EUGENE A. WRIGHT, Circuit Judge:
In cases involving complex computer sale/leaseback transactions, we consider whether the tax court erred in finding that the transactions were shams for federal income tax purposes and in applying the at risk provisions of the Internal Revenue Code, 26 U.S.C. § 465.
BACKGROUND
These appeals arise from the tax court’s disposition of five test cases. Four of the cases are before us.1 The fifth was appealed to the Eighth Circuit, which recently affirmed the tax court’s decision. Shriver v. Commissioner, 899 F.2d 724 (8th Cir.1990).
All transactions at issue in these cases involve the same basic set of facts. . In the late 1970s, Finalco, a leasing company, purchased computer equipment which was then leased to end-users.2 It then sold the equipment to the taxpayers, taking recourse and installment nonrecourse notes. The taxpayers also assumed part of the nonrecourse debt owed by Finalco to the banks which financed Finalco’s original purchases.
The taxpayers then leased the equipment back to Finalco. The amount of Finalco’s monthly rent payment was equal to the taxpayers’ payment on the installment non-recourse notes. The taxpayers also executed remarketing and residual sharing agreements with Finalco.3
[1362] Based on these transactions, the taxpayers deducted depreciation and interest expenses on their federal income tax returns. The IRS disallowed the deductions because the underlying transactions were shams, and assessed deficiencies against the taxpayers. The taxpayers challenged these assessments in the tax court.
Larsen was involved in four transactions, named for the end-users: Hon, Anaconda, Irving 1, and Irving 2.4 After a trial, the tax court found that the Hon and Anaconda transactions were shams. Larsen v. Commissioner, 89 T.C. 1229, 1254 (1987). Although it found that the Irving transactions were not shams, it held that Larsen was not at risk under 26 U.S.C. § 465 on his assumption of a portion of Finaleo’s debt owed to Citicorp Bank. Id. at 1273-74. The court also found that' he was not entitled to elect the half-year convention method of depreciation for the 1979 tax year. Id. at 1278.
Sturm, Casebeer, and Moore were each involved in separate transactions.5 The tax court found that each of these transactions were shams, and disallowed depreciation and interest deductions.
The taxpayers appeal. We have jurisdiction pursuant to 26 U.S.C. § 7482(a)(1).
STANDARD OF REVIEW
We review the tax court’s ultimate conclusion that the transactions were shams for clear error.6 Sochin v. Commissioner, 843 F.2d 351, 353, 355 (9th Cir.), cert. denied, 488 U.S. 824, 109 S.Ct. 72, 102 L.Ed.2d 49 (1988); Bail Bonds by Marvin Nelson, Inc. v. Commissioner, 820 F.2d 1543, 1548 (9th Cir.1987). “The clearly erroneous standard is particularly appropriate when, in instances such as this, the sham determination is based in part on the evaluation of conflicting evidence and live testimony.” Sochin, 843 F.2d at 355 (quoting Enrici v. Commissioner, 813 F.2d 293, 295 (9th Cir.1987)). “ ‘The expertise that the Tax Court brings to bear in its consideration of these complex factual situations provides ... further reason to defer to its conclusion.’ ” Sochin, 843. F.2d at 355 (quoting Thompson v. Commissioner, 631 F.2d 642, 646 (9th Cir.1980), cert. denied, 452 U.S. 961, 101 S.Ct. 3110, 69 L.Ed.2d 972 (1981)).7
We review de novo the legal standard applied by the tax court in making the sham determination. Sochin, 843 F.2d at 353.
ANALYSIS
I. THE TAX COURT’S APPLICATION OF THE SHAM TRANSACTION ANALYSIS
A. Collapse of the Two-Prong Test for Sham Transactions
In all four appeals, the taxpayers argue that the tax court incorrectly collapsed the prongs of the two-part test for determining [1363] whether transactions are' shams for tax purposes. We review this legal argument de novo.
In Frank Lyon Co. v. United States, 435 U.S. 561, 98 S.Ct. 1291, 55 L.Ed.2d 550 (1978), the Supreme Court held:
where ... there is a genuine multiple-party transaction with economic substance which is compelled or encouraged by business or regulatory realities, is imbued with tax-independent considerations, and is not shaped solely by tax-avoidance features that have meaningless labels attached, the Government should honor the allocation of rights and duties effectuated by the parties.
Id. at 583-84, 98 S.Ct. at 1303-04. This language was subsequently construed by this court to create a two-part test for determining whether a transaction is a sham:
1) has the taxpayer shown that it had a business purpose for engaging in the transaction other than tax avoidance? 2) has the taxpayer shown that the transaction had economic substance beyond the creation of tax benefits?
Bail Bonds, 820 F.2d at 1549. The application of the “business purpose” prong is a subjective test, whereas the application of the “economic substance” prong is an objective test. Sochin, 843 F.2d at 354.
Relying on the Fourth Circuit’s decision in Rice’s Toyota World, Inc. v. Commissioner, 752 F.2d 89 (4th Cir.1985), the appellants argue that a court must find that the taxpayer was motivated by no business purpose other than obtaining tax benefits and that the transaction has no economic substance in order to hold that a transaction is a sham. See id. at 91. This argument has no merit.
Several courts, including this circuit, have found that the Court’s holding in Frank Lyon was not intended to outline a rigid two-step analysis. See, e.g., Sochin, 843 F.2d at 354; Shriver, 899 F.2d at 727; James v. Commissioner, 899 F.2d 905, 908-09 (10th Cir.1990); Rose v. Commissioner, 868 F.2d 851, 854 (6th Cir.1989); see also Freidman v. Commissioner, 869 F.2d 785, 792 (4th Cir.1989) (the Fourth Circuit, applying its own decision in Rice’s Toyota, held that, although the tax court did not apply the “exact” two-part test, there was nevertheless ample support for the court’s finding of a sham transaction).
Instead, the consideration of business purpose and economic substance are simply more precise factors to consider in the application of this court's traditional sham analysis; that is, whether the transaction had any practical economic effects other than the creation of income tax losses.
Sochin, 843 F.2d at 354.
In each of the cases here, the tax court considered both the taxpayers’ subjective business motivation and the objective economic substance of the transactions in making its sham determinations.8 Its approach was entirely consistent with Ninth Circuit law and that of other circuits. The Eighth Circuit reached the same conclusion. Shriver, 899 F.2d at 727. There was no error.
B. Application of the Subjective “Business Purpose” Test
The tax court found that all four taxpayers did not meet the subjective “business purpose” prong of the two-part sham transaction analysis. These are factual findings which we review for clear error. Id. at 726.
Under the business purpose test, the court determines whether the taxpayers have shown that they had a business purpose for engaging in the transaction other than tax avoidance. Bail Bonds, 820 F.2d [1364] at 1549. “The business purpose factor often involves an examination of the subjective factors which motivated a taxpayer to make the transaction at issue.” Id. In these cases, the tax court considered several factors in evaluating the taxpayers’ business purpose, including their experience in computer leasing transactions, the extent of their investigation into the market value and residual value of the equipment prior to investing, the extent of such investigation by their professional advisors, and their testimony at trial regarding their motivation for entering into the transactions.
As a preliminary matter, the Eighth Circuit’s analysis of this issue in the related test case is relevant here. In Shriver, the tax court had found that the taxpayer had relied solely • on his CPA, the CPA’s inquiries and analysis were flawed, and much of the testimony as to Shriver’s motivation for the investment was “self-serving.” Shriver, 899 F.2d at 726. The Eighth Circuit found no clear error, noting:
It is enough, we believe, that the tax court considered the question, weighed the evidence before it — including the credibility of Shriver’s witnesses — and specifically determined on the record that no subjective non-tax business purpose existed. Although a finder of fact could have concluded otherwise, this is not sufficient grounds for reversal.
Id. As shown below, we reach the same result here.
1. Overview of Taxpayers’ Arguments The taxpayers rely primarily on two tax court cases to support their argument that the tax court’s findings on business purpose were clearly erroneous. First, they cite Lansburgh v. Commissioner, 54 T.C.M. (CCH) 691 (1987),' for the proposition that appraisals are not necessary prior to entering into leveraged equipment investments. The tax court here found that the taxpayers’ failure to investigate independently the equipment’s fair market value or its expected residual value was relevant in determining whether they manifested a subjective business purpose apart from tax considerations. See, e.g., Larsen, 89 T.C. at 1253.
Lansburgh is distinguishable because the taxpayer in- that case made detailed inquiries into the legitimacy of the deal, and even reviewed an appraisal of the equipment’s current fair market value and its expected residual value at the end of the lease term. None of the taxpayers here made such inquiries. Moreover, the tax court did not impose an appraisal requirement; rather, it found that the taxpayers’ failure to make independent inquiries into the properties’ values was a relevant factor in analyzing their business purpose. There was no error.
The second case relied upon by the taxpayers is Estate of Thomas v. Commissioner, 84 T.C. 412 (1985). In that case, the taxpayer’s net tax savings from a leveraged computer leasing transaction did not exceed the cash investment. Based on this fact, the tax court held that the transaction was not a sham because the investor could not have entered into the transaction solely for its tax benefits. Id. at 438. A crucial finding by the court in Thomas was that the potential for profit was “substantial.” Id. There was no such finding of profit potential by the tax court here, so Thomas is inapposite.
In addition to relying on these cases, the taxpayers contend that the tax court erred in attacking their reliance on professional advisors, including lawyers, accountants, and investment advisors. The tax court, however, did not attack their reliance on these advisors. Instead, it found that the advice sought was motivated solely by tax considerations. See, e.g., Larsen, 89 T.C. at 1253. In addition, after examining these advisors’ actions on their clients’ behalf, it found that their inquiries also were insufficient to meet the business purpose test. See, e.g., id. at 1253-54. There was no error.
The arguments above were made by all or most of the taxpayers. We next consider the specific arguments made by each of them on the business purpose issue.
2. Larsen
In addition to Larsen’s contention that he properly relied on his lawyer, he argues [1365] that the tax court erred in attacking his testimony as “self-serving.” Larsen, 89 T.C. at 1253. While he is correct when he says that such testimony is necessary “to prove that he had a bona fide profit motive,” there is no requirement that the tax court believe him. We find no clear error.
3.Sturm
Sturm relies primarily on a trip he took in 1981 to inspect his equipment at the end-user’s place of business to show that he had a business purpose. This trip has no relevance to the sham determination because it took place two years after he entered into the transactions.
He also relies on the facts that the sale of his insurance business was pursuant to a tax-free reorganization, and that his income after the sale was less than his income before the sale, to show that he was not motivated by tax considerations. These facts, however, do not establish that Sturm did not need deductions to shelter income. Moreover, the other facts present here, including advice from his investment advisor that he not enter into this transaction and his failure to seek an independent opinion of the equipment’s fair market value and residual value, support the tax court’s conclusion that he was motivated solely by the promised tax benefits of the transaction.
4.Casebeer
Casebeer presents the strongest case for manifesting a business purpose because he actually called the end-user of the equipment before investing to verify its existence, its value, and its useful life. The tax court, however, found his testimony to be “self-serving,” and his inquiry of the end-user to be of a general nature. It found that other factors, including his failure to consult an attorney and his lack of understanding of the transaction, indicated that had no business purpose at the time of entry into the transaction other than tax avoidance. These findings were not clearly erroneous.9
5.Moore
Moore argues principally that he relied entirely on his accountant for advice in deciding to enter into this transaction. The tax court analyzed the accountant’s actions, and found that he made no effort to verify independently the transaction’s profit potential. This finding was not clearly erroneous.
C. Application of the Objective “Economic Substance” Test
The tax court found that Larsen’s Anaconda transaction, and Sturm’s, Casebeer’s, and Moore’s transactions, lacked economic substance under the second prong of the sham analysis.10 These are factual findings which we review for clear error. See Keane v. Commissioner, 865 F.2d 1088, 1090 (9th Cir.1989).
Under this test, the court determines whether the transaction had economic substance beyond the creation of tax benefits. Bail Bonds, 820 F.2d at 1549. “The economic substance factor involves a broader examination of whether the substance of a transaction reflects its form, and whether from an objective standpoint the transaction was likely to produce economic benefits aside from a tax deduction.” Id.
In each transaction here, the tax court compared the taxpayer’s potential economic return with his investment in the transaction. See, e.g., Larsen, 89 T.C. at 1263-64 n. 22. It computed the taxpayers’ economic return by adding the interim revenue to be received from Finalco under the residual sharing arrangement and the equipment’s residual value, and deducting the fee which would be paid by the taxpayers to Finalco [1366] to remarket the equipment at the end of the lease under the remarketing agreement.11 Id. at 1264 n. 22. In computing the taxpayers’ investment, the court added the cash contribution, the amount of the recourse notes, and the interest on those notes.12 Id. If the return did not exceed the investment, the court found that the transaction did not have economic substance. See Rice’s Toyota, 752 F.2d at 94 (suggesting approval of this approach).
For example, it calculated Larsen’s return in the Anaconda transaction as follows:
Interim revenue $ 7,500
Residual value 15,000
Less: Remarketing fee (1,500)
$21,000
Larsen, 89 T.C. at 1264 n. 22. his investment as follows: a o EL <t> CL
Recourse note o O o CO co €©•
Interest on recourse note o rH 00
$31,710
Id. Because the return, $21,000, did not exceed the investment, $31,710, the court found that the transaction did not have economic substance. Id. at 1263.
The taxpayers do not fault this basic approach used by the tax court in determining whether the transactions had economic substance.13 Instead, they challenge the inclusion of certain items in the computations, and the values used by the tax court.
1. Inclusion of Interest on Recourse Notes in the Calculation of Taxpayers’ Investment
All of the taxpayers argue that the tax court erred by including interest on the recourse notes in determining their investment in the transactions under the “economic substance” test. The substance of the taxpayers’ argument is that their decisions to incur recourse debt was based on an analysis of their entire portfolio rather than on isolated transactions. Including interest on this debt therefore “penalizefs] those investors who utilize debt or leverage techniques so as to maximize their wealth.”
We reject this argument. At least one court has used interest on recourse notes to determine whether a transaction had economic substance. See Rice’s Toyota, 752 F.2d at 94. Moreover, as the government points out, such interest is a fixed cost of the transaction. In determining whether a transaction is a sham for tax purposes, it is proper for the court to look at all elements of the transaction at issue. In these case, it seems unlikely that the taxpayers incurred liability on recourse debt in complex computer leasing transactions in order to leverage their entire investment portfolios. Interest on the recourse notes was properly included in the court’s economic substance analysis.14
2. Tax Court’s Factual Findings Regarding Taxpayers’ Economic Return
The taxpayers argue that the tax court did not determine correctly their economic return from the transactions in applying the economic substance test. We consider each of their challenges in turn.
a. Larsen-Anaconda
Larsen argues first that the court erred in choosing the low end of the range of residual values proposed by his expert in determining his economic return from the [1367] Anaconda transaction. The court accepted a $15,000 residual value figure after comparing the range of values proposed by Larsen’s expert with the residual value proposed by the government’s expert, zero. Larsen, 89 T.C. at 1262-63. This finding was not clearly erroneous.
He also argues that the court erred in reducing the economic return figure by the 10% remarketing fee to be paid to Finalco at the end of the lease. Specifically, he argues that “it is not a foregone conclusion that Finalco will provide the remarketing service giving rise to the remarketing fee.” The court did not make a specific finding as to why it deducted the remarketing fee in Larsen. Even if the remarketing fee is not deducted, Larsen fails the economic substance test. There was no error.
b. Casebeer
Casebeer also argues that the court erred in deducting the remarketing fee in computing his transaction’s return. In this case, unlike Larsen, the court made specific findings as to why it deducted that fee:
Petitioner is an opthalmologist with experience and resources woefully inadequate to remarket the Equipment without assistance. We believe that petitioner as individual owner of peripheral computer equipment such as the Equipment comprising certain components to a computer system on lease to the End User must in substance rely on Finalco for the sale or the re-leasing of the Equipment. The record does not indicate that other dealers of computer equipment in the secondary market were available and capable to remarket Equipment under conditions more favorable to petitioner. Fi-nalco knew the location, status, and condition of the Equipment. We find it a fair inference that Finalco was relied upon by petitioner and was the sole source of expected assistance in any re-marketing effort of petitioner at the termination of the Owner Lease.
Casebeer argues that, contrary to the tax court’s finding, Comdisco, another leasing company involved in his transaction, offered him “an alternate and very competitive remarketing agency.”
In Rice’s Toyota, the Fourth Circuit indicated that a deduction for a remarketing fee was appropriate when its payment was “the more likely development in view of [the taxpayer’s] lack of computer marketing experience.” 752 F.2d at 94. The tax court in this case found that the remarket-ing fee would most likely be paid to Final-co, and this finding was not clearly erroneous.
c. Moore
Moore argues first that the tax court erred by selecting its expert’s low end of the range of possible residual values. The court accepted a $55,000 residual value figure after comparing the range of values proposed by Moore’s expert with the residual value proposed by the government’s expert. This finding was not clearly erroneous.
Second, he argues that the tax court erred in selecting a residual value for purposes of computing interim revenue, $16,-500, which was less than the residual value accepted for the termination of the owner lease, $55,000. Assuming, without deciding, that this argument has merit, the economic return on the investment would not be sufficient to make the transaction profitable.15
Finally, he argues that the court erred in deducting the remarketing fee in its economic substance analysis. The court found that Moore “must rely on any remarketing or residual revenue to be generated by Finalco.” As with Casebeer, this finding was not clearly erroneous. Moreover, even if the remarketing fee is not deducted, the transaction lacks economic substance under [1368] the court’s analysis.16
We hold that the tax court’s determinations that the taxpayers’ transactions were shams for income tax purposes were not clearly erroneous.
II. APPLICATION OF AT RISK RULES IN THE LARSEN-IRVING TRANSACTION
Larsen argues that the tax court misapplied the at risk rules in denying deductions related to the Irving transactions.17 In these transactions, he personally assumed a portion of Finalco’s nonrecourse debt owed to Citicorp in an attempt to increase his amount at risk.18 We review this legal issue de novo.
A person engaged in an activity to which § 465 of the Internal Revenue Code applies may deduct losses from that activity only to the extent he or she is “at risk.” 19 26 U.S.C. § 465(a)(1)(A). The leasing of de-preciable personal property is a § 465 activity. Id. § 465(c)(1)(C).
Generally, a taxpayer is at risk for the amount of cash invested in the activity and for amounts borrowed for which there is personal liability. Id. § 465(b)(1), (2). As an exception to the general rule that such amounts are at risk, however, § 465(b)(4) provides:
Notwithstanding any other provision of this section, a taxpayer shall not be considered at risk with respect to amounts protected against loss through nonre-course financing, guarantees, stop loss agreements, or other similar arrangements. (emphasis added)
The tax court held that Larsen was protected against loss through “other similar arrangements” in his assumption of Final-co’s debt to Citicorp based on the terms of his purchase agreement with Finalco.20 Larsen, 89 T.C. at 1272-74. We agree.
The term “other similar arrangements” is not defined in the tax code. Melvin v. Commissioner, 894 F.2d 1072, 1074 (9th Cir.1990). However, the legislative history of the Tax Reform Act of 1976, which created the at risk rules, sheds light on what Congress meant by this term:
[UJnder these rules, a taxpayer’s capital is not “at risk” in the business, even as to the equity capital which he has contributed to the extent he is protected against economic loss of all or part of such capital by reason of an agreement or arrangement for compensation or reimbursement to him of any loss which he may suffer. Under this concept, an investor is not “at risk” if he arranges to receive insurance or other compensation for an economic loss after the loss is sustained, or if he is entitled to reimbursement for part or all of any loss by reason of a binding agreement between himself and another person.
[1369] S.Rep. No. 94-938, 94th Cong., 2d Sess. 49 (emphasis added), reprinted in 1976 U.S. Code Cong. & Admin.News 2897, 3485.
Under the terms of Larsen’s agreement to purchase the computer equipment from Finalco, he was entitled to reimbursement from Finalco if he ever was required to pay Citicorp on the assumption. Paragraph (9) of the purchase agreements states:
[Finalco] covenants to [Larsen] that [Fi-nalco] shall (i) fully pay (or cause to be paid) and perform any and all obligations, liabilities and indebtedness (collectively “Debts”) secured by any liens (other than those arising by or through [Larsen]) against Equipment (“Liens”), when due, ...
Although Larsen assumed personal liability on a portion of Finalco’s debt to Citicorp, it is undisputed that he would not become liable on this debt unless Finalco defaulted.21 This debt, secured by the computer equipment, therefore remained a liability of Finalco under Paragraph (9) of the purchase agreement, and Finalco agreed to perform this obligation when due.
Paragraph (12) of Larsen’s purchase agreements states in relevant part:
[Finalco] hereby indemnifies [Larsen] and holds [Larsen] harmless from and against (i) any and all loss, cost, damage, injury or expense ... whatsoever and howsoever arising which [Larsen] or any of its agents or employees may incur by reason of any breach by [Finalco] of any of the covenants, warranties or representations by, or obligations of [Finalco] set forth in this Agreement ...
Finalco therefore agreed to indemnify Larsen if he was ever called upon to pay on the Citicorp debt. The effect of these clauses in the purchase agreements was to protect Larsen against loss through an “other similar arrangement.” He was not at risk under § 465(b)(4).22
Our recent decision in American Principals Leasing Corp. v. United States, 904 F.2d 477 (9th Cir.1990), buttresses this result. We said:
[T]he purpose of subsection 465(b)(4) is to suspend at risk treatment where a transaction is structured — by whatever method — to remove any realistic possibility that the taxpayer will suffer an economic loss if the transaction turns out to be unprofitable_ A theoretical possibility that the taxpayer will suffer economic loss is insufficient to avoid the applicability of this subsection. We must be guided by economic reality....
American Principals, 904 F.2d at 483 (citing Melvin, 894 F.2d at 1075, and Pritchett v. Commissioner, 827 F.2d 644, 647 (9th Cir.1987)).
The economic reality of Larsen’s transactions indicates that he was not at risk. During the trial, the lawyer who advised Larsen at the time he entered into the [1370] transactions testified as to his understanding of the purchase agreement clauses discussed above. In referring to Larsen’s liability under a virtually identical assumption in the Anaconda transaction, he said:
[I]f it turned out that [Larsen] had to pay some of the [amount] which he assumed, then it would be because Finalco breached their agreement and didn’t pay their obligations. And, therefore, [Larsen] would have recourse back to Finalco and so he would be able to recover from Finalco ...
Larsen therefore understood that he was protected against loss under the purchase agreements.
Moreover, although it appears from the documents that Larsen assumed primary personal liability on Finalco’s debt to Citi-corp, the tax court found that, in reality, this was not the case. Larsen, 89 T.C. at 1274. It found that Citicorp did not sign his assumption agreement and that Larsen’s liability for the assumed amount became operative only where Finalco defaulted on the Citicorp note. Id. Citicorp therefore viewed Finalco as the primary obligor. Id. These findings were not clearly erroneous, and support a conclusion that Larsen was not at risk based on these assumptions. See Pritchett, 827 F.2d at 646-47 (indicating that when agreements create secondary rather than primary liability, a taxpayer is not personally liable for the debt under the at risk provisions).
Based on the plain language of the purchase agreements and the economic realities surrounding the Irving transactions, the tax court found correctly that Larsen was protected against loss under § 465(b)(4).
III. TAX COURT’S REJECTION OF THE HALF-YEAR DEPRECIATION CONVENTION
Finally, Larsen argues that the tax court erred in finding that he was not entitled to use the half-year convention method of depreciation for the 1979 tax year. See Larsen, 89 T.C. at 1276-78. The government concedes this issue on appeal, and we reverse this part of the tax court’s decision.
CONCLUSION
The tax court’s decision in Larsen is REVERSED AND REMANDED as to the depreciation computation in 1979. In all other respects, its decisions are AFFIRMED. Costs are awarded to the Commissioner.