[PUBLISH]

             IN THE UNITED STATES COURT OF APPEALS

                     FOR THE ELEVENTH CIRCUIT                   FILED
                      ________________________           U.S. COURT OF APPEALS
                                                           ELEVENTH CIRCUIT
                                                               JUNE 6 2000
                              No. 99-6206
                                                            THOMAS K. KAHN
                       ________________________                  CLERK

                    D. C. Docket No. 87-01223-CR-T-N

JOHN DILLARD, DAMASCUS CRITTENDEN, JR., et al.,

                                                      Plaintiffs-Appellees-
                                                      Cross-Appellants,

                                   versus

CITY OF GREENSBORO,

                                                      Defendant-Appellant-
                                                      Cross-Appellee.

                       ________________________

                Appeals from the United States District Court
                    for the Middle District of Alabama
                      _________________________
                              (June 6, 2000)


Before COX, BIRCH and BARKETT, Circuit Judges.

PER CURIAM:
       The City of Greensboro, the defendant in this action under § 2 of the Voting

Rights Act,1 appeals an order awarding the plaintiffs attorney fees. The plaintiffs

cross-appeal, seeking to increase the fee. We vacate and render judgment for a

reduced fee.

                                         I. Background

       The plaintiffs here sued the town of Greensboro, Alabama (1990 population:

3047) on a claim that the City’s at-large system for electing city councilmembers

violated § 2 by diluting votes cast by the City’s black citizens, who are a majority of

the City’s voting-age population.2 Greensboro agreed to the entry of a consent decree

establishing liability in 1987, not long after the action was filed. By separate

judgment, the City agreed to pay $5712 of the plaintiffs’ attorney fees and expenses

to that date.3




   1
       42 U.S.C. § 1973.
   2
       Greensboro was one of over a hundred local government units in Alabama that were
defendants in this action.
   3
        This sum represented only Greensboro’s share, among the dozens of local government
units found liable, of the plaintiffs’ fee.
        The consent judgment also required Greensboro to pay attorney fees for all
“extraordinary work” — defined to mean all work over one hour spent in enforcement
proceedings — “that [the plaintiffs] contend is allowed by law.” (R.7-237 at 13.) This decree
did not make it into the record on this appeal (we are quoting the district court’s quotation), and
the plaintiffs do not rely on it; we therefore consider whatever right to fees it awarded to be
equivalent to that statutorily available.

                                                 2
       But this supposed death-knell turned out to be the starting bell for a long fight:

While the 1987 consent decree settled the question whether there would be single-

member city-council districts, and the parties have always agreed that there would be

five districts, the decree did not settle the question of the five districts’ boundaries.

Even within that dispute, however, there was common ground. With the exception of

two alternative plans proposed by the City early in the remedy phase, every plan

advanced by both sides featured two districts whose voting-age population had a black

supermajority.4 And again with the exception of two of Greensboro’s early alternative

plans, all the plans had two districts whose voting-age populations were majority

white. It was the third majority-black district — called the “swing district” — that lay

at the heart of the dispute. The quarrel over the swing district was not about its

boundaries as such (for instance, how they grouped neighborhoods), but centered on

the racial composition of the district. As explained in more detail below, the trend of

the litigation was that the plaintiffs wanted a very black district, while Greensboro

wanted one with a simple black majority or a bare black supermajority in the voting-

age population. The plaintiffs repeatedly accused the City of seeking, through its

proposed boundaries, to preserve white hegemony in Greensboro.



   4
       “Supermajority” may be a loose term; we mean over 65% of the relevant (usually voting-
age) population.

                                              3
      The parties proposed their first plans at two hearings in 1988. The plaintiffs’

plan had a swing district whose population (we think total population; the record is

not clear) was over 83% black. The City’s 1988 swing district was 61.3% black. In

the alternative, the City urged adoption of a plan in which each district, like the City

as a whole, had a black majority of just over 60%.

      The 1990 census came and went without a court ruling. In 1992, the parties

presented new 1990-data plans at a third evidentiary hearing. The plaintiffs proposed

a plan whose swing district’s voting-age population was 76.4% black. In the City’s

1992 plan, blacks comprised 58.5% of the swing district’s voting-age population.

The magistrate judge recommended adoption of the City’s 1992 plan, subject to

preclearance by the Department of Justice (DOJ) under § 5 of the Voting Rights Act.

The magistrate judge worried, however, that black-voter turnout would be so low in

the swing district that blacks would not elect the “councilperson[] of their choice,”

(R.1-62 at 4), and he therefore recommended retaining jurisdiction to see if the City’s

plan remedied the admitted § 2 violation.            The district court adopted the

recommendation and ordered Greensboro to use the City’s plan on an interim basis

for elections in August 1992, while the plan awaited DOJ preclearance.

      The swing district in Greensboro’s plan elected a white to the city council over

a black candidate. Late in 1992, following the election, the DOJ refused to preclear


                                           4
the plan because that district’s black majority had elected a white rather than the

“black-supported candidate.” Greensboro now had no voting plan in effect, and the

plaintiffs returned to court and moved for imposition of a plan like those they had

proposed before, again with a swing district whose voting-age population was over

75% black.

      Following a hearing in 1993, the court permitted the City to seek preclearance

of another plan it had proposed. Although this plan had a blacker swing district (with

a 62% black voting-age majority), the DOJ rejected this plan, as well, because the

swing district was not black enough. The City gave up and, in June 1994, asked the

court by letter to draw a plan that would save the City from having to seek

preclearance. The City’s counsel offered as an example a plan with a swing district

whose voting-age population was 66.3% black.              The plaintiffs responded by

continuing to urge a plan like the one they had proposed in 1992, in which the swing

district’s voting-age population was over 75% black.             The plan proposed by

Greensboro’s counsel, they argued, was seeking a “‘sweet spot’ where the district is

high enough in black percentage to please the Court . . .[,] but low enough that whites

will still be able to control the district and defeat the black-preferred candidate.” (R.2-

100 at 4.) The district court rejected the plan proposed by the City’s counsel,

concluding that there was no evidence that the 66%-black district was black enough


                                            5
to ensure effectuation of black-voter preferences; nor had the City explained why in

the light of other available plans it had “limit[ed] black electoral opportunities in [the]

manner that it did.” (R.2-103 at 9 (internal quotation omitted).) The court thus

adopted the plaintiffs’ plan and ordered a special election.

      The City appealed. This court observed that race appeared to have been a

paramount concern in the choice of districts and questioned the soundness of the

district court’s plan under the equal-protection principles announced in Shaw v. Reno,

, 

(1993), and elaborated in Miller v. Johnson, 

, 

(1995), a case that interprets the Equal Protection Clause to

forbid race-dominated districting. Because the Shaw/Miller issue was raised for the

first time on appeal, the court stopped short of concluding that the adopted plan

violated the Equal Protection Clause. The court instead vacated the district court’s

judgment and remanded for the district court to scrutinize the plan under Miller.

      On remand, Greensboro argued that the court-adopted plan was unconstitutional

and continued to ask the court to draw a plan. For their part, the plaintiffs contended

that race did not predominate in the drawing of their proposed plans, and that the court

should readopt their plan with the 75%+ black swing district. In the end, the court

sidestepped Miller scrutiny of the plaintiffs’ plan and decided to appoint a special




                                            6
master to draw a plan for the court. In light of Miller, the court instructed the master

to eschew race as a factor unless it was necessary to remedy the § 2 violation.

      After visiting Greensboro and reviewing the record of party-proposed plans, the

master concluded that race had figured into all of the prior proposed plans, but that

Greensboro’s black majority and residential segregation made it possible to draw a

race-blind plan that would remedy the § 2 violation.            The master found that

Greensboro was too small to have any remarkable geographical communities of

interest, but that street configurations formed discrete neighborhoods. He thus relied

on major streets and other salient barriers to draw up the plan, crossing them only

when necessary to maintain equipopulation or avoid isolating a subdivision or small

neighborhood. After minor revisions to accommodate two incumbents, the master

proposed a plan that has two districts in which blacks constitute over 75% of the

voting-age population (one 76% and the other 80%), a swing district whose voting-

age population is 66.4% black, and two where whites are a majority (64 and 72%) of

the voting-age population. No party objected, and the district court adopted the plan.

      Then the plaintiffs moved for $253,530 in attorney fees and expenses for their

counsel, Ed Still and Jim Blacksher. Greensboro countered, among other arguments,

that the plaintiffs had not “prevailed” because they did not get a swing district as black

as the one they wanted. The district court rejected this argument, concluding that the


                                            7
plaintiffs had partially prevailed and were entitled to a fee. The district court reduced

the hours billed by 30%, however, on the reasoning that while the plaintiffs prevailed

in obtaining a swing district whose population had a higher percentage of blacks than

Greensboro originally proposed, the swing district was not as black as they would

have liked. The court found that Still and Blacksher’s services were worth $290 an

hour, because that it is about what the court had awarded them before in other cases.

After other adjustments not relevant here, the fee award came to $139,310.20, or

$45.72 for every man, woman, and child in Greensboro (according to 1990 census

figures).

       Greensboro appeals.5 We review the district court’s award for abuse of

discretion, but this deferential standard of review does not prevent us from revisiting

questions of law de novo, or from reviewing subsidiary findings of fact for clear error.

See ACLU v. Barnes, 

, 427, 436 (11th Cir. 1999).

                                         II. Discussion

       42 U.S.C. § 1973l(e) authorizes an award of attorney fees to the prevailing party

“[i]n any action or proceeding to enforce the voting guarantees of the fourteenth or

fifteenth amendment.” This circuit has implicitly construed a violation of § 5 of the



   5
       The plaintiffs cross-appeal, challenging the district court’s reduction of their requested
fee. A rejection of their appeal is implicit in our disposition of Greensboro’s appeal.

                                                 8
Voting Rights Act to be a violation of the “guarantees of the fourteenth or fifteenth

amendment,” see Maloney v. City of Marietta, 

, 1026 (11th Cir. 1987),6

and no one argues here that the same implicit extension of § 1973l would not embrace

§ 2 claims, as well. Its standards are identical to those of 42 U.S.C. § 1988, and we

therefore address the two statutes together as bases for this fee. See Brooks v. Georgia

State Bd. of Elections, 

, 860-61 (11th Cir. 1993). Determining a

plaintiff’s entitlement to attorney fees entails a three-step process. First, a court asks

if the plaintiff has “prevailed” in the statutory sense. See Hensley v. Eckerhart, 

, 433, 

, 1939 (1983). Second, the court calculates the

“lodestar,” which is the number of hours (tempered by billing judgment) spent in the

legal work on the case, multiplied by a reasonable market rate in the local area. See

; Duckworth v. Whisenant, 

, 1396 (11th Cir.

1996). Finally, the court has the opportunity to adjust the lodestar to account for other

considerations that have not yet figured in the computation, the most important being

the relation of the results obtained to the work done. See 

,




   6
        In so doing, we were implicitly effectuating the intent the Senate expressed in its
judiciary committee’s report: “Section [1973l(e)] allows a court, in its discretion, to award
attorneys’ fees to a prevailing party in suits to enforce the voting guarantees of the Fourteenth
and Fifteenth amendments, and statutes enacted under those amendments.” S. Rep. No. 94-295,
at 39 (1975), reprinted in 1975 U.S.C.C.A.N. 774, 807.

                                                

. Greensboro has challenged the district court’s conclusions at

every step of the process, and we address the challenges in turn.

                                  A. Prevailing Party

       Greensboro first argues that the relief the plaintiffs got fell so far short of what

they wanted — guaranteed proportional representation on the city council, according

to Greensboro — that they have not prevailed. “[A] plaintiff ‘prevails’ when actual

relief on the merits of his claim materially alters the legal relationship between the

parties by modifying the defendant’s behavior in a way that directly benefits the

plaintiff.” Farrar v. Hobby, 

, 111-12, 

, 573 (1992). In

the context of an injunction, “a party ‘need not obtain relief identical to the relief [that

it] specifically demanded, as long as the relief obtained is of the same general type.’”

Ensley Branch, N.A.A.C.P. v. Siebels, 

, 1583 (11th Cir. 1994) (quoting

Ashley v. Atlantic Richfield Co., 

, 131 (3d Cir. 1986)). Nor does the

plaintiff need to obtain relief to the extent demanded; getting something suffices to

authorize an award of fees. See 

, 113 S. Ct. at 573.

       We agree with the district court that the plaintiffs have at least partially

prevailed here. Of course they prevailed early on when they exacted from Greensboro

an admission of liability; they have already been compensated for that. But they

prevailed to some extent as well in this marathon remedy phase: the district court


                                            10
concluded, and no one here disputes, that the special master’s plan “effects a complete

remedy for the city’s acknowledged § 2 violation.” (R.6-208 at 15.) We would be

reluctant to hold, and we can find no precedent, to conclude that securing a “complete

remedy” for an only claim does not necessarily make the plaintiff at least a partially

prevailing party. We therefore conclude that the plaintiffs have crossed the threshold

to entitle them to some fee.

                                     B. Lodestar

      The first step in computing a reasonable fee is to determine the lodestar, which

is the product of the number of hours reasonably worked by a lawyer and a reasonable

hourly rate. See ACLU v. Barnes, 

, 427 (11th Cir. 1999). Greensboro

challenges the district court’s conclusions on both factors. First, Greensboro argues

that the number of hours is unnecessarily inflated by over-lawyering and

noncompensable appellate argument preparation. The district court did, however,

deduct the hours attributable to the appeal. Greensboro does not, moreover, point to

the individual line items that it believes to indicate poor billing judgment. It would

be surprising, incidentally, to find any great lapses of billing judgment; a total of

around 700 hours of lawyer time for nearly ten years of litigation and several

evidentiary hearings does not on its face look unreasonable. We therefore accept as




                                          11
within its discretion the district court’s calculation of the number of hours reasonably

spent, which amounted to 403.05 hours for Still and 283.20 for Blacksher.

       The district court’s reasonable hourly rate is, however, another matter.

Plaintiffs are entitled to have their lawyers compensated at a reasonable hourly rate,

which is the “prevailing market rate in the relevant legal community for similar

services by lawyers of reasonably comparable skills, experience, and reputation.”

(quoting Norman v. Housing Authority, 

, 1299

(11th Cir. 1988)). Establishing a claimed market rate is the plaintiff’s burden. See 

Because it is a factual matter, we review for clear error. 

.

       The evidence the plaintiffs adduced here was sparse. Still swore in his affidavit

only that he charges paying clients $200 an hour, but that a judge of the Middle

District of Alabama and a judge of the Northern District of Alabama had previously

awarded him $290 to $350 an hour in voting rights cases.7 What Still charges clients

is powerful, and perhaps the best, evidence of his market rate; that is most likely to be

what he is paid as “determined by supply and demand.” Blum v. Stenson, 

, 895 n.11, 

, 1547 n.11 (1984); see Norman v. Housing Auth., 836


   7
       Still stated that the $200 rate was for “cases where [he] charge[s] a fee rather than take a
contingent risk,” (R.6-221 Still Aff. ¶ 9), but he offered no other evidence of the market value of
his own or similar services.

                                                

, 1299 (11th Cir. 1988). On the other hand, a court should hesitate to give

controlling weight to prior awards, even though they may be relevant, see 

(observing that the twelve-factor standard of Johnson v. Georgia

Highway Express, Inc., 

(5th Cir. 1974), one of whose factors is prior

awards, may have been subsumed into the lodestar rule adopted in Hensley v.

Eckerhart, 

, 

(1983)). The reason is obvious: Prior

awards are not direct evidence of market behavior; the court is not a legal souk. Of

course there is some inferential evidentiary value to the prior award, because in theory

the prior court based the award on the market rate. But giving prior awards

controlling weight over the superior evidence of a lawyer’s actual billing rate equates

to giving the prior awards issue-preclusive value against a party whose interests were

not even arguably represented in the prior litigation.8 The scanty evidence that Still

adduced thus does not support a finding of an hourly rate higher than $200.

       Blacksher did even worse than Still, introducing no evidence of the market

value of his services. His affidavit declares only that his “lodestar” rate — whatever

that means — is $300 an hour; perhaps it means the rate at which courts have

compensated him, but even reaching that conclusion requires an inferential leap.


   8
        Applying issue preclusion against parties without an interest in the prior litigation could
raise due-process concerns. See Richards v. Jefferson County, Ala., 

, 800, 116 S.
Ct. 1761, 1766-67 (1996).

                                                 13
There was, however, on-point evidence offered by the defendants. Montgomery

lawyer David Boyd (who practices in the civil rights field) affirmed that $190-225 an

hour is a typical rate in the Middle District of Alabama for lawyers of Blacksher’s

caliber. This being the only evidence available, the district court clearly erred in

awarding Blacksher more than $225 an hour.

      The question remains whether we must remand for a finding as to the

appropriate rate. This court has decided that we may determine for ourselves, once

we conclude that the district court has abused its discretion, how many hours were

reasonably spent in litigation. See 

. That panel rested this

decision both on the ability of this court to make its own reasonableness assessments

and on the Supreme Court’s command that we not allow a request for attorney fees to

“result in ‘a second major litigation.’” 

(quoting Hensley v. Eckerhart, 

, 437, 

, 1941 (1983)). An analogous rationale leads us to the

conclusion that, in a case such as this one where the record is sparse and finding a

reasonable market rate does not require any credibility determinations, we may settle

on the highest rate justified by the evidence. Here, those rates would be $200 an hour

for Still and $225 an hour for Blacksher.

      The product of the hours totaled by the district court and the rate we settle on

here yields lodestar fees of $80,610 for Still and $63,720 for Blacksher.


                                            14
               C.     Lodestar Adjustment

       Greensboro then argues that even if the plaintiffs are entitled to a fee, the

district court abused its discretion in awarding them a full 70% of the lodestar.9 The

lodestar is adjustable to arrive at a reasonable fee in light of the relation of the results

obtained to the work performed. See Perkins v. Mobile Housing Bd., 

,

739 (11th Cir. 1988). We agree with Greensboro that the plaintiffs’ fee should have

been reduced by a far greater percentage.

       The district court’s approach to adjustment was to examine the effectiveness of

the plaintiffs’ work in removing the obstacles to full relief, and the court credited them

with moving the City from steadfast opposition to acquiescence in a complete remedy.

This approach to adjustment is reasonable, and on this record, we cannot disagree with

the district court’s finding. In the early years of the litigation, the City’s goal was

arguably to thwart a remedy by maintaining black voting strength at the same majority

level that black voters had enjoyed before the admission of § 2 liability. But by June

1994, the City wrote to the court a letter announcing that it was prepared to accept a



   9
        The district court adjusted the fee “above the line” by reducing the number of hours by
30% across the board before calculating the lodestar. In the end, of course, the result is the same
whether you adjust the fee itself or the number of hours spent, and the Supreme Court has
approved of both approaches. See 

, 103 S. Ct. at 1941 (“There is no
precise rule or formula for making these determinations. The district court may attempt to
identify specific hours that should be eliminated, or it may simply reduce the award to account
for limited success.”).

                                                15
court-drawn plan imposing three supermajority black districts, a plan that was

materially similar to the plan the special master ultimately drew. We could attribute

this success at least in part to the plaintiffs’ successful lobbying of the DOJ,10 which

prevented the City from permanently implementing any of the City’s chosen plans.

Perhaps, too, the plaintiffs’ steadfast backing of their plan amounted to a war of

attrition against the City that by 1994 had sapped its resources and resolve.

        But we are troubled by the small scale of the district court’s adjustment, given

that the plaintiffs apparently did not act as litigants in any way that specifically

contributed to their success. In an ordinary action, we can fairly assume that the

plaintiffs had everything to do with their prevailing because it is they who framed the

claims, gathered the evidence, and presented the case. Not so here. The plaintiffs

never got any of the relief that they asked for in court: The hearings and plans

proposed in 1988 did not produce a ruling. The hearing in 1992 produced a result

unfavorable to the plaintiffs because the magistrate judge permitted the City to use a

plan adopted by the city council. The same unfavorable result followed the 1993

hearing. The 1994 hearing yielded adoption of the plaintiffs’ proposed plan, but this

court vacated that order because the plan was constitutionally suspect. In the end, not



   10
        DOJ lobbying time is not inherently noncompensable. See Brooks v. Georgia State Bd.
of Elections, 

, 866 (11th Cir. 1993).

                                             16
only did the City get what it asked for postremand — a plan drawn by the court

without regard to race — but the relief was awarded based on the special master’s own

investigation and work, not the plaintiffs’.

      Thus, in short, the plaintiffs’ work had no articulable value, even as part of a

war of attrition, after the City accepted the possibility of a fully remedial plan in June

1994. The lion’s share of Still and Blacksher’s time was expended after 1994. Still

spent only 100.7 hours, or 20.1% of his total hours, before June 16, 1994, when his

billing records show that he was served the letter from City’s counsel to the court.

Blacksher spent even less time before then; only 0.8% of his total hours predated June

16, 1994.

      The retrospective futility of work does not per se make it noncompensable. But

it does cast doubt on the reasonableness of requiring the other side to pay for it. The

proportion of time spent fruitlessly informs our judgment that Greensboro was due a

considerable reduction in the plaintiffs’ attorney fees, at least half for Still and two-

thirds for Blacksher. We hasten to emphasize that the sort of analysis we have

engaged in was not mandatory; it is simply a way in this case to determine, in a

principled manner, how much of the fee is reasonable. The district court of course had

a range of percentages within its discretion, and it could well have awarded both

lawyers less.


                                           17
        Again, as was the case with the reasonable hourly rate, the question presents

itself whether we must remand for the district court to exercise appropriately

circumscribed discretion, or whether we may finally lay this action to rest. We again

think that rather than prolong the proceedings, we may reduce the lodestar to an

amount that would have been the upper limit of the district court’s discretion.

Reducing the lodestar by the ratios explained above produces a fee of $40,305 for Still

and $21,664.80 for Blacksher.

                                      III. Conclusion

        For the foregoing reasons, we vacate the award of fees and render judgment for

$61,969.80 in attorney fees.11

        VACATED AND RENDERED.




   11
      Our mandate does not disturb the district court’s February 3, 1999 award of certain
expenses.

                                              18