the amount of fees ... not be reduced because the rights involved may be nonpecuniary in nature. ’ ” Rivera, 106 S.Ct. at 2695 (quoting S.Rep. No. 94-1011, p. 6 (176), reprinted in, U.S. Code Cong. & Admin.News 1976 pp. 5908, 5913 (emphasis added by Supreme Court)). Even in cases seeking only monetary relief, “a successful civil rights plaintiff often secures important social benefits that are not reflected in nominal or relatively small damage awards.” Id. at 2694. Therefore, it is inappropriate for a district court to reduce a fee award below the lodestar simply because the damages obtained are small. Permitting such reductions would create an incentive to bring only those civil-rights cases that would produce large damage awards. This incentive conflicts with the purposes of section 1988. See id. at 2695-97.

The reasons offered by the district court at the fee hearing cannot support reduction of the fee below the lodestar. We therefore proceed to consider the additional reason offered by the court in support of its decision: the contingent-fee arrangement between Quesada and his attorney.

II. Consideration of the Contingent-Fee Arrangement

In deciding to lower the fee award below the lodestar, the district court relied on the contingent-fee agreement Quesada made with his attorney. Whether courts may lower fees on this ground raises an issue of first impression in this Circuit. Of course, whether the fee is fixed or contingent is the sixth Kerr factor, and therefore presumably should be considered by courts in setting a reasonable fee. However, because the sixth Kerr factor has not often been construed in this circuit, the uses to which it may be put have not yet been determined. Additionally, several recent Supreme Court cases affect our interpretation of this factor.

A. The Sixth Kerr Factor

The sixth Kerr factor has developed two meanings in separate lines of precedent. Some courts have understood the sixth factor as supporting increases in the lodestar because the contingent nature of the fee agreement creates inherent risks of non-payment. See, e.g., Blum, 465 U.S. at 902-04, 104 S.Ct. at 1550-51 (Brennan, J., concurring); Clark v. City of Los Angeles, 803 F.2d 987, 991 (9th Cir.1986). Under this theory, the existence of a contingent-fee agreement might justify increasing the lodestar, though it could never justify decreasing the lodestar. Of course, Quesada supports this “contingency-multiplier” interpretation of the sixth factor.

Other courts, however, use the sixth Kerr factor by considering the amount a private fee agreement would produce for the attorney, and then giving that fee some weight as a candidate for a reasonable fee under section 1988. See, e.g., Blanchard v. Bergeron, 831 F.2d 563 (5th Cir.1987). On this understanding, the existence of a contingent-fee agreement does not necessarily support raising or lowering the lodestar in the abstract. Rather, district courts may look to the fee an attorney would be entitled to under the contingent-fee agreement in setting a reasonable statutory fee.

This “reasonable-expectation” interpretation of the sixth factor was first advanced in Johnson v. Georgia Highway Express, Inc., 488 F.2d 714 (5th Cir.1974), the Fifth Circuit case from which Kerr quoted the 12 *541 factors. See id. at 718 (stating that “[t]he fee quoted to the client or the percentage of the recovery agreed to is helpful in demonstrating the attorney’s fee expectations when he accepted the case.”); see also Moore v. JAS. H. Mathews, 682 F.2d 830, 840 n. 12 (9th Cir.1982) (stating that the fixed or contingent nature of the fee is an element to be considered in setting the customary fee).

In this case, the district court used the “reasonable-expectation” understanding of the sixth Kerr factor. It explained the relevance of the contingent-fee agreement by telling Quesada’s attorney “you agreed to accept as attorneys’ fees for your representation one third of any recovery if the case did not go to judgment, and 40 percent if it went to judgment. Seventy-five hundred dollars in attorneys’ fees is more than you bargained for. The case was settled for $17,500, and I think you did okay.” Under the contingent-fee agreement, the attorney was entitled to only $5,833.00 from his client. He requested $22,095.00 as a statutory fee. Therefore, the “reasonable-expectation” understanding of the sixth Kerr factor counseled lowering the fee. We must decide whether this understanding of the sixth Kerr factor is correct. We conclude that the purposes of section 1988, recent Supreme Court cases, and our own precedents do not support this understanding. 2

Section 1988 promotes the effective protection of civil rights. Because the victims of civil-rights violations often cannot afford to pay attorneys, Congress enacted this fee shifting statute so that lawyers would accept civil-rights cases, even if those cases would not produce large damage awards. See S. Rep. No 94-1011, p. 6, reprinted in, U.S. Code Cong. & Admin.News 1976 p. 5913 (1976). Following this policy choice, the Supreme Court has interpreted section 1988 so that prevailing plaintiffs’ lawyers recover a reasonable hourly fee for all hours reasonably worked on a case, see, e.g., Hensley, 461 U.S. at 433, 103 S.Ct. at 1939, even if plaintiffs’ monetary recovery falls substantially below the attorneys’ fees. See Rivera, 106 S.Ct. at 2694. This rule reflects a conviction that society may reasonably encourage plaintiffs to spend more litigating a civil-rights dispute than the plaintiff can recover in light of the important nonmonetary, social benefits of vindicating civil rights. See Id. at 2694.

In Rivera, the Supreme Court declined to adopt a per se rule limiting fees to some percentage of monetary recovery. The Court expressed concern that section 1988 was designed to facilitate civil-rights claims, and that if attorneys could only recover a percentage of the damages in fees, fewer civil rights cases would be brought, and only those cases promising high recovery would be pursued. These same considerations lead us to conclude that reasonable fees should not be reduced *542 simply because a lawyer agreed to represent a civil-rights client in exchange for a percentage of any recovery.

Some commentators and courts argue that the reasoning of Rivera does not preclude limiting statutory awards to a percentage of recovery if the attorney agreed to represent a civil-rights plaintiff on the basis of a contingent-fee agreement. See Blanchard, 831 F.2d at 564; Hagge v. Bauer, 827 F.2d 101 (7th Cir.1987); Kirchoff v. Flynn, 786 F.2d 320 (7th Cir.1986); see generally Note, Nonpayment Risk Multipliers: Incentives or Windfalls, 53 U.Chi.L.Rev. 1074, (1986). Four main arguments have been advanced to support this position. First, the presence of a contingent-fee agreement should limit fee recoveries because, if the attorney agreed to represent the plaintiff for a percentage of the damage award, the court need not worry that plaintiffs similarly situated will be unable to obtain adequate counsel. See Cooper v. Singer, 689 F.2d at 929, 931, reconsidered en banc, 719 F.2d 1496 (10th Cir.1983). But see Hamner v. Rios, 769 F.2d 1404, 1407 (9th Cir.1985) (rejecting this view). By definition, anyone who has obtained counsel based on a contingent-fee agreement is in a position to obtain adequate counsel without promising an hourly wage. Second, limiting statutory fees to the amount set by private agreement presents no danger of undercompensating the attorney, because the attorney was willing to represent the plaintiff for exactly the amount set by the agreement. See Johnson, 488 F.2d at 718. Third, awarding statutory fees above the amount the plaintiff is obligated to pay under a private agreement creates a windfall either for the plaintiff, or for plaintiffs attorney, depending on who keeps the difference between the statutory and private fees. See Kirchoff, 786 F.2d at 325; Blanchard, 831 F.2d at 564. Finally, limiting statutory fees to the amount of private debt is especially appropriate in cases that seek money damages, and present settled questions of law, because in these cases no concern arises for deterring civil-rights claims; the only suits that might be deterred are those that will predictably cost more to litigate than they will produce in damages. Because the benefits of these suits are limited to compensation for the individual plaintiff, these suits should not be brought. They are inefficient and wasteful. Cf. Kirchoff, 786 F.2d at 325 (stating that in such cases, limiting statutory awards to a percentage of damages “ensure[s] a reasonable proportion between the recovery and the fees assessed to the defendants.... [N]o client, however wealthy, pays a lawyer more than a dollar to pursue a dollar’s worth of recovery.... Congress [did not mean] to give civil rights plaintiffs a sort of representation that wealthy people would not purchase for themselves.”).

We find these arguments unpersuasive. The first and second misconstrue the significance of contingent-fee agreements in civil-rights cases. Fees under section 1988 do not mirror those available in the private market. See, Rivera, 106 S.Ct. at 2696 n. 9. When an attorney signs a contingent-fee agreement in a civil-rights case, the attorney has not agreed to represent the client for a percentage of the damage award. The attorney signs the agreement knowing that statutory attorneys’ fees will be available to a prevailing plaintiff under section 1988. The fact that an attorney, believing this to be true, will represent a client, does not mean that the same attorney would still represent the same client knowing that she could not receive more in statutory fees than the fee agreement provided.

If we announced a rule allowing district courts to lower hourly fees because of contingent-fee agreements, fewer successful civil-rights suits would be brought. Furthermore, attorneys accepting civil-rights suits would focus on the damage remedy, rather than on injunctive relief, and would select cases most likely to produce large damage awards. See Cooper, 719 F.2d at 1503. These effects do not comport with the goals of section 1988. Additionally, such a rule would create undesirable incentives for billing practices. See, e.g., Id. at 1503 (warning that this limitation might encourage lawyers to negotiate for larger *543 shares in contingent-fee agreements in civil-rights cases); Note, 53 U.Chi.L.Rev. at 1091 (warning of incentive to conceal private agreements).

The third argument advanced in favor of limiting statutory fees to the amount due under a private agreement suggests that an attorney who is compensated for more than the private agreement would have paid receives a windfall. However, the Supreme Court has made clear that paying a reasonable hourly fee is not a windfall merely because the attorney was willing to work for less. See Blum, 465 U.S. at 892-96, 104 S.Ct. at 1545-46 (holding that public-interest attorneys are entitled to the same statutory fees as private attorneys, even though they are willing to represent these plaintiffs for a lower fee). Whatever fee might at a minimum satisfy a lawyer, attorneys are generally entitled under section 1988 to receive a reasonable hourly fee for all hours reasonably spent. See id. at 896 n. 11, 104 S.Ct. at 1547 n. 11; Ackerley Communications, Inc. v. City of Salem Oregon, 752 F.2d 1394, 1397 (9th Cir.), cert. denied, 472 U.S. 1028, 105 S.Ct. 3503, 87 L.Ed.2d 634 (1985). Of course, no real concern for windfalls to plaintiffs arises. Circuits permitting statutory awards in excess of private fee agreements uniformly hold that the higher statutory fee should be used to compensate the attorney in full. See, e.g., Omni Promotions, 815 F.2d at 1423; Cooper, 719 F.2d at 1507.

Finally, we reject the fourth argument— that if the plaintiff seeks only monetary damages, section 1988 should discourage those suits that will recover less than they cost to litigate. The value to society of successful civil-rights claims cannot be measured by the monetary recovery, even if the plaintiff does not seek equitable relief. See Rivera, 106 S.Ct. at 2694. The deterrent effect of these suits, and the publicity brought to the importance of respecting civil rights, as well as the personal satisfaction of having one’s rights vindicated are all nonmonetizable benefits that justify the costs of civil-rights claims that produce small damage awards. See generally id. at 2686.

We conclude that permitting district courts to lower otherwise reasonable statutory fees because the attorney agreed to represent the plaintiff for a percentage of damages that turns out to be less than a reasonable hourly rate would deter vigorous pursuit of civil-rights claims, would deprive civil rights attorneys of a reasonable compensation as defined by the Supreme Court.

A contrary rule would offer an unjustified windfall to defendants. In Hamner, 769 F.2d at 1407-10, we held that if a contingent-fee agreement produces an award greater than would be available under section 1988, the defendant must pay a reasonable hourly fee, and the plaintiff can then pay to the attorney the difference between the statutory fee and the plaintiffs private debt. This rule ensured that defendants would not be penalized by the existence of a private fee agreement. They needed only to pay for hours reasonably spent preparing for trial. We now complete this rule. Just as defendants should not be penalized for the existence of a private fee agreement, so too they should not benefit from the private agreement by being permitted to pay anything less than a reasonable hourly wage. We therefore reject the claim that a contingent-fee agreement can justify lowering an otherwise reasonable lodestar fee. We REVERSE and REMAND for a new calculation.

Notes

. The Kerr factors are (1) the time and labor required; (2) the novelty and difficulty of the questions involved; (3) the skill requisite to perform the legal service properly; (4) the preclusion of other employment by the attorney due to acceptance of the case; (5) the customary fee; (6) whether the fee is fixed or contingent; (7) time limitations imposed by the client or the circumstances; (8) the amount involved and the results obtained; (9) the experience, reputation, and ability of the attorneys; (10) the "undesirability” of the case; (11) the nature and length of the professional relationship with the client; and (12) awards in similar cases.

. Our search for decisions from other circuits indicates a split of opinion over whether statutory fees may be reduced because an otherwise reasonable hourly rate exceeds the attorney’s expected fee under a private agreement. One circuit has decided that fees may not be reduced on this basis. See Sargeant v. Sharp, 579 F.2d 645, 649 (1st Cir.1978). Another circuit has suggested that it holds the same position. See Cooper v. Singer, 719 F.2d 1496, 1503 n. 12 (10th Cir.1983) (suggesting that the only legitimate use for a contingent-fee agreement in calculating statutory awards is for purpose of finding a contingency multiplier). One Circuit has held that the payment due under a contingent fee agreement may be used either to increase or to decrease an otherwise reasonable fee in certain circumstances. See Hagge v. Bauer, 827 F.2d 101 (7th Cir.1987); Kirchoff v. Flynn, 786 F.2d 320 (7th Cir.1986).

Two circuits have taken the position that district courts must limit statutory fees to the amount due under a private fee agreement. See Blanchard v. Bergeron, 831 F.2d 563 (5th Cir.1987); Pharr v. Housing Authority, 704 F.2d 1216 (11th Cir.1983). But see TIC-X-Press Inc. v. Omni Promotions Co., 815 F.2d 1407 (11th Cir.1987) (permitting a statutory award in excess of a private fee agreement). However, many other circuits have refused to set this cap on statutory awards. See Sisco v. J.S. Alberici Const., 733 F.2d 55 (8th Cir.1984); United Slate v. G & M Roofing, 732 F.2d 495 (6th Cir.1984); Cooper v. Singer, 719 F.2d 1496 (10th Cir.1983); Sullivan v. Crown Paper, 719 F.2d 667 (3d Cir.1983); Sanchez v. Schwartz, 688 F.2d 503 (7th Cir.1982); Manhart v. Dep’t of Water, 652 F.2d 904, 909 (9th Cir.1981), vacated and remanded on other grounds, 461 U.S. 951, 103 S.Ct. 2420, 77 L.Ed.2d 1310 (1983); Sargeant v. Sharp, 579 F.2d 645 (1st Cir.1978).