JERRY E. SMITH, Circuit Judge:

This appeal concerns the tax consequences of two transactions undertaken by Dow Chemical Company ("Dow") and a number of foreign banks1 from 1993 through 2006. During those years, Dow and the foreign banks purported to operate two partnerships that generated over one billion dollars in tax deductions for Dow. After a five-day trial, the district court disregarded the partnerships for tax purposes on three grounds: (1) The partnerships were shams; (2) the transactions lacked economic substance; and (3) the banks' interests in Chemtech Royalty Associates, L.P. ("Chemtech"), were debt, not equity. The court also imposed substantial understatement and negligence penalties but refused to impose substantial-valuation or gross-valuation misstatement penalties. Because, under these specific facts, the court did not clearly err in holding that Dow lacked the intent to share the profits and losses with the foreign banks, we affirm its sham-partnership holding. In light of United States v. Woods, ___ U.S. ___, 134 S.Ct. 557, 187 L.Ed.2d 472 (2013), however, we vacate and remand as to the penalty award.

I.

In the early 1990s, Goldman Sachs developed a financial product called Special Limited Investment Partnerships ("SLIPs"), which it promoted as a tax shelter. A series of steps typically had to be executed to create this type of product. First, the American corporation had to identify a valuable group of assets with a tax basis2 at or near zero. Second, the corporation needed to create or designate subsidiaries through which it would participate in the transaction. Those subsidiaries would then contribute the assets to the partnership. Third, the corporation had to entice foreign entities to participate in the transaction. The tax benefits generated by the partnership could be attained only if the partnership's income could be assigned to a tax-indifferent party. Fourth, the parties would need to enter into various agreements that would govern the transaction.3 In 1992, Dow decided to pursue this transaction, endeavoring to create an asset-backed equity financing vehicle.

Following these steps, Dow selected 73 patents to contribute to the partnership. The district court found that Dow did not select "patents that would be attractive to a third party." Instead, it contributed those patents that (1) "had the highest value (in order to reduce the total number of patents)," (2) had a zero or near zero tax basis, and (3) were actively used by one of Dow's businesses. For most patents, Dow "did not contribute all technology that would have been necessary for third party licensees," requiring a potential third-party licensee to obtain licenses from both the partnership and Dow. In line with these findings, Dow selected patents valued at roughly $867 million, with 71 of the 73 patents having zero tax basis.4

Next, Dow created two domestic subsidiaries — Diamond Technology Partnership Co. ("DTPC") and Ifco, Inc. ("Ifco") — and used a wholly — owned foreign subsidiary — Dow Europe, S.A. ("DESA") — to carry out this transaction. Through these subsidiaries, Dow formed Chemtech as a Delaware limited partnership with its principal place of business in Switzerland.5 Again through these subsidiaries, Dow contributed to Chemtech I6 the identified 73 patents, $110 million, and all of the stock of Chemtech Portfolio, Inc. ("CPI"), a pre-existing shell corporation owned by Dow.

Five foreign banks decided to participate as limited partners in Chemtech, investing a total of $200 million in the partnership. The entry of the foreign banks forced Ifco's partnership share to be retired. By October 1993, Chemtech was owned 1% by DESA (the general partner), 81% by DTPC, and 18% by the foreign banks.

Finally, Dow and the foreign banks entered into various agreements to govern the transaction, including a patent license agreement, a partnership agreement, and various indemnity agreements. The patent license agreement allowed Dow to continue to use the patents contributed to Chemtech. Under that agreement, Dow bore responsibility for all costs related to the patents and paid a royalty to Chemtech, regardless of Dow's use of the patents. Chemtech did not change Dow's use of its patents. The partnership agreement, in relevant part, (1) required the maintenance of capital accounts for each partner,7 (2) governed the allocation of profits and losses among the partners,8 (3) limited the types of assets the partnership could hold,9 (4) included the conditions that triggered the right to liquidate the partnership,10 and (5) provided the manner in which assets would be allocated on liquidation.11 Finally, because the foreign banks conditioned their willingness to participate in Chemtech on being indemnified against any liability arising from the assets or any tax liability, Dow indemnified them against those risks.

Chemtech I operated from April 1993 through June 1998, during which time Dow's royalty payments served as Chemtech's primary source of income, totaling $646 million. Because Chemtech claimed $476.1 million of book depreciation on the patents contributed to it, it reported book profits12 of only $61.7 million,13 out of which it paid (1) the 6.947% priority return to the foreign banks, (2) a 1% distribution to DESA, and (3) a relatively small distribution to each partner to pay its Swiss tax obligation. Chemtech then contributed the remaining cash to CPI, which loaned the bulk of the cash to Dow. Chemtech allocated the overwhelming majority of its income to the foreign banks and only a fraction of its income to Dow. As the district court observed, "[w]hile Dow claimed royalty expense deductions for the money flowing to Chemtech, it did not take into account the income of the bulk of the money flowing from Chemtech."14

In December 1997, DESA informed the foreign banks that new tax regulations could potentially subject the banks' priority return to a 30% withholding tax for which Dow would be responsible under the tax indemnity. In February 1998, Dow terminated Chemtech I. The foreign banks received the sum of their capital account balances, the early liquidation amounts, 1% of the increase in value of the contributed patents, and the priority return for one month. Ifco also bought out DESA's interest as general partner.

Shortly after terminating Chemtech I, Dow began planning a similar transaction that would operate essentially the same way. Dow again sought to identify a high — value, low tax basis asset to contribute to Chemtech II. Dow decided to use one of its Louisiana chemical plants. The chemical plant was valued at $715 million but had a tax basis of only about $18.5 million.

As in Chemtech I, Dow utilized a subsidiary to participate in the transaction-this time Dow Chemical Delaware Corporation ("DCDC"). In June 1998, DCDC contributed the chemical plant and all of the stock of a shell subsidiary, Chemtech Portfolio Inc. II ("CPI II"), to Chemtech II. Dow entered into a lease with Chemtech II for continued use of the chemical plant. Under the lease, Dow remained responsible for all expenses associated with the plant and was required to pay rent regardless of its use of the plant. As with the patents in Chemtech I, Chemtech II did not change Dow's use of the chemical plant. During the transition from Chemtech I to Chemtech II, Dow retired DTPC as a partner.15

In June 1998, RBDC, Inc. ("RBDC"), a U.S. affiliate of Rabo Mercent Bank N.V., purchased a limited interest in Chemtech II for $200 million. Rabo Mercent Bank N.V. was one of the foreign banks that invested in Chemtech I. At this point, Chemtech II was owned 6.37% by Ifco, 20.45% by RBDC, and 73.18% by DCDC. Ifco served as the general partner and RBDC and DCDC served as limited partners. Chemtech II operated similarly to Chemtech I:(1) Dow entered into similar agreements with substantially similar terms; (2) the cash flows displayed similar patterns;16 and (3) Dow enjoyed substantial tax savings through a similar but not identical mechanism.17

Chemtech II's partnership agreement permitted RBDC to elect to liquidate its interest in March 2003. At that time, Dow and RBDC negotiated a new partnership agreement that reduced RBDC's priority return to 4.207%. Dow and RBDC continued to operate Chemtech II through June 2008.

The Internal Revenue Service ("IRS") issued Chemtech Final Partnership Administrative Adjustments ("FPAAs") for tax years 1993 through 2006. It also asserted accuracy-related penalties under I.R.C. § 6662 for 1997 through 2006. Dow sued to contest the FPAAs. After a five-day trial, the court disregarded the partnership for tax purposes on three grounds: (1) The partnerships were shams; (2) the transactions lacked economic substance; and (3) the banks' interests in Chemtech were debt, not equity.18 The court also assessed a twenty-percent penalty for each of the relevant tax years, awarding substantial-understatement and negligence penalties. The court, however, believed it could not impose a gross-valuation misstatement penalty.

As to its sham partnership holding, the court found that Dow lacked both the intent to act in good faith for some genuine business purpose other than tax avoidance and the intent to share profits and losses with the foreign banks. As to the second intent, the court found that "[t]he foreign banks were not true partners" because "the banks were [essentially] guaranteed a return just under 7% each year" and "[a] valid partnership is not formed where, among other things, one partner receives a guaranteed, specific return."

On appeal, Dow avers that the district court erred in finding the partnerships to be shams. Because Dow believes the foreign banks' interest cannot be classified as debt under United States v. South Georgia Railway Co., 107 F.2d 3 (5th Cir.1939), it claims that it must have provided the foreign banks with equity in Chemtech. It reasons that because the foreign banks received equity, Dow entered into a valid tax partnership, regardless of any other criteria. As to the penalty award, Dow concedes that the court erred in foreclosing the availability of a gross-valuation misstatement penalty.19

The government contends, for three reasons, that Dow did not intend to share the profits or losses with the foreign banks: First, the agreement allocated essentially all of the risk — bearing to Dow. "The banks [] insisted that they bear no liability" — product, tax, or any other type — "for the patents or the chemical plant." Second, "[t]he banks did not have bona fide equity interests in Chemtech." And third, "[t]he evidence is [] clear that the banks did not view themselves as joining with Dow to manage such assets."

II.

"The starting point for our analysis is the cardinal principle of income taxation: A transaction's tax consequences depend on its substance, not its form." Southgate Master Fund, L.L.C. ex rel. Montgomery Capital Advisors, LLC v. United States, 659 F.3d 466, 478-79 (5th Cir.2011). That maxim "is the cornerstone of sound taxation." Estate of Weinert v. Comm'r, 294 F.2d 750, 755 (5th Cir.1961). "`Tax law deals in economic realities, not legal abstractions.'" Id. (quoting Comm'r v. Sw. Exploration Co., 350 U.S. 308, 315, 76 S.Ct. 395, 100 L.Ed. 347 (1956)). "This foundational principle finds its voice in the judicial anti-abuse doctrines, which prevent taxpayers from subverting the legislative purpose of the tax code by engaging in transactions that are fictitious or lack economic reality simply to reap a tax benefit." Southgate, 659 F.3d at 479 (internal quotation marks omitted).

A taxpayer may not be able to claim the "tax benefits of a transaction — even a transaction that formally complies with the black-letter provisions of the Code and its implementing regulations — if the taxpayer cannot establish that `what was done, apart from the tax motive, was the thing which the statute intended.'" Id. (quoting Gregory v. Helvering, 293 U.S. 465, 469, 55 S.Ct. 266, 79 L.Ed. 596 (1935)). "Because so many abusive tax-avoidance schemes are designed to exploit the Code's partnership provisions, our scrutiny of a taxpayer's choice to use the partnership form is especially stringent." Id. at 483-84 (footnote omitted).

"In an appeal from a bench trial, we review the district court's findings of fact for clear error and its conclusions of law de novo. Id. at 480. Specifically, a district court's characterization of a transaction for tax purposes is a question of law subject to de novo review, but the particular facts from which that characterization is made are reviewed for clear error." Id. (internal quotation marks omitted). "Under the clearly erroneous standard, we will uphold a finding so long as it is plausible in light of the record as a whole," United States v. Ekanem, 555 F.3d 172, 175 (5th Cir.2009) (internal quotation marks omitted), or so long as this court has not been "left with the definite and firm conviction that a mistake has been made," Streber v. Comm'r, 138 F.3d 216, 219 (5th Cir.1998).

III.

A partnership "may be disregarded [for tax purposes] where it is a sham or unreal."20 In order not to be a sham, or to be a valid partnership for tax purposes, "persons [must] join together their money, goods, labor, or skill for the purpose of carrying on a trade, profession, or business and [] there [must be a] community of interest in the profits and losses." Comm'r v. Tower, 327 U.S. 280, 286, 66 S.Ct. 532, 90 L.Ed. 670 (1946).21 This court has recently reaffirmed the test announced in Tower and repeated in Culbertson: "As the Supreme Court [has] explained..., whether a partnership will be respected for tax purposes depends on whether the parties in good faith and acting with a business purpose genuinely intended to join together for the purpose of carrying on the business and sharing in the profits and losses." Southgate, 659 F.3d at 483 (internal quotation marks omitted). "The fact that a partnership's underlying business activities had economic substance does not, standing alone, immunize the partnership from judicial scrutiny." Id. at 484.

As Tower, Culbertson, and Southgate demonstrate, the parties, to form a valid tax partnership, must have two separate intents: (1) the intent to act in good faith for some genuine business purpose and (2) the intent to be partners, demonstrated by an intent to share "the profits and losses." If the parties lack either intent, then no valid tax partnership has been formed. To determine whether the parties had these intents, a court must consider "all the relevant facts and circumstances," including (a) "the agreement," (b) "the conduct of the parties in execution of its provisions," (c) the parties' statements, (d) "the testimony of disinterested persons," (e) "the relationship of the parties," (f) the parties' "respective abilities and capital contributions," (g) "the actual control of income and the purposes for which it is used," and (h) "any other facts throwing light on their true intent." Id. at 483 (internal quotation marks omitted). Consistent with these directives, we limit our consideration and decision here to the specific facts and transactions that are presented.

Southgate relied on TIFD III-E, Inc. v. United States (Castle Harbour II), 459 F.3d 220 (2d Cir.2006), which involved a scheme very similar to the one involved in this case. There, acting through various subsidiaries, General Electric Capital Corporation ("GECC") formed a partnership with two Dutch banks. Both GECC and the banks contributed assets to the partnership.22 Although "the Dutch banks[] contributed about 18% of the partnership's capital and [] nothing to its management, [they] were allocated ... 98% of most of its taxable income." Id. at 227. The banks' actual receipts, however, were much smaller: "[T]he reimbursement of their investment, plus an annual return at an agreed rate near 9%, plus a small share in any unexpectedly large profits," which was capped at less than 2.5% of the banks' investment. Id. at 227, 229. In response to the IRS's issuance of two FPAAs, GECC sued to challenge their validity. The district court found that the transactions had economic substance and that Castle Harbour was a valid tax partnership.

The Second Circuit reversed, determining "that the Dutch banks [were not] equity partners in the Castle Harbour partnership because they had no meaningful stake in the success or failure of the partnership." Id. at 224. The Dutch banks neither shared in the profits nor the losses. "As a practical matter," GECC capped "the Dutch banks' opportunity to participate in unexpected and extraordinary profits (beyond the reimbursement of their investment at the Applicable Rate of return)...." Id. at 235.23 And second, the Dutch banks faced no meaningful risk of loss: "[F]eatures of the Castle Harbour agreements combined to provide the Dutch banks with not only a reasonable expectation, but an ironclad assurance that they would receive repayment of their principal at the Applicable Rate of return, regardless of the success of the Castle Harbour venture." Id. at 239-40.24

In Castle Harbour II, in conducting the sham-partnership inquiry, the Second Circuit considered it helpful first to address whether the interest has "the prevailing character of debt or equity." Id. at 232.25 Dow insists that we must first determine whether an interest qualifies as debt or equity before we can address whether there is a sham partnership under Culbertson. To support that proposition, Dow points to South Georgia Railway, which it believes demonstrates that debt requires (1) a fixed maturity date on which fixed amounts are due and (2) a holder's legal right to enforcement in the event of default. The reasoning continues that the foreign banks were not legally entitled to repayment of their investment even if the banks could recover the value of their partnership share when terminating the partnership. Therefore, Dow avers that the parties must have held equity in Chemtech and must have been valid tax partners under Culbertson regardless of what else the record may demonstrate.

Dow's argument fails. First, it has not identified any precedent that requires us to (a) classify an interest as debt or equity before conducting the Culbertson inquiry and (b) find a valid partnership solely because the parties did not have a legal right to demand repayment of their principal investment on any fixed future date. Even assuming that South Georgia Railway correctly describes when we must classify an interest as debt,26 that case does not have any bearing on the sham-partnership inquiry. Southgate certainly did not rely on South Georgia Railway in any respect.27

Second, such a requirement would run afoul of Culbertson and Southgate. In essence, Dow wants us to limit our sham-partnership inquiry to two considerations: whether the parties executed a legal document expressly (1) allowing the foreign banks to demand repayment of their principal investment (as opposed to the value of their partnership share) and (2) specifying a fixed date on which they could do so. Even further assuming Dow can demonstrate the transactions lacked one of these criteria,28 Southgate does not restrict our inquiry in that manner. "The sine qua non of a partnership is an intent to join together for the purpose of sharing in the profits and losses of a genuine business." Southgate, 659 F.3d at 488. Accepting Dow's suggestion would require us to elevate the transaction's form over its substance, contrary to long-standing doctrine.

Therefore, in assessing whether the district court erred in its sham-partnership holding, we express no opinion as to whether the interests should be classified as debt. Instead, we limit our inquiry to whether Dow possessed the intent to be partners with the foreign banks, focusing on whether Dow had the intent to share the profits and losses with the foreign banks. To make this determination, we consider all relevant "facts throwing light on their true intent," id. at 484 (quoting Culbertson, 337 U.S. at 742, 69 S.Ct. 1210), and review only for clear error, id. at 480.29

IV.

As we explain, we consider the court's finding on both the intent to share profits and the intent to share losses to be plausible in light of the record as a whole and therefore not clear error. First, the transactions were structured to ensure that Dow paid the foreign banks a fixed annual return on their investment "regardless of the success of the [Chemtech] venture," just as in the transaction in Castle Harbour II. The agreement entitled the foreign banks to 99% of Chemtech's profits until the banks received the priority return, but only 1% after that. Even if Chemtech did not generate sufficient profits to pay the return, itself a highly unlikely situation,30 the foreign banks were still entitled to 97% of the priority return. Moreover, the banks received compensation even if the partnership did not last the length anticipated by the parties, again demonstrating that the banks were compensated regardless of profitability. Dow even insulated the banks from bearing transactional costs incurred by participating in Chemtech.31

Second, Dow agreed to bear all of the non-insignificant risks arising out of the Chemtech transactions, which further shows that the parties did not intend to share any possible losses. The transaction created only three possible sources of loss: (1) tax liability, (2) liability arising from ownership of the patents or chemical plant, and (3) loss of the banks' initial investment. Dow has not identified any other possible source of loss. Because Dow indemnified the foreign banks for any liability arising from the patents and the chemical plant and for any tax liability, Dow did not intend to share that risk with the foreign banks. In fact, the foreign banks would not have participated in Chemtech if they had to bear any of that risk.

Furthermore, just as in Castle Harbour II, the agreement included four significant "ironclad" assurances to ensure that Dow would not misappropriate or otherwise lose the banks' initial investment: One, requiring Chemtech to hold 3.5 times the unrecovered capital contributions of the bank, ensured that if anything happened, the banks would be able to get back their money. Two, by severely limiting the assets Chemtech could hold, the agreement again minimized the possibility that the foreign banks would lose their initial investment. Three, in light of all of the possible voluntary conditions that triggered the right to terminate, if the banks perceived any risk to their investment, the agreement allowed the banks to terminate the partnership and recoup effectively their full initial investment with minimal transaction costs. Four, Dow guaranteed that its subsidiaries would perform their obligations under the various agreements. All of these features worked together to ensure that the foreign banks faced effectively no risk to their initial capital investment or to their priority return.

Third, just as in Castle Harbour II, the foreign banks did not meaningfully share in any potential upside. The possibility that the foreign banks could possibly obtain a fraction of residual profits does not make the finding on intent clearly erroneous. This is true because residual profits were possible only if a patent portfolio performed well enough to trigger Dow's obligation to pay variable royalties. Dow, however, does not contend — and nothing in the record suggests — that Dow or the foreign banks expected the contributed patents to increase in value. In fact, Dow does not even claim that it created Chemtech for the purpose of managing its patents. The parties could not have intended to share profits through a means no one expected or designed to be profitable. Even assuming arguendo (a) the intent to share profits can be demonstrated in a way not contemplated to be profitable at the time of the agreement, or (b) Dow and the foreign banks believed the patents would increase in value, we would still not consider the district court's finding clearly erroneous. The agreement (a) allocated only 1% of the increased value of a given patent portfolio to all of the foreign banks collectively and (b) allowed Dow effectively to control Chemtech's ability to earn such additional profits by giving Dow the ability to remove profitable patents.

All of these considerations demonstrate that the district court did not clearly err in determining that Dow lacked the intent to share the profits and losses of the Chemtech transactions with the foreign banks. We therefore affirm the district court's sham partnership holding and do not reach its economic substance holding or its holding classifying the interest as debt.

V.

Section 6662 of the Internal Revenue Code imposes a twenty-percent penalty to "the portion of any underpayment which is attributable to 1 or more of the following: (1) [n]egligence or disregard of rules or regulations[,] (2) [a]ny substantial understatement of income tax[, or] (3) [a]ny substantial valuation misstatement under chapter 1...." 26 U.S.C. § 6662(a), (b)(1)-(3). The Code increases the penalty to forty percent of the underpayment for a "gross" valuation misstatement.32 Id. § 6662(h). The Code does not allow penalties to be stacked, even if more than one penalty applies.33

The district court imposed twenty-percent penalties for negligence and substantial understatement but declined to impose either the substantial-valuation or gross-valuation misstatement penalties. The court believed that it could not impose a valuation-misstatement penalty when an entire transaction had been disregarded (here under the economic substance doctrine). The court relied on Heasley v. Commissioner, 902 F.2d 380 (5th Cir. 1990).34

After the district court issued its order, the Supreme Court decided United States v. Woods, ___ U.S. ___, 134 S.Ct. 557, 187 L.Ed.2d 472 (2013), which rejects the Heasley rule:

Id. at 566. Therefore, the district court erred in foreclosing the applicability of both the substantial-valuation and gross-valuation misstatement penalties. We remand for the court to determine whether to impose either or both of those penalties. We express no opinion on whether the court erred in imposing the negligence and substantial-understatement penalties. On remand, the court should consider the extent to which imposing those penalties remains consistent with this opinion.

The judgment is AFFIRMED in part and VACATED and REMANDED in part. In so deciding, we limit our reasoning to the specific facts and transactions at hand.