UNPUBLISHED

UNITED STATES COURT OF APPEALS
                FOR THE FOURTH CIRCUIT


SHADOW CREEK APARTMENTS, L.L.C.,      
               Plaintiff-Appellee,
                v.                              No. 01-2212
HARTFORD FIRE INSURANCE COMPANY,
              Defendant-Appellant.
                                      
SHADOW CREEK APARTMENTS, L.L.C.,      
              Plaintiff-Appellant,
                v.                              No. 01-2251
HARTFORD FIRE INSURANCE COMPANY,
               Defendant-Appellee.
                                      
          Appeals from the United States District Court
         for the District of South Carolina, at Greenville.
               Margaret B. Seymour, District Judge.
                         (CA-00-2025-6-24)

                       Argued: June 5, 2002

                     Decided: August 22, 2002

      Before MICHAEL, Circuit Judge, Robert R. BEEZER,
    Senior Circuit Judge of the United States Court of Appeals
         for the Ninth Circuit, sitting by designation, and
      Benson E. LEGG, United States District Judge for the
           District of Maryland, sitting by designation.



Affirmed in part, vacated in part, and remanded by unpublished per
curiam opinion.
2         SHADOW CREEK APARTMENTS v. HARTFORD FIRE INS.
                             COUNSEL

ARGUED: Edward Kriegsmann Pritchard, III, PRITCHARD &
BERLINSKY, L.L.C., Charleston, South Carolina, for Appellant.
James Lafon Rogers, Jr., LEATHERWOOD, WALKER, TODD &
MANN, P.C., Greenville, South Carolina, for Appellee. ON BRIEF:
Paul E. Hammack, LEATHERWOOD, WALKER, TODD & MANN,
P.C., Greenville, South Carolina, for Appellee.



Unpublished opinions are not binding precedent in this circuit. See
Local Rule 36(c).


                              OPINION

PER CURIAM:

   Shadow Creek LLC (Shadow Creek) sued its insurer, Hartford
Insurance Company (Hartford), seeking coverage for rent it says it
lost when several of its apartment buildings, then under construction,
were damaged by fires. A jury returned a verdict for Shadow Creek
in the amount of $500,000, the limit of the lost rent provision of the
insurance policy. The district court denied Hartford’s motion for judg-
ment as a matter of law or for a new trial. The court awarded prejudg-
ment interest to Shadow Creek, but denied Shadow Creek’s motion
for attorneys’ fees. Hartford appeals the denial of judgment as a mat-
ter of law and the award of prejudgment interest, and Shadow Creek
cross-appeals the denial of its motion for attorneys’ fees. For the fol-
lowing reasons, we affirm the district court’s denial of judgment as
a matter of law, vacate the prejudgment interest award and remand for
recalculation, and affirm the denial of Shadow Creek’s motion for
attorneys’ fees.

                                   I.

  The relevant facts of the case are not in dispute. Shadow Creek is
a South Carolina company that owns property in Anderson County,
            SHADOW CREEK APARTMENTS v. HARTFORD FIRE INS.                  3
South Carolina, on which it built an apartment complex consisting of
nine apartment buildings and one free-standing clubhouse. Construc-
tion on the apartment complex began in September of 1998, and the
first building was scheduled to be completed in March of 1999. The
apartment complex was insured by Hartford Insurance Company dur-
ing construction. On January 28, 1999, Shadow Creek purchased a
rider covering lost rent and soft costs for the entire complex with a
total coverage limit of $500,000.1 The lost rent rider provides, in rele-
vant part, that Hartford "will pay for the actual loss of ‘Rental
Income’ you sustain as a result of delay" caused by a covered loss.
It further explained that "[t]he amount of ‘Rental Income’ will be
determined based on: (1) the Net ‘Rental Income’ of similar rental
properties before the direct physical ‘loss’ occurred; [and] (2) the
likely Net ‘Rental Income’ of the property if no ‘loss’ occurred."

   On February 1, 1999, a series of fires burned several of the build-
ings. One of the buildings was completely destroyed, two were
severely damaged, and three were somewhat damaged. The cause was
arson. One of the arsonists, a volunteer fireman, was apprehended and
convicted, and none of the Shadow Creek principals were implicated
in any way. Roughly speaking, the nine apartment buildings were
scheduled to be completed on a staggered basis between March and
August of 1999. Because of the delay caused by the fires, the first
building (Building 9) was not completed until mid-May of 1999 and
the other buildings were completed on a staggered basis mainly from
July of 1999 to January of 2000.2 Hartford reimbursed Shadow Creek
  1
     For the sake of brevity we will refer to this rider as covering "lost
rent," but we acknowledge that it also covers soft costs associated with
such lost rent. Soft costs are costs such as architectural fees and refinanc-
ing costs. The parties stipulated to the amount of soft costs; this case con-
cerns only the amount of lost rent.
   2
     Even though only some of the buildings were damaged by the fires,
construction on all nine buildings was delayed. The buildings’ originally
scheduled completion dates and the actual completion dates after the
fires are as follows:
                 Scheduled completion:         Actual completion:
      Bldg. 1:   March 1, 1999                 January 13, 2000
      Bldg. 9:   March 26, 1999                May 15, 1999
4           SHADOW CREEK APARTMENTS v. HARTFORD FIRE INS.
for much of its loss, and the parties do not dispute the reimbursement
amount except for the amount of lost rent. Hartford calculated the lost
rent at $200,000 and sent Shadow Creek a check in that amount.3
Shadow Creek estimated its lost rent at about $527,000 and submitted
a claim for $500,000, the policy limit. Hartford denied this claim, and
Shadow Creek brought this suit in South Carolina circuit court to
compel payment of the claim. Hartford removed the case to federal
district court based on diversity of citizenship. Both parties agree that
South Carolina law controls.

   After a three-day trial, the jury returned a verdict for Shadow Creek
in the amount of $500,000, the policy limit. The parties had entered
into an agreement whereby any jury verdict on the total amount of
lost rent would be reduced by the $200,000 already paid by Hartford,
so the district court entered judgment in the amount of $300,000. The
court awarded Shadow Creek prejudgment interest at the statutory
rate calculated from February 1, 1999, the date of the fires. Hartford
moved for judgment as a matter of law or for a new trial, and the
court denied this motion. Shadow Creek moved for attorneys’ fees,
arguing that Hartford unreasonably denied the claim. The district
court denied this motion as well. Hartford appeals the denial of judg-
ment as a matter of law and the award of prejudgment interest, and
Shadow Creek cross-appeals the denial of attorneys’ fees.

                                    II.

    Bldg.   8:   April 23, 1999              July 13, 2000
    Bldg.   7:   May 14, 1999                July 10, 1999
    Bldg.   6:   June 11, 1999               July 31, 1999
    Bldg.   4:   July 23, 1999               October 15, 1999
    Bldg.   5:   July 23, 1999               October 2, 1999
    Bldg.   3:   August 13, 1999             December 15, 1999
    Bldg.   2:   August 21, 1999             January 13, 2000
For reasons that are not clear, Building 8 was not completed until July
of 2000, but Shadow Creek attributed only about eight months of this 15-
month delay to the fires. Hartford does not dispute this aspect of the loss
calculation.
   3
     Specifically, Hartford estimated $138,744.86 in lost rent and
$61,255.14 in soft costs, for a total of $200,000. As noted above, the par-
ties dispute only the amount of lost rent.
          SHADOW CREEK APARTMENTS v. HARTFORD FIRE INS.                 5
   We first address Hartford’s argument that the district court erred in
denying its motion for judgment as a matter of law. We review the
district court’s decision de novo, taking the facts in the light most
favorable to Shadow Creek, the non-moving party. Baynard v.
Malone, 

, 234-35 (4th Cir. 2001). "The question is
whether a jury, viewing the evidence in the light most favorable to
[Shadow Creek], could have properly reached the conclusion reached
by this jury." Benesh v. Amphenol Corp. (In re Wildewood Litiga-
tion), 

, 502 (4th Cir. 1995). As noted above, the parties
do not disagree about the relevant facts of this case; rather, they dis-
agree about the meaning of "actual loss of rental income" as applied
to those facts. The disparity between Hartford’s lost rent estimate and
Shadow Creek’s estimate does not stem from disagreements on esti-
mated occupancy rates, rent amounts, or the like, but from the meth-
ods used to calculate the loss. Essentially, Hartford argues that its
method of calculating lost rent was correct as a matter of law given
the facts presented. We will first lay out Shadow Creek’s method of
calculating its loss and then turn to Hartford’s method, which is some-
what more complicated and requires some discussion.

   Shadow Creek calculated its lost rent as follows. Based on stabi-
lized occupancy rates of similar completed buildings in the area,
Shadow Creek assumed a 92 percent occupancy rate. It then calcu-
lated the amount of rent it would have received at 92 percent occu-
pancy for each of the nine buildings during the length of time that the
completion of each building was delayed. Take Building 1 as an
example: a delay of 10 months 13 days multiplied by $11,776 rent per
month (at 92 percent occupancy) equals $122,662.89 in lost rent.

   Hartford’s calculation was significantly different. First of all, Hart-
ford did not calculate the month-by-month loss of rent using a 92 per-
cent occupancy rate. It did not contest the assumption that the
stabilized occupancy rate would likely be about 92 percent. Rather,
it rejected the notion that the stabilized occupancy rate was the rele-
vant occupancy rate for the purpose of calculating this loss. Both par-
ties agreed that the Shadow Creek Apartments, like many apartment
buildings, would not for the most part reach maximum occupancy as
soon as they opened, but would experience a gradually increasing
occupancy rate over a period of months until, at some point, the occu-
pancy rates stabilized at around 92 percent. Thus, for each building
6              SHADOW CREEK APARTMENTS v. HARTFORD FIRE INS.
Hartford calculated Shadow Creek’s lost rent according to Shadow
Creek’s own estimates of its expected occupancy rates — about 35
percent occupancy for the first month opened, 66 percent for the sec-
ond month, and so on. Shadow Creek’s managing member, Woodrow
Wilson, admitted on cross-examination that Shadow Creek would not
have received 92 percent occupancy rates on all nine buildings as
soon as they opened.4 The parties do not dispute that "actual loss of
rental income" means the amount of rental income that Shadow Creek
would have received if the fires had not occurred but did not receive
because of the fires. Because Shadow Creek admits that it would not
have received rental income at a 92 percent occupancy rate for each
of the nine buildings as soon as they opened, we agree with Hartford
that Shadow Creek is overestimating its loss of rental income, at least
during the first months of its loss on each building. There was no evi-
dence presented from which a reasonable factfinder could conclude
that absent the fires, Shadow Creek would have received rental
income at a 92 percent occupancy rate for each building as soon as
it opened. Rather, Hartford is correct to use Shadow Creek’s esti-
mated occupancy rates, which gradually increased over time towards
occupancy stabilization. We will return to this conclusion later in the
opinion to address its impact on our ultimate holding. For now, we
turn to the other important aspect of Hartford’s lost rent calculation.

   In addition to using a gradually increasing occupancy rate instead
of a 92 percent rate from day one, Hartford also ended any recovery
for lost rent as of July 31, 1999, even though the apartment buildings
opened on a staggered basis roughly from May of 1999 to January of
    4
     The exchange was as follows:
          Q: But had there been no fire would you have leased those
        buildings at a ninety-two percent occupancy rate the day they
        opened up?
           A: [Wilson]: The first three buildings would have been prob-
        ably in that range, and then it would have dropped off through
        the rest of the property.
           Q: But you haven’t done that [in your calculation] there,
        have you? You’ve [calculated] it for all nine buildings [at]
        ninety-two percent the day it would have opened?
          A:   [Wilson]: That’s correct.
           SHADOW CREEK APARTMENTS v. HARTFORD FIRE INS.                 7
2000. Hartford reasoned as follows. When the first building (Building
9) opened in May, it was fully occupied immediately. Likewise, when
the second building (Building 7) opened, it was fully occupied.5
When the third building (Building 6) opened on July 31, 1999, it was
not completely occupied. As the other buildings were completed,
from July 31 on there were always available rental units in at least
one of the completed buildings.6 Hartford argues that as soon as
Shadow Creek had available rental units that were sitting vacant, as
a matter of logic it suffered no further loss of rental income. The fact
that there were apartments available for rent conclusively demon-
strates, Hartford argues, that even if the other buildings had been up
and running as initially scheduled, Shadow Creek would not have
been able to rent any more units than it actually rented. The idea is
that the presence of vacant rental units demonstrates that there were
no renters who would have rented but for the arson — if they would
have rented, they would simply have filled up the vacant units. Thus,
Hartford argues, Shadow Creek did not lose rental income after July
31, 1999, because of the arson.

   Hartford claims that this method of calculating Shadow Creek’s
lost rent is correct as a matter of law because it follows necessarily
from the meaning of "actual loss of rental income" in the policy. The
principle on which Hartford relies — that supply in excess of demand
precludes any claim for lost income — does have support in the case
law. For example, in one case a cement manufacturing plant was shut
down by a fire. Northwestern States Portland Cement Co. v. Hartford
Fire Ins. Co., 

(8th Cir. 1966). The plant produced
  5
     The parties agreed that, generally speaking, occupancy rates would
start low and gradually increase, as described above. However, they also
agreed that the first one or two buildings to open would probably be
immediately filled, as there would be an initial period of high demand.
As Wilson put it, "It’s not uncommon in the first building if you’ve done
pre-leasing to be able to have that building leased out or at least ninety-
two [percent occupancy] — that is not uncommon. . . . The rest of the
complex we didn’t project it to be leased out [at full occupancy immedi-
ately]."
   6
     There was a period between August and December of 1999 when no
one-room units were available, so Hartford included payment for lost
rent on one-room units during that period.
8         SHADOW CREEK APARTMENTS v. HARTFORD FIRE INS.
clinker, an ingredient used to make cement. The business had a large
stockpile of both finished cement and of clinker, so it was able to
meet all customer demand while the factory was down. 

the company had inventory in excess of customer demand,
the Eighth Circuit held that it had lost no income because of the fire
and was not entitled to insurance proceeds for lost profits.7 Id.; see
also Lyon Metal Prod., LLC v. Protection Mutual Ins. Co., 

, 505 (Ill. App. Ct. 2001) ("[I]f there is enough inventory
to meet the projected sales for the period of interruption there is no
loss of earnings.").

   The principle that a plaintiff cannot recover for lost profits when
it has inventory in excess of demand is based on several assumptions.
Among other things, the principle asks us to assume that the loss of
inventory has not affected consumer demand. If the loss reduced not
only supply but also consumer demand, then of course we could not
assume that the level of demand after the loss occurred reflects what
demand would have been absent the loss. In Northwestern States
Portland Cement there was no reason to think that customer demand
for cement was lower because of the fire at the plant. Thus it was rea-
sonable to assume that had the fire not occurred, Portland Cement
would not have sold more cement. In this case, Hartford’s argument
rests on the assumption that the actual demand for apartments in the
Shadow Creek complex on July 31, 1999, was the same as what the
demand would have been on July 31 absent the fires.

   The uncontested facts in this case render this assumption question-
able to say the least. As explained above, both parties agree with the
general proposition that the Shadow Creek apartment complex would
not immediately reach its stabilized occupancy rate. Rather, as each
building opened, occupancy rates would slowly increase until at some
point the rate stabilized at around 92 percent. At oral argument,
Shadow Creek referred to the time between the initial opening of an
apartment building and the point at which occupancy stabilizes as the
"ramp-up period," and we think this is as good a term as any to
    7
   There is a difference between coverage for lost profits and coverage
for lost rent. A lost profits provision covers income less expenses,
whereas a lost rent provision simply covers income. This difference is
not relevant to our analysis, so we discuss the two interchangeably.
          SHADOW CREEK APARTMENTS v. HARTFORD FIRE INS.                 9
describe this period. Presumably, the ramp-up period exists because
demand for apartment units is spread out over time. Over the course
of a year, there may be 200 people interested in renting a Shadow
Creek apartment, but not all 200 will want to move in on the first day
the apartments are available. Thus, when the apartment buildings
were delayed in opening, the ramp-up period — the period of gradu-
ally rising occupancy rates — was also delayed. Because of the delay
caused by the fire, on July 31 Building 6 was experiencing the begin-
ning, low-occupancy stage of the ramp-up period. If Building 6 had
opened on June 11, as planned, by July 31 it would have been seven
weeks further along in the ramp-up period. In August, instead of get-
ting rents at the Month 3 occupancy rate, Shadow Creek got rents at
the lower Month 1 occupancy rate. This type of loss, caused by delay-
ing the opening of a business, was recognized by the Georgia Court
of Appeals in Chronicle Bldg. Co. v. New Hampshire Fire Ins. Co.,

(Ga. Ct. App. 1918). In that case, the court observed that
"losses [in the form of lost rent] . . . might accrue by reason of neces-
sary delay attendant upon the retenanting of the building subsequently
to such period after its actual reconstruction." 

court
recognized that when an apartment building is damaged and must be
rebuilt, the owner will probably lose rent not only while the apartment
building is being reconstructed, but also after it opens during the
period of time it takes to return the building to its stabilized occu-
pancy rate.8 In this case, a reasonable jury could have found that the
demand for Shadow Creek’s apartments on July 31 and thereafter was
lower than it would have been absent the fires because the fires
delayed the ramp-up period. Thus the district court properly denied
Hartford’s motion for judgment as a matter of law seeking to cut off
all recovery after July 31, 1999. On the facts presented, a reasonable

  8
   The court in Chronicle Building, while recognizing that the insured
might actually suffer lost rent in this manner, denied coverage because
the policy expressly limited lost rent recovery to "the necessary period
of reconstruction." Thus, losses occurring after reconstruction was fin-
ished, whether actually incurred by the insured or not, were not covered
under the policy. 

similar to Chronicle Building except
that the policy here contains no language limiting actual lost rent to the
time period during construction.
10         SHADOW CREEK APARTMENTS v. HARTFORD FIRE INS.
jury could determine that Shadow Creek continued to suffer loss of
rental income even after it had vacant units available for rent.9

   We concluded above that Shadow Creek’s method of calculating
its loss resulted in overestimating its loss during the period of delay
for each building. This is because Shadow Creek assumed a 92 per-
cent occupancy rate immediately for all buildings, which Shadow
Creek admits would not have actually happened. Rather, the evidence
supports the conclusion that Shadow Creek lost rent at a lower
amount per month than Shadow Creek claimed. But both Hartford
and Shadow Creek are underestimating Shadow Creek’s loss of rental
income after each building opened — Hartford by cutting off loss as
soon as rental units were available in any unit, and Shadow Creek by
not claiming any lost rent on each building after that building opened.
The evidence supports the conclusion that Shadow Creek’s losses
continued after each building was completed. This is because the
effect of the delay in the ramp-up period continued even after each
building had opened. For example, Building 6 opened July 31. By
September, Shadow Creek was still losing rental income on Building
6 even though it had been open for over a month. In September Build-
ing 6 had Month 2 occupancy rates when without the fires it would
have had Month 4 occupancy rates. Not only did Shadow Creek lose
rental income on each building during that time before each building
opened, but it continued to lose rental income after each building
opened. The losses ceased only when the occupancy rate stabilized at
92 percent. It is only at this point that one can confidently say that
  9
    Hartford’s position rests on yet another assumption that is question-
able in this case, namely that the goods in question are fungible. The
principle assumes that customers will be content to receive the excess
inventory in place of the inventory that was destroyed. But apartments
are not generally considered fungible. If a potential lessee had his heart
set on a third-floor, south-facing apartment, he may not be willing to rent
a vacant apartment on the first floor that faces north. The vacancy in the
first-floor north-facing apartment does not necessarily imply that there is
no demand for the third-floor south-facing apartment. Hartford claims
that these particular apartments were interchangeable and that Shadow
Creek presented no evidence to the contrary. We need not resolve this
issue because Hartford’s argument fails on the independent ground, dis-
cussed above, that the loss can reasonably be said to have reduced
demand by delaying the ramp-up period.
          SHADOW CREEK APARTMENTS v. HARTFORD FIRE INS.                11
Shadow Creek was renting its units at the same rate that it would have
been renting them absent the fires. Accordingly, while we observed
above that Shadow Creek is overestimating its actual losses during the
period of each building’s delay, Shadow Creek is likewise underesti-
mating its actual losses during the period following each building’s
completion.

   Interestingly, Shadow Creek’s method of calculating the loss and
the method we have described above are equivalent — they yield the
exact same loss amount. That is, Shadow Creek would have lost the
same amount of money whether it lost several month’s rent on all of
its apartments operating at their stabilized occupancy rates or whether
the apartments were delayed the same number of months during con-
struction. In either case, Shadow Creek loses an amount equal to its
revenue at its average maximum capacity for the period of the shut-
down.10 The difference is that when the shutdown occurs at the open-
ing of a business rather than after the business is established, the loss,
while equal in amount, is simply spread out over a longer period of
time. Thus, even though we agree with Hartford that the evidence
does not support the conclusion that Shadow Creek lost rent at a 92
percent occupancy rate in the first months, we are able to affirm the
  10
     Showing how this principle works using one of the actual buildings
in this case would involve irregular figures and thus would not provide
a clear illustration. For clarity’s sake, we use imaginary Building A as
an illustration. Assume that Building A was delayed by three months,
and that Building A would experience occupancy rates of 25 percent the
first month, 50 percent the second month, 75 percent the third month,
and 100 percent the fourth month and thereafter. If Building A has $4000
in rent at full occupancy, then the lost rent in month one would be $1000,
since it expected 25 percent occupancy ($1000 in rent) but had zero
occupancy because of the delay. The losses for the following months
would be: $2000 in month two, $3000 in month three, $3000 in month
four, $2000 in month five, $1000 in month six, and no loss in month
seven. The loss levels off in month four and declines thereafter because
by month four Building A has opened and is being occupied, just at a
lower level than it would have absent the delay. These figures, added
together, yield a total loss of $12,000 over six months. This is the same
total loss as the total loss assuming full occupancy for the three month
period of delay: $4000 x 3 = $12,000.
12        SHADOW CREEK APARTMENTS v. HARTFORD FIRE INS.
district court’s denial of Hartford’s motion for judgment as a matter
of law.11

                                  III.

   Apart from the court’s denial of its motion for judgment as a matter
of law, Hartford appeals the district court’s calculation of prejudg-
ment interest. Shadow Creek, for its part, appeals the district court’s
denial of its motion for attorneys’ fees. We address these issues in
turn.

                                   A.

   The district court awarded Shadow Creek prejudgment interest
starting on February 1, 1999, the date of the fires. Hartford argues that
even assuming that the damages award is justifiable, the district court
should not have calculated interest starting at the date of the fires.
First, Hartford claims that interest does not begin to accrue on the
date of the loss, but rather 60 days after the loss. Second, even if
interest begins on the date of loss rather than 60 days after the loss,
Hartford argues that Shadow Creek’s loss of rental income did not
occur on the date of the fires, but rather on the date that each building
was scheduled to open. We address these contentions in turn.

   Hartford is correct that ordinarily prejudgment interest does not
begin on the date of loss but only after some reasonable period (usu-
ally 60 days) following the loss. This 60 days represents the time that
the insurer is given to investigate the loss and make a coverage deter-
  11
    Because these two methods of calculating lost rent result in the same
total figure, it may seem that we are being hyper-technical to draw the
distinction. The method we have described for calculating lost rent, how-
ever, is necessary to make sense of the competing claims in this case (in
addition to simply being technically correct). Our method explains why
Shadow Creek’s ultimate loss figure is correct even though Hartford is
obviously right when it points out that Shadow Creek did not, in fact,
lose rental income at a 92 percent occupancy rate during the first few
months that each building would have been open. While Shadow Creek
did not lose rent in that manner, the amount of its actual loss of rental
income is the same as if it had.
           SHADOW CREEK APARTMENTS v. HARTFORD FIRE INS.                 13
mination. However, in Flynn v. Nationwide Mut. Ins. Co., 

, 821 (S.C. App. Ct. 1984), the South Carolina Court of Appeals
held that when the insurance company denies a claim and must be
sued before it will pay, the court will compute interest starting on the
date of loss, not on the date that the payment was demandable by the
insured. Likewise, in Varnadore v. Nationwide Mut. Ins. Co., 

, 715 (S.C. 1986), the South Carolina Supreme Court cited
Flynn with approval and affirmed the trial judge’s computation of
interest starting at the date of the loss. Accordingly, as we read South
Carolina law, prejudgment interest for an award of insurance cover-
age is computed from the date of the loss, absent contractual language
to the contrary.12

   Even assuming that prejudgment interest accrues from the date of
the loss rather than 60 days after the loss, Hartford argues that the dis-
trict court erred in determining when the loss actually occurred. As
stated above, the district court set the date of loss as February 1, 1999,
the date of the fires. Hartford argues that the lost rent (which is the
only loss at issue here) did not occur until, at the very earliest, the
date when the apartment buildings would have opened absent the
fires. Hartford argues that interest should begin running on each
building based on the date when each would have opened (i.e., Build-
ing 1 on March 1, 1999, Building 9 on March 26, 1999, and so on).
We agree with Hartford on this point. Shadow Creek did not suffer
  12
    Hartford cites Jacobs v. American Mut. Fire Ins. Co. of Charleston,

(S.C. 1986), for the proposition that interest begins 60
days after the loss occurs rather than on the date of the loss. Indeed, in
Jacobs the South Carolina Supreme Court did affirm the trial court’s
computation of interest starting 60 days after the loss occurred. 

in Jacobs only the insurer appealed the interest computation.
The Court rejected the insurer’s appeal, noting that the trial court’s com-
putation was actually "more favorable to the Insurer than the rule enunci-
ated in Flynn." 

the only error in the trial court’s
ruling benefitted the insurer, and because only the insurer appealed, the
Supreme Court simply affirmed the trial court’s judgment. If the insured
had cross-appealed the interest award, the Supreme Court would
undoubtedly have corrected the trial court’s error and instructed that the
interest be calculated from the date of the loss. Jacobs, like Varnadore,
clearly adopts the rule articulated in Flynn, namely that interest is calcu-
lated from the date of loss.
14         SHADOW CREEK APARTMENTS v. HARTFORD FIRE INS.
a loss of rental income on February 1, 1999, the date of the fires.
Instead, Shadow Creek’s loss of rental income for each building did
not begin until each building was originally scheduled to open.
Accordingly, we vacate the district court’s prejudgment interest
award and remand for recalculation. On remand, the district court
should calculate interest on a building-by-building basis, using the
scheduled opening date of each building as the date on which Shadow
Creek’s loss of rental income for that building occurred.13

                                    B.

   Finally, Shadow Creek appeals the district court’s denial of its
motion for attorneys’ fees. Attorneys’ fees are available in South Car-
olina when the insurer denies a claim in bad faith or without reason-
able cause. S.C. Code Ann. § 38-59-40. Shadow Creek primarily
takes issue with a several month delay in reconstruction that it says
was caused by unreasonableness on the part of Hartford. Hartford’s
adjustor, Sanford, maintained for several months that the concrete
slabs beneath Buildings 1 and 2 were repairable. Finally, after several
months he conceded that they were not repairable and that the build-
ings had to be completely rebuilt. Shadow Creek presented Sanford’s
own computer log as evidence of unreasonableness. In the log San-
  13
    Actually, Shadow Creek did not lose all of its rental income for each
building on the first day that each building was scheduled to open.
Rather, it lost rental income on a month-by-month basis, as it would have
received the rent. For example, on Building 1 Shadow Creek lost one
month’s rent at 35 percent occupancy for March, one month’s rent at 66
percent occupancy for April, and so on. Nonetheless, we will not instruct
the district court to separate the loss out on a month by month basis for
the purpose of calculating interest separately for each month of lost rent.
While Hartford contested the February 1 date of loss, it requested (both
in its brief and in oral argument) only that the loss be allocated to the
date of each building’s scheduled opening, not that it be separated out for
each month. Calculating the interest separately for each month of rent
would undoubtedly be quite complicated and would not yield a figure
significantly lower than calculating the interest for each building from
the date that building opened. Because Hartford requests only that the
interest on each building’s rent be calculated from the date that building
was scheduled to open, the district court is free to do just that.
          SHADOW CREEK APARTMENTS v. HARTFORD FIRE INS.              15
ford apparently had noted as early as a week after the fires that these
slabs were "totaled."

   Even if we accept Shadow Creek’s version of events, it is not enti-
tled to an award of attorneys’ fees here because the attorneys’ fees
were not expended as a result of the allegedly unreasonable behavior.
Hartford paid to have the buildings rebuilt, and Shadow Creek did not
have to sue them to recover for the rebuilding. Thus, Shadow Creek
incurred no attorneys’ fees that are in any way related to the delay
attributable to Sanford’s behavior. The only question with regard to
attorneys’ fees is whether Hartford’s denial of the lost rent in excess
of $200,000 was in bad faith or was unreasonable, because that is the
only issue over which any attorneys’ fees were expended. See Taylor
v. Nix, 

, 622 (S.C. 1992) (statutory attorneys’ fees can-
not be recovered for fees spent on matters unrelated to the claim for
which a fee award is authorized). While we have affirmed the judg-
ment against Hartford, its reason for denial of the extra lost rent was
both factually and legally supportable. We therefore affirm the district
court’s denial of attorneys’ fees.

                                  IV.

   In sum, we affirm the district court’s denial of Hartford’s motion
for judgment as a matter of law or for a new trial. We vacate the dis-
trict court’s prejudgment interest award and remand for recalculation.
Finally, we affirm the district court’s denial of Shadow Creek’s
motion for attorneys’ fees.

                                    AFFIRMED IN PART, VACATED
                                       IN PART, AND REMANDED