In the

    United States Court of Appeals
                 For the Seventh Circuit
                     ____________________
No. 18-2169
UNITED STATES OF AMERICA,
                                                   Plaintiff-Appellee,
                                 v.

ISHAIHU HARMELECH,
                                               Defendant-Appellant.
                     ____________________

         Appeal from the United States District Court for the
           Northern District of Illinois, Eastern Division.
             No. 15-CR-724 — Elaine E. Bucklo, Judge.
                     ____________________

    ARGUED NOVEMBER 28, 2018 — DECIDED JUNE 24, 2019
               ____________________

   Before ROVNER, HAMILTON, and BRENNAN, Circuit Judges.
    BRENNAN, Circuit Judge. A federal grand jury indicted
Ishaihu Harmelech on two counts of mail fraud under 18
U.S.C. § 1341. Harmelech pleaded guilty to the first count, and
the government dismissed the remaining count. In pleading
guilty, Harmelech, who owned and operated multiple cable
installation companies, admitted to setting up hundreds of
DIRECTV accounts under a fraudulent scheme and pocketing
the money that should have been paid for servicing those
2                                                  No. 18-2169

accounts. He now appeals his sentence, arguing the district
court erred in calculating DIRECTV’s losses and in applying
a four-level sentencing enhancement pursuant to Sentencing
Guideline § 3B1.1(a). Because we see no error in the district
court’s loss calculation and sentencing determination, we af-
firm.
                        I. The Scheme
    Between 2005 and 2011, Harmelech owned and operated
three companies that installed cable television services at sin-
gle-family residences and in multi-dwelling unit properties,
such as hospitals, nursing homes, motels, senior living facili-
ties, and apartment buildings. Harmelech managed nine
different employees across his three companies, and he was
responsible for the day-to-day operations. He continuously
held himself out as an authorized dealer and distributor of
cable installation services, despite his authorization having
lapsed in 2005. As an “authorized” dealer, Harmelech con-
tracted directly with cable providers—including DIRECTV—
to install receivers in over 150 multi-dwelling unit properties.
    Unlike single-family residences, multi-dwelling units typ-
ically have one master antenna system with multiple cable re-
ceivers placed in a rack called a “headend.” Each receiver in
the headend is tuned to one channel, and the signal is then
distributed throughout the building. Each television set in the
building may access multiple channels through the headend.
For multi-dwelling unit buildings, it is usually the building
owner or manager (rather than individual unit residents) who
contracts with a cable provider for the entire building’s ser-
vice. A multi-dwelling receiver is generally more expensive
than the same receiver installed in a single-family residence
because it provides service to the entire building.
No. 18-2169                                                    3

    DIRECTV provides satellite television programming to
single-family and multi-dwelling customers by installing a
satellite receiver on the customer’s property and charging a
subscription fee. Each receiver bears a unique identification
code, which helps tie that receiver to an individual account.
The cost of a subscription depends on the type of account (sin-
gle-family or multi-dwelling) and the channels provided. As
is standard industry practice, DIRECTV charges more for
multi-dwelling subscriptions than for single-family subscrip-
tions. This rate structure reflects the different installation and
service requirements for multi-dwelling unit buildings de-
scribed above.
    Authorized dealers and distributors routinely contract
with DIRECTV to install receivers in multi-dwelling unit
buildings. Despite lacking authorization, Harmelech installed
DIRECTV receivers and exploited the multi-dwelling receiver
configuration to commit fraud. Harmelech misrepresented
his authority to customers to gain their trust and access their
personal information to set up fraudulent accounts, as well as
used fictitious names, to open approximately 384 single-fam-
ily residential accounts. Once a fraudulent single-family ac-
count was created, he would place the receiver meant for that
account into the headend at a multi-dwelling building, tuning
the receiver to a channel not already included in the build-
ing’s subscription—usually a premium channel with a higher
price. Harmelech would then charge the building customer
for the fraudulently-installed channel as part of a multi-dwell-
ing subscription. The building customer would pay
Harmelech directly for all channels installed under a multi-
dwelling subscription, despite being billed by DIRECTV
(through Harmelech) at a single-family rate. Harmelech
would then pay DIRECTV the lower-billed rate, pocketing the
4                                                   No. 18-2169

difference between what was charged and what the building
customer paid. Ultimately, Harmelech’s scheme caused
DIRECTV to provide multi-dwelling buildings with channels
for which neither Harmelech nor the customers were paying.
   This scheme continued for over six years and involved
several other participants who acted at Harmelech’s direction.
For instance, Harmelech’s secretary assisted him in opening
fraudulent accounts and used her personal credit card to
make payments for those accounts on Harmelech’s behalf.
Another employee aware of the fraud also opened fraudulent
accounts for Harmelech and installed single-family receivers
in multi-dwelling buildings under Harmelech’s control.
Harmelech also directed the activities of seven other employ-
ees and convinced four separate companies that were author-
ized DIRECTV dealers to provide additional receivers to
increase the number of fraudulent accounts he could open.
   In August 2009, suspecting Harmelech was committing
fraud, DIRECTV opened an internal investigation, and hired
an outside firm, Signal Audit, to assist. As part of the investi-
gation, DIRECTV and Signal Audit attempted to locate and
access the headend at each multi-dwelling building that
Harmelech serviced. Once inside a building, investigators in-
spected each receiver in the headend to determine which type
of account (single-family or multi-dwelling) was associated
with the unique identification code on the receiver. This al-
lowed investigators to determine whether receivers tied to
single-family accounts were being used to provide program-
ming for the entire building.
   DIRECTV and Signal Audit were able to inspect five multi-
dwelling buildings in the Chicago area. For those five build-
ings, investigators identified the account type tied to each
No. 18-2169                                                  5

receiver in the headends and found the receivers did not cor-
respond to the buildings’ DIRECTV accounts. Investigators
also obtained from each building manager a list of the
channels the building’s residents received. The investigation
produced no evidence that residents were aware of any fraud
occurring on their accounts.
    Once Harmelech learned of the investigation, he in-
structed the building managers to not cooperate and directed
his employees to remove receivers from the remaining build-
ings he serviced before they could be inspected. Harmelech
eventually stopped making payments to DIRECTV on behalf
of all buildings with fraudulent accounts, causing hundreds
of accounts to become delinquent. Within three to four
months, DIRECTV terminated programming for all accounts
serviced by Harmelech.
    Although Harmelech prevented DIRECTV from inspect-
ing all impacted buildings, the five buildings it did inspect
were used to determine the quantifiable value of its losses. At
sentencing, a DIRECTV representative testified its investiga-
tors compared the large numbers of channels the five building
managers reported that residents actually received with the
smaller numbers of channels for which DIRECTV received
subscription payments. This comparison showed DIRECTV
should have been paid $2,006 per building per month for the
full programming provided. In its internal calculations,
DIRECTV credited Harmelech for all monthly payments
made on behalf of the five buildings, resulting in a loss to
DIRECTV of approximately $1,477 per month per building.
Altogether, DIRECTV’s investigation revealed an average an-
nual loss in programming costs of approximately $20,000 for
each of the five buildings.
6                                                 No. 18-2169

               II. District Court Proceedings
    Harmelech pleaded guilty to defrauding DIRECTV and, in
his plea declaration, acknowledged the charge of mail fraud
carries a maximum sentence of 20 years’ imprisonment.
Harmelech also agreed to pay “any other penalties or restitu-
tion” imposed by the judge, but he did not stipulate to a res-
titution amount.
    At sentencing, the court and the parties discussed how to
calculate DIRECTV’s losses. Harmelech argued he should be
credited with all payments made, and he should not owe any
restitution because he brought customers to DIRECTV. Alt-
hough he admitted DIRECTV was not paid what it should
have been because of his fraudulent scheme, Harmelech
claimed his scheme benefited the company, and any losses
suffered were a result of additional business he brought. In-
stead of proposing a calculation method, Harmelech argued
DIRECTV suffered zero losses.
    The government, in contrast, proposed two different loss
calculation methods. The first was an estimate of the value of
the programming Harmelech’s customers received, but for
which DIRECTV was never paid. That required taking the
average monthly cost of the stolen channels from the five in-
spected buildings and multiplying it by 150, the number of
multi-dwelling buildings that Harmelech serviced. This cal-
culation yielded a total loss of approximately $3,511,917 a
year, or $21,071,502 for years 2005-2010. The government con-
ceded, though, that “[w]hile the value of those stolen channels
would best capture the true loss suffered by DIRECTV,” the
$21 million estimate was speculative because it was based on
a small sample size of only the five inspected buildings.
No. 18-2169                                                              7

   The second loss calculation was based on actual losses
from identified fraudulent accounts. Under this method, the
government identified the fraudulent accounts and calculated
the losses associated with those accounts based on three cate-
gories: (1) the cost of account delinquencies for fraudulent ac-
counts that became delinquent when Harmelech stopped
making payments; (2) the value of DIRECTV receivers that
were not returned to DIRECTV when it terminated service on
fraudulent accounts; and (3) the value of promotional credits
that DIRECTV gave to fraudulent accounts that would not
have otherwise qualified for such credits but for Harmelech’s
fraudulent scheme. This second calculation did not consider
the value of the stolen channels; it focused only on account
delinquencies, lost equipment, and promotional credits, mak-
ing it conservative even in Harmelech’s view.
    The district court rejected Harmelech’s position, charac-
terizing the relevant question as whether DIRECTV was paid
what it should have been for the services provided. The dis-
trict court adopted the government’s second, conservative
loss calculation and found Harmelech owed: (1) $108,000 in
account delinquencies; (2) $39,000 in unrecovered DIRECTV
receivers; and (3) $29,600 in promotional customer credits.
The court also assessed $166,0001 in stolen channels for the
five inspected buildings and $35,000 for the price DIRECTV


    1 To accurately calculate DIRECTV’s known programming losses, the
district court began with the $20,000 baseline estimate for annual pro-
gramming costs per account. It then considered the account opening dates
for each of the five buildings; the longer an account was open, the greater
the programming costs. Based on the age of each account, the district court
estimated the total programming loss for the audited buildings as
$166,000.
8                                                            No. 18-2169

paid to Signal Audit to assist in the internal investigation. In
sum, the district court ordered $372,600 in restitution.2
    After calculating DIRECTV’s losses, the court considered
Harmelech’s sentencing range. The court assessed a four-level
sentencing enhancement for Harmelech’s role as the organ-
izer and leader of an otherwise extensive fraudulent scheme.
U.S.S.G. § 3B1.1(a). Combined with his category II criminal
history, the sentencing range came out between 46 to 57
months. The court sentenced Harmelech to 48 months.
    While Harmelech conceded he acted as the manager of the
scheme, he argued he should have received only a two-level
enhancement under the Guidelines because the scheme
involved fewer than five participants3 and was not otherwise
extensive. U.S.S.G. § 3B1.1(c). In response, the government
conceded they had identified only two employees who knew
of the scheme, but that it was otherwise extensive based on
several factors: (1) it spanned six years and involved 384
fraudulent accounts across 150 multi-dwelling buildings; (2)
the number of people Harmelech needed to maintain the
scheme—nine employees, four separate authorized dealers,
DIRECTV representatives, building owners and managers,
and hundreds of unknowing residents; and (3) the dishonest
conduct Harmelech displayed in using unknowing

    2The district court’s calculation was $5,000 less than what the figures
add up to, but neither party noticed or corrected the district court’s math-
ematical error at sentencing. Because the error did not affect the court’s
sentencing calculation under the Guidelines, and because neither party
seeks to correct the court’s error on appeal, we accept the restitution
amount as calculated.
    3See U.S.S.G. § 3B1.1, cmt. n.1 (defining a “participant” as someone
who is “criminally responsible for the commission of the offense”).
No. 18-2169                                                              9

participants to further his scheme, including in some cases
stealing their identities.
    On appeal, Harmelech pursues two principal arguments:
(1) the district court erred in calculating DIRECTV’s loss
amount when it declined to assess zero losses; and (2) the
district court erred in applying a four-level sentencing en-
hancement instead of a two-level enhancement.4 For the rea-
sons below, we reject these arguments and affirm the district
court.
                        III. Loss Calculation
    We review the district court’s loss calculation for clear er-
ror. United States v. Rosen, 

, 1024 (7th Cir. 2013).
Clear error exists only when we are “left with the definite and
firm conviction that a mistake has been made.” United States
v. Vivit, 

, 914 (7th Cir. 2000). The loss calculation
must be “not only inaccurate but outside the realm of permis-
sible computations.” See United States v. Gumila, 

,
834 (7th Cir. 2018) (quoting United States v. Littrice, 

, 1060 (7th Cir. 2012)).
    Harmelech fails to make such a showing. He does not ar-
gue the court’s calculation was inaccurate or outside the realm
of permissible computations; indeed, he admits that the
calculation is accurate and even produces a conservative esti-
mate of DIRECTV’s losses. Rather, Harmelech objects to


    4 Harmelech’s brief also notes his opposition to the inclusion of evi-
dence at sentencing showing his personal finances. The district judge
noted she had either not seen the relevant evidence or it had not influ-
enced her calculations. Because neither party asserts this evidence was the
reason for the district court’s loss and sentencing determinations, we do
not address it on appeal.
10                                                   No. 18-2169

calculating any loss at all. He argues that all payments he
made to DIRECTV on the fraudulent accounts should be
“fully credited” to him—without further calculation—be-
cause “[l]oss cannot include the value of services a defendant
legitimately performed for the victims of his fraud.” United
States v. Swanson, 

, 513 (7th Cir. 2007).
    This reasoning overlooks a critical step in analyzing loss.
“Nominally legitimate payments are not offset against in-
tended loss when they are ‘intertwined with and an ingredi-
ent of [an] overall fraudulent scheme.’” United States v. Stochel,

, 890 (7th Cir. 2018) (quoting United States v.
Marvin, 

, 665 (7th Cir. 1994)). When reviewing a
loss calculation this court does not credit payments made in
furtherance of the scheme. See Stochel, 901 F.3d at 890 (holding
that defendant should not receive a credit against loss for pay-
ments covering legitimate expenses because the payments
made were “essentially the cost of perpetuating the scheme,”
“designed to lull his victims so he could avoid detection.”).
We decline to adopt Harmelech’s position which would re-
quire us to skip this step in our review of the district court’s
loss calculation.
    While Harmelech admits he made payments in further-
ance of the scheme, he claims the payments were nevertheless
“legitimate” because DIRECTV financially benefitted. We
again reject his argument. In United States v. Lane, 

, 585 n. 4 (7th Cir. 2003), we declined to credit “gains made
by successful investors in a fraudulent investing scheme, as
those gains [were] only intended to lure and defraud other
investors.” And in Stochel, 901 F.3d at 890, we refused to credit
a defendant who “stole receivership funds and covered his
tracks with money from other sources for the purpose of
No. 18-2169                                                    11

throwing the [victims] off his scent and keeping the scam
alive.”
    Notably, in Stochel, we relied on the district court’s finding
that the defendant was not entitled to any credit “for the value
of the services he provided” regardless of the “substantial fi-
nancial benefit” the victims gained. Id. Here, the district court
did credit Harmelech for the value of the payments he made
to DIRECTV over the course of the scheme (based on the fig-
ures produced by the internal investigation), but those pay-
ments were less than the full amount owed. That was, in fact,
the entire structure of Harmelech’s scheme: make less-than-
full payments to skim money off the top without DIRECTV
noticing. Harmelech now asks this court to apply an unprec-
edented and extraordinary remedy, beyond what we contem-
plated in Stochel. He asks the court to find not only that his
payments were legitimate (as they may be credited without
being legitimate), but also that they were sufficient to fully
compensate DIRECTV for its losses such that no further loss
calculation is required. Harmelech’s request conflicts with ap-
plicable case law and fails to acknowledge the extent of the
harm he caused DIRECTV.
    Here, the district court calculated DIRECTV’s losses by as-
sessing whether “DIRECTV was getting paid what it should
have been paid.” In calculating this amount, the parties pre-
sented the court three alternative theories: (1) Harmelech’s
theory that DIRECTV’s loss was zero; (2) the government’s
first proposal calculating losses for all impacted buildings by
projecting costs across the board based on a sample size of
five buildings; and (3) the government’s second proposal add-
ing together the concrete losses for all known fraudulent
accounts.
12                                                  No. 18-2169

    The court chose option three, plus two other known losses:
the value of the stolen channels (for just the five inspected
buildings) and the cost of the internal investigation. By both
parties’ admissions, the court’s calculation was conservative.
It did not assess speculative losses, but instead focused on
specific, concrete loss figures: $108,000 in account delinquen-
cies; $39,000 in unrecovered DIRECTV receivers; $29,600 in
promotional customer credits; $166,000 in stolen channels for
five buildings; and $35,000 to compensate for the internal in-
vestigation costs.
    Because the district court’s loss calculation was concrete,
specific, conservative in its results, and consistent with this
court’s existing precedent, we see no error in the calculation.
We further conclude the calculation was neither inaccurate
nor outside the realm of permissible computations—indeed,
it was one of three alternative options the parties presented to
the court, and each component of the court’s chosen figure ac-
counted for a known and quantifiable loss.
IV. Organizer or Leader of “Otherwise Extensive” Scheme
    Under the Sentencing Guidelines, a four-level enhance-
ment is applied to a defendant’s offense level when the de-
fendant is an “organizer or leader” of criminal activity that
“involved five or more participants or was otherwise exten-
sive.” U.S.S.G. § 3B1.1(a). If the defendant is an organizer or
leader of criminal activity with fewer than five participants
and the activity is not considered otherwise extensive, then a
two-level enhancement applies. U.S.S.G. § 3B1.1(c). On these
issues, we review the district court’s factual findings for clear
error and legal conclusions de novo. See Stochel, 901 F.3d at
888; see also Rosen, 726 F.3d at 1024 (reviewing defendant’s
role as leader or manager for clear error); United States v.
No. 18-2169                                                     13

Hussein, 

, 156, 162 (7th Cir. 2011) (reviewing “oth-
erwise extensive” element for clear error); United States v.
Arojojoye, 

, 737 (7th Cir. 2014) (reviewing whether
§ 3B1.1 applies de novo).
    Harmelech admits he directed his fraudulent scheme, but
claims it was not otherwise extensive. A fraudulent scheme is
“otherwise extensive” if the defendant “made a substantial
portion of [his] income” from the fraud scheme, the scheme
“continued in operation” for an extended period, or the
scheme “used many people,” including unknowing individ-
uals “to make the profit from the scheme.” United States v.
Sheikh, 

, 688-89 (7th Cir. 2004); see also Hussein, 664
F.3d at 161-62. If the “otherwise extensive” element is satis-
fied, the defendant need only have exercised actual control
over one other participant for the four-level enhancement to
apply. United States v. Blaylock, 

, 621 (7th Cir.
2005); see U.S.S.G. § 3B1.1, cmt. n.3 (“In assessing whether an
organization is ‘otherwise extensive,’ all persons involved
during the course of the entire offense are to be considered.
Thus, a fraud that involved only three participants but that
used the unknowing services of many outsiders could be con-
sidered extensive.”).
    Harmelech concedes the scheme spanned over six years
and involved hundreds of fraudulent accounts from 150
multi-dwelling buildings, often using the personally identify-
ing information of residents without their knowledge. He per-
sonally owned the three businesses used to perpetuate the
fraud and made his living operating them. He supervised
nine employees, two of whom knew of the fraud and facili-
tated it at Harmelech’s direction. He lied to four separate au-
thorized dealers, DIRECTV representatives, and building
14                                                No. 18-2169

owners and managers to open fraudulent customer accounts,
gain access to the buildings, and increase the number of re-
ceivers he could fraudulently install. His customers identified
him as their sole point of contact for cable services, and they
refused to work with other cable providers, or DIRECTV or
Signal Audit employees. Harmelech’s scheme was for his own
financial benefit—and, for six years, he did benefit. Based on
this evidence, we see no error in the district court’s decision
to find the scheme was otherwise extensive. The four-level en-
hancement was properly applied.
     For the reasons above, we AFFIRM.