Defendant-appellant Torrence James pleaded guilty to wire fraud in violation of 18 U.S.C. § 1343 and engaging in a monetary transaction in property derived from specified unlawful activity in violation of 18 U.S.C. § 1957, both in connection with a scheme to fraudulently obtain mortgage loans to purchase twenty residential homes in and around Denver, Colorado. Mr. James appeals his 151-month sentence, arguing that the district court erroneously: (1) found him to be a leader or organizer of the criminal enterprise under United States Sentencing Guidelines Manual (“U.S.S.G” or “Guidelines”) § 3Bl.l(a); and (2) calculated the loss sustained by the victim lenders under U.S.S.G. § 2B1.1. Exercising jurisdiction under 28 U.S.C. § 1291 and 18 U.S.C. § 3742, we REVERSE in part, AFFIRM in part, and *1111remand to the district court for resentencing.
I. BACKGROUND
The fraudulent mortgage scheme took place from April 2004 until May 2005. As part of the scheme, Mr. James and co-defendant Ronald Fontenot would locate residential properties for sale. They would then procure appraisals for the properties that exceeded the sales price and submit materially false statements from straw buyers to mortgage lenders in order to obtain loans to purchase the properties at the inflated appraisal prices. The straw buyers put no money down, and the loans were secured by the properties. The straw buyers, through Mr. James and Mr. Fontenot, would contract with the seller to have the excess loan amount over the sales price disbursed to the straw buyers at closing, purportedly for improvements to the property. These excess funds, however, were actually disbursed to Mr. James, Mr. Fontenot, and others for their personal use.
Mr. James participated in the purchase of twenty properties involving ten lenders. Although he was able to make some payments on the mortgages, Mr. James could not keep up with the scheme and each property eventually went into foreclosure. However, most of the original ten mortgage lenders had sold the loans to other loan servicers prior to the foreclosure proceedings. Thus, the PSR ultimately reported that in most instances, “the actual loss was sustained by successor lenders.”
A superseding indictment was filed in August 2008, charging Mr. James and six others with various counts stemming from the fraud. Pursuant to a January 2007 plea agreement, Mr. James pleaded guilty to one count of wire fraud and one count of engaging in a monetary transaction in property derived from specified unlawful activity. At the time of the agreement, the government had little information relating to the loss resulting from Mr. James’s conduct, and fewer than half of the properties had been sold at foreclosure. Thus, in the agreement the government took the position that the excess funds (i.e., the amount of the loan above the sales price of the property) from each of the twenty property sales — a figure amounting to $2,298,193 — constituted the total gain from the fraud and was the appropriate measure of loss under the Guidelines. The initial PSR, which was prepared in March 2007, recommended the same. In addition, the PSR reported that the government had not yet provided final restitution figures under the Mandatory Victims Restitution Act (“MVRA”) related to losses on the mortgage loans.
As property records began to reflect foreclosure sales prices, however, the district court instructed the parties to calculate the actual loss sustained by the lenders. Apparently in furtherance of this instruction, as well as to establish restitution figures, the government contacted various lenders to ascertain their losses. Some lenders responded with dollar figures, others responded to state that they did not know the amount of loss, and still others did not respond at all. The government, however, did not submit this evidence to the district court. There was also no evidence presented to the district court as to what the successor lenders paid for the loans, or how much the original lenders gained or lost from the sales.
Ultimately, the PSR was amended eight times over the course of a year to report, for each of the twenty properties: (1) the name of the original lender; (2) the total amount of the original loan; and (3) the foreclosure sales price. The addenda to the PSR calculated the actual loss by subtracting, for each property, the foreclosure sale price from the original loan amount; after adding all twenty figures, the final *1112addenda to the PSR arrived at an actual loss amount of $3,731,839.
Mr. James submitted written objections to this calculation of loss, contending that the PSR’s method to calculate loss was flawed:
Mr. James disagrees with the methodology used by the probation officer to determine the actual loss to the victim lenders on the properties.... The public foreclosure and real estate sale records may accurately describe the ultimate sale price but do not provide enough information to determine the total loss suffered by the original lenders victimized by the fraudulent conduct. Many of these properties were purchased using a first mortgage and a home equity loan (HELOC) as a second mortgage. The total amount financed is reflected in the chart as a combined number for both loans. The property records do not reveal the amount recovered for each mortgage. Nor do the records account for the loans that were sold by the original lenders or the effect of mortgage insurance, payments by the borrowers before default, rents received by the lenders prior to the sales or other income or expenses related to the defaults. A better methodology is to use the restitution amounts actually claimed by the victims and reviewed for legitimacy and accuracy.
Sentencing hearings for Mr. James and his co-defendants took place on August 24, 2007, and January 4, March 14, and March 26, 2008. During these hearings, counsel for Mr. James objected to the loss amount recommended by the PSR, stating in relevant part that “the probation officer’s work does not answer the question of what the loss was that was suffered by the original lender.... [M]ost of the loans were sold.... What did they sell them for?” According to defense counsel, the amounts received at foreclosure were not relevant to the loss sustained by the original lenders because those lenders were generally not the ones that foreclosed on the properties. Counsel explained: “[W]e don’t believe the property records tell the Court what the actual loss was, because you can’t tell what happened to these loans, who foreclosed, and who suffered the loss, and what the loss amount was.” Counsel contended that loss could not be determined, and that Mr. James’s gain should be used as an alternative measure of loss.
The district court agreed with the position advocated by the government and the addenda to the PSR. Relevant to this appeal, the court found that Mr. James was a leader or organizer in criminal activity involving at least five participants and added four levels under § 3B1.1(a)(1). The court also added two levels because the offense involved ten victims — the ten original lenders. See U.S.S.G. § 2Bl.l(b)(2)(A)(i). Importantly, the district court refused to consider the successor lenders as victims, reasoning that “the losses of those entities do[ ] not constitute reasonably foreseeable pecuniary harm, because the decision to resell [the original loans] was made entirely by the [original] lender without the knowledge or input from the defendant.” Finally, the court found that the actual loss was $3,731,839— the amount produced after subtracting the foreclosure prices from the original loan amounts' — and therefore added eighteen points. See U.S.S.G. § 2Bl.l(b)(l)(J). Notwithstanding this estimate of actual loss, the district court refused to order restitution for any lender, original or successor, because doing so would require a determination of complex issues of fact that would complicate or prolong the sentencing process to a degree that the burden outweighed the need to provide restitution.1 See 18 U.S.C. § 3663A(c)(3)(B); *1113Combined with other enhancements, adjustments, and Mr. James’s criminal history score of III, the advisory Guidelines range was 135-168 months’ imprisonment. The court sentenced Mr. James to 151 months.
II. DISCUSSION
A. Role in the Offense
The district court applied a four-level enhancement based on its finding that Mr. James was an organizer or leader of a criminal activity that involved at least five participants. See U.S.S.G. § 3Bl.l(a). “We review for clear error the district court’s finding that [the defendant] acted as a leader or organizer for purposes of § 3B1.1.” United States v. Wilfong, 475 F.3d 1214, 1218 (10th Cir.2007). Under this standard, we will not reverse the district court’s finding unless, “on the entire evidence, we are left with the definite and firm conviction that a mistake has been committed.” Id.
To determine whether an individual qualifies as an organizer or leader, as opposed to merely a manager or supervisor subject to a lesser enhancement, the Guidelines instruct the sentencing court to consider
the exercise of decision making authority, the nature of participation in the commission of the offense, the recruitment of accomplices, the claimed right to a larger share of the fruits of the crime, the degree of participation in planning or organizing the offense, the nature and scope of the illegal activity, and the degree of control and authority exercised over others. There can, of course, be more than one person who qualifies as a leader or organizer of a criminal association or conspiracy.
U.S.S.G. § 3B1.1 cmt. n. 4.
In this case, the district court found the following facts to support the organizer-or-leader enhancement: (1) Mr. James, along with Mr. Fontenot, was primarily responsible for arranging and negotiating the purchase of the properties; (2) they were primarily responsible for recruiting and overseeing the straw purchasers; (3) they either collectively or separately retained the largest share of the cash proceeds from the closings — their share “dwarfed” the share retained by the other co-defendants; and (4) they were responsible for the part of the scheme involving the five straw buyers.
On appeal, Mr. James does not argue that these facts are insufficient to support the § 3Bl.l(a) enhancement; rather, he contends that the facts themselves are clearly erroneous. We have carefully reviewed the record and cannot say with conviction that the district court erred in finding these facts. We reject Mr. James’s attempt to substitute the court’s view of the evidence with his own express admissions regarding his involvement in the crimes, and we further disagree with Mr. James’s position that other, uncharged individuals were actually the organizers of the scheme. See U.S.S.G. § 3B1.1 cmt. n. 4 (“There can, of course, be more than one person who qualifies as a leader or organizer of a criminal association or conspiracy.”). In short, the district court’s application of the § 3Bl.l(a) enhancement is not clearly erroneous.
B. Calculation of Actual Loss
Under U.S.S.G. § 2Bl.l(b), a defendant’s base offense level for a fraud *1114conviction is increased according to the loss. The court should use the greater of actual or intended loss. U.S.S.G. § 2B1.1 cmt. n. 3(A). If there is a loss but it cannot reasonably be determined, the court may use gain that resulted from the offense as an alternative measure of loss. Id. cmt. n. 3(B).
Based on the evidence presented at the sentencing hearings, the district court found that the actual loss attributable to Mr. James’s conduct was $3,731,839. This increased his offense level by 18. See U.S.S.G. § 2B1.1 (b)(1)(J) (18-level increase for loss between $2,500,001 and $7,000,000). On appeal, Mr. James contends that actual loss in this case cannot reasonably be determined, and that the district court should have considered his gain instead.2 Using the $2,298,193 of gain reported in the plea agreement and the original PSR would have increased his offense level by only 16. See id. § 2Bl.l(b)(l)(I) (16-level increase for loss between $1,000,001 and $2,500,000). In support of his contention, Mr. James argues that the district court did not consider that the original lenders sold the loans to successor lenders, and that the loss figure was not supported by any evidence regarding payments made on the loans.
We review factual findings regarding calculation of loss under a clearly erroneous standard. United States v. Smith, 951 F.2d 1164, 1166 (10th Cir.1991). In this case, however, Mr. James challenges the methodology the district court used to calculate loss. This is a legal question we review de novo. See United States v. Lara, 956 F.2d 994, 998 (10th Cir.1992) (“The question of what factors ... the court [may] consider in computing the amount of a loss is a legal question that we review de novo.”); United States v. Haddock, 12 F.3d 950, 961 (10th Cir.1994) (“[W]e review de novo what may be included in computing loss.”); see also United States v. Goss, 549 F.3d 1013, 1016 (5th Cir.2008) (methodology used in computing loss is reviewed de novo); United States v. Staples, 410 F.3d 484, 490 (8th Cir.2005) (same).
“ ‘Actual loss’ means the reasonably foreseeable pecuniary harm that resulted from the offense.” U.S.S.G. § 2B1.1 cmt. n. 3(A)(i). “The [sentencing] court need only make a reasonable estimate of the loss.” Id. cmt. n. 3(C). In cases where the defendant has pledged collateral to secure a fraudulent loan, “[a]ctual loss should be measured by the net value, not the gross value, of what was taken.” Smith, 951 F.2d at 1167. This loss is calculated by subtracting the value of the collateral — or, if the lender has foreclosed on and sold the collateral, the amount of the sales price — from the amount of the outstanding balance on the loan. See United States v. Swanson, 360 F.3d 1155, 1169 (10th Cir.2004) (stating that in “fraudulent loan cases ... the loss is the amount of the loan not repaid at the time the offense is discovered, reduced by the amount the lending institution has recovered (or can expect to recover) from any assets pledged to secure the loan”); Haddock, 12 F.3d at 961 (“[T]he net loss to a lender is the unpaid amount of the loans minus the value of the collateral at the time of sentencing.”); Smith, 951 F.2d at 1167 (“[I]f the fraud consists of an unequal exchange of property, the loss ... consists only of the difference in value between what was given and what was obtained.”).
In this case, the district court subtracted the foreclosure sales prices of the properties from the amount of the original loans *1115in order to calculate actual loss. But, as Mr. James emphasized below and on appeal, the original lenders were generally not the lenders who foreclosed on the properties. Rather, most of the original lenders sold the loans to successor lenders before the foreclosure sales. Thus, the successor lenders — not the original lenders — received the proceeds from the foreclosure sales. Accordingly, to the extent any original lender sustained an actual loss, that loss is the difference between the outstanding balance on the original loan and what the lender received when it sold the loan. See Smith, 951 F.2d at 1167 (“[T]he loss ... consists only of the difference in value between what was given and what was obtained.”). The foreclosure sales price is irrelevant to that computation because the original lender was not the recipient of those proceeds and because there is no evidence suggesting those figures were a reasonable estimate of what the original lenders received when they sold the loans to the successor lenders. Indeed, the PSR — which was prepared with the investigative assistance of the government — explicitly recognized this reality when it noted that “[t]he victim lenders ... [are] the original lenders. In most cases, however, the loans were subsequently sold to other loan servicers and the actual loss was sustained by successor lenders.”
Nor is it reasonable in this case to subtract the foreclosure sales prices of the properties from the amount of the original loans in order to calculate the actual loss sustained by the successor lenders. As noted, actual loss should be measured by the net value of what was taken — -in other words, the difference in value between what was given and what was obtained.
See Smith, 951 F.2d at 1167. The successor lenders’ actual loss, then, is the difference between what they paid the original lenders for the loans (less principal repayments by borrowers, if any) and what they received for the properties at the foreclosure sales, plus reasonably foreseeable expenses relating to the foreclosure proceedings. Using the amount financed by the original lender as a starting point in this calculation is improper when there is no evidence suggesting that the loans were sold to the successor lenders for sums approximating the original loan amount. Moreover, in determining that the successor lenders were not victims under U.S.S.G. § 2Bl.l(b)(2), the district court explicitly found that “the losses of those entities does not constitute reasonably foreseeable pecuniary harm.... ” That finding precluded the district court from considering their losses as part of the actual loss calculation for purposes of § 2Bl.l(b)(l).3 See U.S.S.G. § 2B1.1 cmt. n. 3(A)(i) (“ ‘Actual loss’ means the reasonably foreseeable pecuniary harm that resulted from the offense.”).
We recognize “[t]he district court need not calculate actual or intended loss with exact precision[;] it need only make reasonable estimates.” United States v. Galloway, 509 F.3d 1246, 1251 (10th Cir.2007). But a loss estimate is reasonable only when it is calculated under a reasonable method. In this case, the district court’s finding that the successor lenders did not sustain reasonably foreseeable pecuniary harm limited the court’s actual loss calculation to the losses sustained by the original lenders only.4 And on this record, subtracting the foreclosure sales prices from the original loan amounts is not a *1116reasonable method of calculating the original lenders’ actual loss. Those lenders never received the proceeds from the foreclosure sales, and there is no evidence that those figures are a reasonable estimate of what those lenders received when they sold the loans to the successor lenders; thus, the loss sustained by the original lenders could not, on this record, be calculated with reference to the foreclosure sales prices.
We are aware that today’s banking realities — the bundling of mortgages into securities, for example' — may make it difficult to identify precisely the proceeds a lender received for a specific mortgage loan. The Guidelines, however, contemplate such circumstances and thus permit a district court to estimate loss “based on available information.” U.S.S.G. § 2B1.1 cmt. n. 3(C) (emphasis added). It is not necessary for us to set forth an exhaustive list of the types of “available information” the court may use in a case in which the victim lender sells downstream a fraudulently obtained loan; nor do we suggest that it is always improper to estimate such a lender’s loss based on the foreclosure price information provided to the district court in this case. It is enough to say that this particular record included no evidence to support an inference that the foreclosure sales prices were appropriate estimates of what the original lenders received when they sold the loans to the successor lenders. Thus, those figures could not be used to determine the original lenders’ actual losses.5
We therefore remand with instructions for the district court to recalculate the actual losses of the ten original lenders it identified as victims of Mr. James’s conduct. See United States v. Kristi, 437 F.3d 1050, 1055 (10th Cir.2006) (a non-harmless error in determining the applicable Guidelines range warrants a remand for resentencing). Because the court previously found, however, that the successor lenders’ losses were not reasonably foreseeable pecuniary harm — a finding that the government does not challenge in this appeal — the court shall not include their losses, if any, for purposes of the § 2Bl.l(b)(l) enhancement. If the district court finds that the original lenders have suffered an actual loss, but that the loss cannot reasonably be determined, it shall explain this finding, see Galloway, 509 F.3d at 1252, and shall use gain as an alternative measure of loss. In that case, we express no opinion on the appropriate calculation of gain.6
III. CONCLUSION
The district court appropriately applied a four-level enhancement under U.S.S.G. § 3Bl.l(a) based on its finding that Mr. James was an organizer or leader of the offense, but the court erred in calculating the actual loss resulting from the offense for purposes of § 2Bl.l(b)(l). We therefore remand for resentencing in accordance with the instructions set forth above. Mr. James’s pro se motion to file a supplemental brief is denied.
. Although the court did order Mr. James to pay restitution in the amount of $1,203 to *1113Marli Carani and $25,433 to the Homeowners Association of The Villas at Cherry Creek, neither party was a lender. Ms. Carani’s claim stemmed from identify theft and the Homeowners Association’s claim arose out of unpaid homeowners dues, fines, and accompanying legal fees.
. The parties’ sole disagreement is whether the $3,731,839 figure reasonably represents the actual loss resulting from Mr. James’s conduct or whether gain should be used as an alternative measure of loss. Accordingly, we do not address the issue of intended loss.
. We express no opinion on the propriety of this finding, which is unchallenged on appeal.
. We do not suggest that a district court may never consider successor lenders' actual loss in calculating the total loss under U.S.S.G. § 2B 1.1(b). So long as that harm is reasonably foreseeable, it is properly considered actual loss under the Guidelines.
. On this point, we respectfully disagree with the concurrence’s view that, in all cases, "[wjhen a mortgage is resold, a successor lender pays the amount of the original mortgage plus a markup equal to the original lender's profit” and thus "the total loss caused by the fraud can always be calculated as the original mortgage amount minus the final foreclosure price.”
. Mr. James also argues that the district court’s loss calculation is unreasonable given the court's position that it was too difficult to order restitution. Although we need not reach this issue, we note that the calculation of loss for sentencing purposes does not necessarily establish loss for the purpose of awarding restitution under the MVRA. United States v. Gallant, 537 F.3d 1202, 1247 (10th Cir.2008).