Duquesne Light Holdings Inc v. Commissioner of Internal Reven

861 F.3d 396, 2017 WL 2802758, 120 A.F.T.R.2d (RIA) 2017, 2017 U.S. App. LEXIS 11595

• Court: Court of Appeals for the Third Circuit
• Decided: June 29, 2017
• Docket: 14-1743
• Status: Published

Opinions

OPINION OF THE COURT

AMBRO, Circuit Judge

This appeal concerns the continued vitality of the so-called Ilfeld doctrine for interpreting the Internal Revenue Code. Taken from Charles Ilfeld Co. v. Hernandez, 292 U.S. 62, 54 S.Ct. 596, 78 L.Ed. 1127 (1934), this doctrine teaches that “the Code should not be interpreted to allow [the taxpayer] ‘the practical equivalent of a double deduction’ ... absent a clear decía-, ration of intent by Congress.” United States v. Skelly Oil Co., 394 U.S. 678, 684, 89 S.Ct. 1379, 22 L.Ed.2d 642 (1969) (quoting Ilfeld, 292 U.S. at 68, 54 S.Ct. 596). Duquesne Light Holdings, Inc. (“DLH”) and its subsidiaries (DLH and its subsidiaries are variously referred to as the “Du-quesne group,” the “Duquesne entities,” or simply “Duquesne”) appeal the Tax Court’s grant of summary judgment to the Internal Revenue Service based on the Ilfeld doctrine. In particular, Duquesne challenges the Tax Court’s holding that the consolidated entities affiliated with DLH claimed a double deduction for certain losses incurred by its AquaSource subsidiary and thus disallowed $199 million of those losses (all numbers are rounded). As we conclude that the Tax Court properly applied the Ilfeld doctrine, we affirm.¹

I. BACKGROUND

The Duquesne entities, including DLH and AquaSource, filed their tax returns as a consolidated taxpayer, meaning they filed a single tax return reflecting their joint tax liability. Despite allowing corporate groups to file a single return, the applicable tax laws require a mixed approach of calculating some aspects of the group’s taxes as though the entities were a single taxpayer and calculating others as if each member of the group were a separate taxpayer. 13 Mertens Law of Federal Income Taxation § 46:1 (2016 ed.). This approach' — called the “consolidated return regime” — reflects how the IRS has chosen to exercise its broad discretion to issue regulations for consolidated returns “to reflect the income-tax liability” of the group and “to prevent avoidance of such tax liability.” 26 U.S.C. § 1502.

The possibility of separate treatment nonetheless creates the potential for the group to deflect its tax liability by using stock sales to claim a second deduction for a single loss at the subsidiary (such as a loss on the subsidiary’s sale of an asset). See Lawrence Axelrod, 1 Consolidated Tax Returns § 18:2 (4th ed. 2015). This is known as a double deduction, or more technically as loss duplication. It may occur because by definition the parent company in a corporate group owns all or most of the stock in its subsidiaries. See 26 U.S.C. § 1504(a). All else being equal, the value of the parent’s stock depends on the value of the subsidiary’s assets. If the subsidiary’s assets decline in value, the parent’s stock will as well. If the subsidiary sells those assets (which may include stock and other securities) at a loss, it is generally able to claim a deduction for those losses. See 26 U.S.C. § 165(a) (“General rule. — There shall be allowed as a deduction any loss sustained during the taxable year and not compensated for by insurance or otherwise”). If the parent and subsidiary are viewed as separate entities, the parent is able to sell its stock of the subsidiary at a loss and claim a deduction for that loss as well. See Axelrod, swpra, § 18:2. But in fact the overall group has only suffered one economic loss though it was deducted twice. For example, suppose that parent P has a wholly owned subsidiary S and its investment in S’s stock is worth $100. After one of S’s assets declines in value by $50, S sells the asset and deducts a $50 loss under § 165. P’s stock value in S also declines by $50, and if P and S are viewed separately, P is able to sell its stock in S and deduct a further loss of $50 under § 165. The consolidated group is thus able to deduct $100 for a single economic loss of $50 resulting from the decline in value of S’s asset.

To prevent double deductions, the IRS has promulgated numerous regulations requiring that consolidated taxpayers be treated as a single entity for stock sales. Of particular relevance to the events of this case is the former Treas. Reg. § 1.1502-20. Starting in the early 1990s, it prevented, among other things, double deductions when the parent’s loss on its sale of stock occurred before the subsidiary recognized its loss. See Consolidated Return Regulations; Special Rules Relating to Dispositions and Deconsolidations of Subsidiary Stock, 55 Fed. Reg. 9426-01, 9427 (Mar. 14, 1990). In July 2001, however, the Federal Circuit’s decision in Rite Aid Corp. v. United States, 255 F.3d 1357 (Fed. Cir. 2001), invalidated the pertinent portion of § 1.1502-20 as beyond the IRS’s power to issue because, it addressed a problem not specifically attributable to the filing of consolidated returns. Id. at 1360. Though Rite Aid has not been construed to annul any other consolidated return regulation preventing duplicated loss, invalidating § 1.1502-20 meant that in its immediate aftermath there was no regulation expressly preventing a double deduction when the parent’s stock loss occurred before the subsidiary’s asset loss. In contrast, Rite Aid left intact the regulatory prohibition on double deductions where the transactions are structured in such a way that the losses occur in reverse order, ie., the subsidiary’s loss is recognized before the parent’s loss. See Treas. Reg. § 1.1502-32.

In the aftermath of Rite Aid, the Du-quesne group arranged a series of transactions in which DLH incurred a loss on AquaSource stock, and then AquaSource incurred losses on the sale of its assets (which were stock interests that Aqua-Source held directly and indirectly in eight of its own subsidiaries). DLH formed AquaSource in the late 1990s as a wholly owned subsidiary to expand into the water utility business. It funded AquaSource through a series of contributions in return for AquaSource stock. Through February 2001, DLH contributed approximately $223 million in return for 50,000 shares of AquaSource stock. Though DLH contributed a similar amount to AquaSource in the years thereafter, it increased its holdings to 1.2 million shares of AquaSource stock. AquaSource used these ' contributions to purchase more than 50 water utility companies, which became both its subsidiaries and its assets. It began to lose money, however, and in 2000 the Duquesne group decided to explore the sale of Aqua-Source’s assets.

The transactions before us began on December 31, 2001, just before the deadline would expire for the IRS to file a petition for certiorari in Rite Aid. DLH transferred 50,000 shares of AquaSource stock to Lehman Brothers, which Lehman valued at $4 million, as payment for Lehman’s services rendered to AquaSource. DLH determined that these particular 50,000 shares of stock were the shares that it had possessed prior to February 2001 and thus accounted for $223 million of its investment. After various adjustments, DLH claimed a capital loss of $199 million on the transfer of stock to Lehman Brothers (the “2001 loss”). On its 2001 tax return, thé Duquesne group carried back $161 million of that loss to tax year 2000 and claimed a tentative refund of $35 million.

Shortly thereafter, the IRS announced the regulatory response to Rite Aid. It declined to litigate further the validity of Treas. Reg. § 1.1502-20 and instead announced that it would issue new regulations governing stock losses. I.R.S. Notice 2002-11, 2002-1 C.B. 526 (Jan. 31, 2002). It did so in early March 2002 by issuing temporary regulations that included Treas. Reg. § 1.337(d)-2T, which applied to stock losses occurring on or after March 7, 2002. See Loss Limitation Rules, 67 Fed. Reg. 11034-01, 11036-37 (Mar. 12, 2002).² Though these regulations “[did] not disallow [a] stock loss that reflects ... built-in asset losses of a subsidiary member,” I.R.S. Notice 2002-18, 2002-1 C.B. 644 (Mar. 9, 2002), the IRS published as guidance a draft of a new regulation barring double deductions in October 2002. See Guidance Under Section 1502; Suspension of Losses on Certain Stock Dispositions, 67 Fed. Reg. 65060-01 (Oct. 23, 2002). The new regulation was issued in final form as Treas. Reg. § 1.1502-35T and applied retroactively to stock sales occurring on or after March 7, 2002.

In this complex and shifting regulatory environment, the Duquesne group thereafter incurred further losses on the sale by AquaSouree of its assets.³ It did so through a series of transactions after March 2002 and continuing into 2003 in which AquaSouree sold all of its stock in eight subsidiaries, and thus the resulting losses were also stock losses.⁴ In 2002, these transactions resulted in capital losses for AquaSouree totaling $59.5 million (the “2002 loss”), all of which the Duquesne group carried back to tax year 2000 on its consolidated return. Based on the 2002 loss and additional carrybacks, the group received a tentative refund of $12 million. The sale transactions in 2003 yielded $192.8 million in capital losses for Aqua-Source after various adjustments (the “2003 loss”). Duquesne carried all of the 2003 losses back to 2000 on its consolidated return and, based on that loss and additional.carrybacks, received a tentative refund of $40 million. Aggregating the 2002-2003 losses, the Duquesne group deducted approximately $252 million in capital losses in addition to the $199 million it had already claimed for the 2001 loss.

Though the IRS initially declined to challenge Duquesne’s deductions in a 2004 audit, it subsequently determined that the Duquesne group had claimed a double deduction; that is, the 2001 loss and the 2002-2003 losses reflected a single economic loss in the form of an underlying decline in the value of AquaSource’s subsidiaries. Based on its calculations, the IRS concluded that $199 million of these losses were disallowed under the Ilfeld doctrine.⁵ Duquesne strongly disagreed, and as the parties were unable to resolve their dispute out of court, the IRS sent Duquesne a formal notice of deficiency in 2010.

Duquesne began proceedings in the Tax Court by filing a petition for relief. Rather than following the normal pretrial procedure, the parties chose to forgo discovery and filed cross-motions for summary judgment focusing on whether the Ilfeld doctrine applied. While these motions were pending, the Tax Court issued its decision in Thrifty Oil Co. v. Commissioner, 139 T.C. 198 (2012). There the Court relied on Ilfeld to disallow certain duplicative losses that a consolidated taxpayer had claimed based on environmental remediation expenses. As a result, the Tax Court allowed the parties to this case to file supplemental briefs addressing Thrifty Oil.

Thereafter the Tax Court granted the IRS’s motion for summary judgment and concluded that the Ilfeld doctrine remained good law in our Circuit. Turning to the factual record, the Tax Court concluded that the IRS met its burden to show that the 2002-2003 losses duplicated the $199 million deduction taken in 2001, as they reflected the same economic loss. It also rejected Duquesne’s argument that it satisfied the Ilfeld doctrine because its deductions were authorized under the applicable statutes and regulations as well as its further argument that the statute of limitations had, in any event, lapsed. The Tax Court thus disallowed $199 million of the 2002-2003 losses and ordered Duquesne to repay $36.9 million of the refunds it had received based on those losses. This appeal followed.

II. ANALYSIS

Did the Tax Court properly rely on the Ilfeld doctrine to disallow $199 million of the 2002-2003 losses? Duquesne contends that the Court erred for three reasons: (1) the factual record was inadequate to support summary'judgment; (2) the Ilfeld doctrine does not support disallowing the losses; and (3) the IRS’s claims are at least partially barred by the statute of limitations. We consider each contention in turn.

A. Adequacy of the Record

As the Ilfeld doctrine requires a clear declaration of intent to allow a double deduction for a single economic loss, its application is premised on a factual question: Did the deductions claimed by the taxpayer reflect the same economic loss? See Thrifty Oil, 139 T.C. at 212; see also Denton & Anderson Co. v. Kavanagh, 164 F.Supp. 372, 378 (E.D. Mich. 1958). Here the question is whether the 2001 loss and the 2002-2003 losses are truly duplica-tive — that is, did they both reflect the decline in value of the same AquaSource assets? The Tax Court held that the IRS proved the absence of a genuine dispute of material fact on this point and that $199 million of the losses were deducted twice. Though Duquesne contends that the IRS failed to submit sufficient evidence to show the existence of a double deduction or its amount, we disagree in light of Duquesne’s failure to present any evidence to the contrary despite having all of the relevant documents in its possession and the Supreme Court’s decision in McLaughlin v. Pacific Lumber Co., 293 U.S. 351, 55 S.Ct. 219, 79 L.Ed. 423 (1934).

We exercise plenary review over the Tax Court’s grant of summary judgment, Hartmann v. Comm’r of Internal Revenue, 638 F.3d 248, 249 (3d Cir. 2011) (per curiam), and the summary judgment standard in Tax Court Rule 121 is identical to that contained in Federal Rule of Civil Procedure 56, see Rivera v. Comm’r of Internal Revenue, 89 T.C. 343, 346 (1987). The party moving for summary judgment bears the initial burden to show the.absence of a genuine dispute of material fact, see Celotex Corp. v. Catrett, 477 U.S. 317, 323, 106 S.Ct. 2548, 91 L.Ed.2d 265 (1986), and thus all that remains is to resolve legal issues.

Though the initial burden of the IRS may be heavy in some cases, it is far lighter when, as here, it seeks to collect unpaid taxes. “It is well established that[,] as a general matter, the [IRS] ’s determination of deficiency is presumed correct, and the taxpayer bears ‘the burden of proving it wrong.’ ” Cebollero v. Comm’r of Internal Revenue, 967 F.2d 986, 990 (4th Cir. 1992) (quoting Welch v. Helvering, 290 U.S. 111, 115, 54 S.Ct. 8, 78 L.Ed. 212 (1933)). “This presumption is a procedural device that places the burden of producing evidence to rebut the presumption on the taxpayer.” Anastasato v. Comm’r of Internal Revenue, 794 F.2d 884, 886 (3d Cir. 1986). This is so because “an income tax deduction is a matter of legislative grace and ... the burden of clearly showing the right to the claimed deduction is on the taxpayer.” INDOPCO, Inc. v. Comm’r of Internal Revenue, 503 U.S. 79, 84, 112 S.Ct. 1039, 117 L.Ed.2d 226 (1992). The IRS is thus essentially in the position of a defendant in a civil case who may meet its initial burden merely by pointing to the absence of evidence supporting essential elements of the non-moving party’s case. See Celotex, 477 U.S. at 322-23, 106 S.Ct. 2548; see also Gal-Farm, Ins. Co. v. United States, 647 F.Supp. 1083, 1086-87 (E.D. Cal. 1986) (collecting cases stating that the presumption alone is sufficient to support summary judgment when the taxpayer fails to rebut it); Mitchell v. Comm’r of Internal Revenue, 38 T.C.M. (CCH) 854 (T.C. 1979) (describing the IRS’s deficiency determination as “prima facie correct” and concluding that it must be sustained where “[the taxpayers] have presented no evidence to show that they are entitled to additional deductions”) (citing Tax Ct. R. 142(a)).

Duquesne now challenges the adequacy of the record before the Tax Court, yet it made numerous tactical decisions to limit that record. While conceding that it has possession of the records necessary to determine whether its deductions reflect the same economic loss, Oral Arg. Trans, at 19-20 (Nov. 18, 2015), it nonetheless agreed to limit the scope of discovery before summary judgment in order to conserve resources. Duq. Br. at 43. It then filed a motion for summary judgment raising “only legal issues,” notably whether the Ilfeld doctrine applied. Id. at 37 n.10.

The IRS responded with a cross-motion for summary judgment. It pointed to evidence initially filed in an exhibit to Du-quesne’s motion regarding the size and timing of the deductions. While Duquesne challenged the method the IRS used to calculate the amount of duplicated losses, it did not present any evidence rebutting the latter’s claim of a double deduction and merely promised that “[it] w[ould] challenge at trial the [IRS’s] erroneous determination' of the amount of purported duplicated losses.” J.A. at 670. After allowing the parties to submit supplemental briefs on the effect of the Thrifty Oil decision, the Tax Court held that the IRS had pointed to sufficient evidence of duplicate losses; hence Duquesne’s “bald assertions” of a factual dispute were too little to merit a trial. Id. at 45.

Duquesne claims that the Tax Court applied an “irrebuttable presumption” based on Thrifty Oil that the deductions were duplicative. Reply Br. at 18. While we may affirm on any basis supported by the record, it is worth noting that this argument misrepresents the Court’s reasoning. Though it stated that Duquesne’s deduc-. tions “represent the same economic loss[es],” J.A. at 42, borrowing this turn of phrase from Thrifty Oil was not invoking an irrebuttable presumption. Nor did its conclusion that Thrifty Oil was consistent with Third Circuit precedent become a presumption that barred rebuttal. Instead, the Tax Court determined that the IRS had met its burden based on the evidence in the record, including the size and timing of the deductions at issue. There is thus nothing irrebuttable in the Court’s analysis; it concluded that Duquesne simply failed to rebut the IRS’s claims as required by ordinary summary judgment practice.

Though Duquesne contends that it was not required to present any evidence because the IRS’s case contained “unexplained gaps,” Reply Br. at 15 (quoting O’Donnell v. United States, 891 F.2d 1079, 1082 (3d Cir. 1989)), we agree with the Tax Court that the IRS demonstrated the absence of a genuine dispute of material fact. As noted, it could have met its burden here merely by pointing to the absence of evidence supporting Duquesne’s position that the losses were not duplicative.

The Supreme Court’s 1934 decision in McLaughlin also supports the IRS. There a subsidiary between 1920 and 1923 had claimed operating losses three times greater than the capital invested in it by the parent, and the parent had attempted' to claim further investment losses when it liquidated the subsidiary. The Court reasoned that “the circumstances tend strongly to indicate” that the losses had a common source in the failing business of the subsidiary. Id. at 357, 55 S.Ct. 219. In addition, as “[presumably [the taxpayer] had within its control the records showing facts that would fully disclose the relations between such losses,” if they were not duplicative “it reasonably may be presumed that [the taxpayer] would have shown that fact.” Id. at 356-57, 55 S.Ct. 219. The Court thus upheld the determination of the IRS that the parent’s losses should be disallowed.

Here the IRS’s evidence regarding the size and timing of the losses similarly reveals that Duquesne claimed losses significantly greater than its net investment in AquaSource. Aggregating the 2001 loss and the 2002-2003 losses as provided on AquaSource’s Stock Sales Adjustment calculation sheet, J.A. at 744, the IRS determined the Duquesne group deducted far more in aggregate capital losses than its net investment in AquaSource, the difference being $281 million. As in Pacific Lumber, the excess of losses over investment occurred over a few years before the subsidiary ceased operation. This “tend[s] strongly to indicate” a double deduction stemming from the declining value of AquaSource’s assets. Id. at 357, 55 S.Ct. 219. Because Duquesne has possession of the relevant documents, we presume that it would have demonstrated that the losses came from different sources if that were the case.⁶ See id. at 356-57, 55 S.Ct. 219. Though it attempts to distinguish Pacific Lumber on the grounds that it was not decided at summary judgment and involved a less complex set of corporate relationships, that is off target in light of the low burden the IRS faced here.

The IRS also met its burden to show the amount of duplicative losses. As the aggregate excess of deducted losses over net investment implies a double deduction, the IRS compared the amount of excess to each deduction. That excess was greater than the deduction claimed for the 2001 stock loss ($199 million), and thus the IRS concluded that at least the amount of the 2001 stock loss was deducted twice. This calculation of $199 million in duplicated loss thus abandoned any IRS claim in addition to that loss (and, as in Pacific Lumber, it reflected the parent’s claimed loss in the transaction). Duquesne cites no case in which greater proof was required, nor does it present any evidence that the $199 million figure is inaccurate.⁷ We thus see no unexplained gaps in the IRS’s case that renders summary judgment in its favor improper.

Duquesne nonetheless contends that the IRS had to ■ “trace” the losses between Duquesne’s sale of AquaSouree stock and the losses incurred on the sale of particular AquaSouree assets (that is, the sale of particular AquaSouree subsidiaries). Tracing is necessary, Duquesne asserts, to rule out the possibility that unspecified intervening event(s) somehow account for a portion of the 2002-2003 asset losses. But because Duquesne fails to present evidence of any such events or of any other effect that tracing would have on the amount of duplicated loss, this is merely the kind of speculation that does not defeat summary judgment.⁸

In the alternative, Duquesne contends that the District Court abused its discretion by failing to order further discovery before deciding the IRS’s summary judgment motion. Like Federal Rule of Civil Procedure 56(d), Tax Court Rule 121(e) confers discretion on the trial court to order further discovery when the non-moving party files a motion or affidavit stating that more discovery is needed for it to justify its opposition to summary judgment. Duquesne, however, never claimed that it could not justify its opposition without further discovery, and thus it cannot claim the protection of Rule 121(e).

Despite its failure to comply with the Rule, Duquesne contends that our decision in Sames v. Gable, 732 F.2d 49 (3d Cir. 1984), required the Tax Court to order further discovery anyway. There two former police officers filed a civil rights claim against local police officials for retaliatory discharge. While their interrogatory requests were pending, the defendants moved for summary judgment and the plaintiffs failed to request further discovery under Rule 56(d) (then codified as Rule 56(f)). Relying on the well-settled rule that summary judgment should not be granted while the material facts remain in the moving party’s possession, we held “that it was error for the district court to grant defendants’ motion for summary judgment while pertinent discovery requests were outstanding” despite the plaintiffs’ failure to comply with the Federal Rules. Id. at 51-52. Here, however, Du-quesne had possession of the relevant facts and chose to limit discovery for reasons only it knows. In sum, it gambled that its legal arguments were strong enough to win without creating the factual record necessary to rebut the IRS’s position. After that gamble failed, Sames did not require the.District Court to delay summary judgment.

We thus conclude that the Tax Court properly held that $199 million of the 2001 loss and 2002-2003 losses were deducted for the same underlying economic loss. After determining that there was no genuine dispute that $199 million in deducted losses for 2002-2003 were duplicative, the Tax Court disallowed'that amount of those losses under the Ilfeld doctrine.⁹ We thus turn to Duquesne’s next contention: that although its consolidated deductions maybe duplicative, the Ilfeld doctrine nonetheless does not support the Tax Court’s decision to disallow them.

B. The Ilfeld Doctrine

Ilfeld requires a clear declaration allowing double deductions for the same loss on consolidated returns. Duquesne contends that the text of 26 U.S.C. § 165 provides the required authority.to satisfy Ilfeld; if § 165 alone proves insufficient, the combination of it with the applicable regulations in effect provides clear authority; and in any event the regulations alone — particularly § 1.337(d)-2T — suffice. In evaluating these arguments we review anew the Tax Court’s legal conclusions. See Anderson v. Comm’r of Internal Revenue, 698 F.3d 160, 164 (3d Cir. 2012).

1. The Current Status of the Ilfeld Doctrine

Duquesne does not dispute that Ilfeld remains good law. While there is dispute as to the scope of the Ilfeld doctrine — that is, whether it applies outside the consolidated return context — for consolidated taxpayers it continues to require that a statute and/or regulation specifically authorize a double deduction for an underlying economic loss.

Ilfeld, decided in 1934, concerned consolidated tax returns filed by Charles Ilfeld Co. and two wholly owned subsidiaries between 1917 and 1929. During this period, the subsidiaries claimed substantial operating losses and were eventually liquidated. Ilfeld Co. then attempted to deduct losses on its investment in the defunct subsidiaries. Though these investment' losses could not be deducted under the regulations in effect at the time, the Supreme Court emphasized that, because the investment losses were already reflected in the subsidiaries’ operating losses, allowing them “would be the practical equivalent of a double deduction.” Ilfeld, 292 U.S. at 68, 54 S.Ct. 596. Thus, “[i]n the absence of a provision of the [applicable statute] definitely requiring it, a purpose so opposed to precedent and equality of treatment of taxpayers will not be attributed to lawmakers.” Id. “[D]efinitely requiring” a provision to authorize a double deduction for the same economic loss is a very high hurdle, though how high is debatable when Ilfeld itself stated that “[i]n the absence of a provision in the [applicable act] or regulation that fairly may be read to authorize [a double deduction], the deduction claimed is not allowed.” Id. at 66, 54 S.Ct. 596.

How to interpret this disparity in the degree of certainty was taken care of when the Supreme Court restated Ilfeld as requiring a “clear declaration of intent by Congress” to authorize a double deduction in Skelly Oil Co., 394 U.S. at 684, 89 S.Ct. 1379 (citing United States v. Ludey, 274 U.S. 295, 47 S.Ct. 608, 71 L.Ed. 1054 (1927)). Though a declaration by Congress is stated, we do not purport to rule out clear statements of intent set out in regulations the IRS Commissioner is empowered to prescribe. See Ilfeld, 292 U.S. at , 54 S.Ct. 596. Indeed, Ilfeld itself noted that because “[t]here [was] nothing in the [Revenue Act of 1928] that purported] to authorize double deduction of losses or in the regulations to suggest that the commissioner construed any of its provisions to empower him to prescribe a regulation that would permit consolidated returns to be made on the basis now claimed by [Ilfeld Co.],” it could not deduct its dupli-cative investment losses. Id. (emphasis added).

Ilfeld quickly became a key precedent in the consolidated return context. In Pacific Lumber the Supreme Court relied on Il-feld for the principle that “[l]osses of [the subsidiary] that were subtracted from [the group’s] income are not directly or indirectly again deductible.” 293 U.S. at 355, 55 S.Ct. 219. Our Court similarly relied on Ilfeld to disallow deductions for consolidated returns. See Greif Cooperage Corp. v. Comm’r of Internal Revenue, 85 F.2d 365, 366 (3d Cir. 1936); see also Comm’r of Internal Revenue v. National Casket Co., 78 F.2d 940, 941-42 (3d Cir. 1935).

Some of our sister Circuits have extended the Ilfeld doctrine beyond the consolidated-return context — see, e.g., Chicago & N.W.R. Co. v. Comm’r of Internal Revenue, 114 F.2d 882, 887 (7th Cir. 1940) (appropriate method of depreciation for railroad property); Comar Oil Co. v. Helvering, 107 F.2d 709, 711 (8th Cir. 1939) (deductibility of anticipated inventory losses); Marwais Steel Co. v. Comm’r of Internal Revenue, 354 F.2d 997, 997-99 & n.1 (9th Cir. 1965) (investment losses among corporate taxpayers filing separate returns).

All of the cases questioning the continued viability of Ilfeld have occurred outside the consolidated-return context. For example, in a case in our Circuit—Miller’s Estate v. Commissioner, 400 F.2d 407 (3d Cir. 1968) — the IRS argued that certain deductions flowing from estates’ donations to charitable trusts were double deductions. We distinguished Ilfeld on the ground that it concerned “the peculiar income tax context of consolidated corporate income tax reporting,” and thus it “cannot be regarded as a legitimate canon of estate tax interpretation to assist the court in this case.” Id. at 411 (internal footnotes omitted).

The Supreme Court indicated that Ilfeld might apply beyond consolidated returns the year following Estate of Miller in Shelly Oil, where it required a non-consolidated taxpayer to recalculate certain refunds based on oil and gas revenues. 394 U.S. at 684, 89 S.Ct. 1379. But decades later the Court in Gitlitz v. Commissioner, 531 U.S. 206, 121 S.Ct. 701, 148 L.Ed.2d 613 (2001), may have implied that Ilfeld does not apply other than for consolidated returns. Addressing an argument abandoned by the IRS — that, if shareholders of an S corporation (in which income is passed through to shareholders and then taxed as personal income) were permitted to pass through a discharge of indebtedness before reducing tax attributes, the shareholders may have gotten a “double windfall” by being partially exempted from taxation yet able to increase their basis and deduct previously suspended losses — the Court elected not to address this “policy concern.” Id at 219-20, 121 S.Ct. 701. Though Justice Breyer’s dissent criticized the majority for failing to apply Ilfeld, see id. at 224, 121 S.Ct. 701, the majority did not so much as mention that decision or Shelly Oil; it certainly did not purport to overrule them. In any event, after Gitlitz the Ilfeld doctrine thus remains good law in the consolidated-return context.¹⁰

With this in mind, we proceed to what is currently required for a statute to satisfy the Ilfeld doctrine. It reflects “[t]he presumption ... that statutes and regulations do not allow a double deduction” for the same economic loss. United Telecomm. v. Comm’r of Internal Revenue, 589 F.2d 1383, 1387-88 (10th Cir. 1978). This presumption must be overcome by a clear declaration in statutory text or a properly authorized regulation.

2. Deduction of the 2002-2003 Losses Under § 165

Duquesne argues that a specific authorization for duplication of loss here may be found in the text of § 165. In particular, it points to subsections (a) and (f) of that provision. As these subsections “provide! ] for deduction of losses, including capital losses,” and the 2002-2003 losses were incurred on AquaSource’s sale of stock in its subsidiaries, Duquesne contends that the “deductions fall squarely within § 165’s text” even if they are duplicative. Duq. Br. at 15. We conclude otherwise.

The general rule of § 165(a) is that “[t]here shall be allowed as a deduction any loss sustained during the taxable year and not compensated for by insurance or otherwise.” 26 U.S.C. § 165(a). That the subsection allows “a deduction” for “any loss” indicates that it allows a single deduction for a single loss. Id. (emphases added). The brief text of the subsection certainly contains no express authorization of a double deduction, and we are unaware of any evidence of congressional intent to that effect. Moreover, as the Tax Court noted, § 165(a) is quite broad in authorizing a deduction for any loss and is thus a poor candidate to satisfy the Ilfeld requirement of explicit approval for duplicating the underlying economic loss. See J.A. at 46-47.

Section 165(f)’s limitation on the deducti-bility of capital losses also does not support Duquesne’s position. It provides that “[l]osses from sales or exchanges of capital assets shall be allowed only to the extent allowed in sections 1211 and 1212.” 26 U.S.C. § 165(f). 26 U.S.C. § 1211(a) provides that corporations may deduct capital losses only to offset capital gains from the sale of different assets, and section 1212(a)(1)(A) allows corporations to carry back capital losses up to three years. These limitations are irrelevant to the prospect of a double deduction for the same economic loss, and in any event they are not at issue here.

Beyond the plain text of the statute, Ilfeld’s case-specific holding sends a particularly strong signal that § 165 does not authorize a double deduction. At the time Ilfeld was decided, the Revenue Act of 1928 was in effect and contained a provision indistinguishable from the current § 165(a): “In computing net income there shall be allowed as deductions: ... In the case of a corporation, losses sustained during the taxable year and not compensated by insurance or otherwise.” Revenue Act of 1928, ch. 852, § 23(f), 45 Stat. 791, 799-800. The IRS brought this provision to the Court’s attention and specifically argued that it did not allow Ilfeld Co. to claim a double deduction. See Brief for Respondent at 22, Charles Ilfeld Co. v. Hernandez, 292 U.S. 68 (No. 579), 1933 WL 63349 at *22. To uphold that assertion, Ilfeld had to reject any assertion that this provision demonstrated a clear intent to allow it.

While Duquesne contends that any consideration of the Revenue Act by the Supreme Court was a mere dictum because it had already determined that the taxpayer’s deductions were barred by applicable regulations, this contention runs into Pacific Lumber, which also interpreted virtually identical statutory text. As no applicable regulations governed the specific deductions, the Court determined whether they were allowed under the applicable statute, the Revenue Act of 1921. That Act contained the following provision: “[For purposes of the corporate income tax], there shall be allowed as deductions: ... Losses sustained during the taxable year and not compensated for by insurance or otherwise ...” Revenue Act of 1921, ch. 136, § 234(a)(4), 42 Stat. 227, 254-55. The IRS also brought this provision to the Court’s attention in arguing that it did not authorize a double deduction. See Brief for Petitioner at 20, McLaughlin v. Pacific Lumber Co., 293 U.S. 351 (No. 125), 1934 WL 60331 at *20. The bottom line: to hold that the taxpayer’s losses were not deductible a second time, Pacific Lumber rejected a provision materially identical to § 165(a) without any help from the consolidated return regulations. Ilfeld and Pacific Lumber's rejection of such a strikingly similar predecessor to § 165 as a rationale for a double deduction thus provides further support for our conclusion that the current statute does not provide that rationale.

We thus turn to Duquesne’s next argument: that even if § 165 alone does not satisfy Ilfeld’s requirement of a clear statement of intent to authorize a double deduction,’ the combination of that provision with the applicable regulations in effect does so.

3. Deduction of the 2002-2003 Losses Under § 165 and Applicable Regulations

Duquesne asserts that it complied with all applicable regulations in calculating AquaSource’s losses and then deducted them under § 165. In particular, it calculated the amount that AquaSource lost on sales of its subsidiaries’ stock per the regulations governing stock losses, primarily Treas. Reg. §§ 1.1502-32 and 1.337(d)-2T, before deducting those losses under § 165. As Duquesne thus purports to have complied with the requirements of both on-point regulations and an on-point statute, it contends that the Tax Court’s decision in Woods Investment Co. v. Commissioner, 85 T.C. 274 (1985), teaches that Ilfeld’s clear statement rule is satisfied and we should rule in its favor by applying the regulations and the Code “as written.” Id. at 282.

In light of the complexity of the regulatory framework, Duquesne’s argument requires some unpacking. The first step in determining the amount of the 2002-2003 losses was calculating AquaSource’s basis in the stock of its subsidiaries. Under 26 U.S.C. § 1001(a), the amount of loss for tax purposes on the sale of an asset, including stock, is equal to the taxpayer’s basis minus the sale price. For stock sales by consolidated taxpayers, such as Aqua-Source, Tax Regulation § 1.1502-32 requires the parent to adjust its basis to reflect various events at the subsidiary, including gains and losses. It is undisputed that the Duquesne group made all the adjustments required by § 1.1502-32 to AquaSource’s basis in its subsidiaries’ stock and calculated the amount of loss for tax purposes accordingly. ■

The Duquesne group then calculated the amount of allowable loss on AquaSource’s stock sales per § 1.337(d)-2T. Paragraph (a)(1) of the regulation provides the general rule that “[n]o deduction is allowed for any loss recognized by a member of a consolidated group with respect to the disposition of stock of a subsidiary.” Paragraph (c)(2), governing “[ajllowable [floss,” then gives a partial exception to this general rule: “Loss is not disallowed under paragraph (a)(1) of this section ... to the extent the taxpayer establishes that the loss ... is not attributable to the recognition of built-in gain on the disposition of an asset (including stock and securities).” Treas. Reg. § 1.337(d)-2T(c)(2). A built-in gain occurs when an asset increases in value before being sold. As AquaSource’s subsidiaries had declined in value, none of the 2002-2003 losses reflected built-in gains.

With this backdrop to Duquesne’s contention of statutory/regulatory compliance in conjunction with its reading of Woods, the Government asserts that Woods does not control, as here no statute or regulation permitted a double deduction with anything approaching the specificity that compelled the outcome in that case. In Woods the issue was whether the Ilfeld doctrine applied to the calculation of a parent’s basis in its subsidiary’s stock. At the time Regulation § 1.1502-32 required the parent to adjust its basis in line with the subsidiary’s earnings and profits, and 26 U.S.C. § 312(k) required the parent to calculate those earnings and profits using a straight-line method of depreciation. Despite its awareness of the potential for straight-line depreciation to result in the practical equivalent of a double deduction, the IRS had failed to amend § 1.1502-32 for almost twenty years. The Tax Court reasoned that the details of the consolidated return regime were “essentially a legislative and administrative matter,” and thus it was not a court’s institutional role to engage in “judicial interference.” Woods, 85 T.C. at 282. It thus held that it would “apply [the consolidated return] regulations and the statute as written”: if the IRS disliked this result, it “should use [its] broad power to amend [its] regulations.” Id. Ilfeld did not support the IRS’s position because “the detailed rules in [§ 1.1502-32], which were enacted to comprehensively address the problem in [Il-feld], together with section 312(k), can fairly be read to authorizé the result herein ....” Id. at 283 (internal footnote omitted). Woods thus teaches that when an on-point statute and an on-point regulation authorize the taxpayer to act as it did, courts should apply the statute and regulation as written.

The Tax Court clarified in Wyman-Gordon Co. v. Commissioner, 89 T.C. 207 (1987), that Woods may not apply in the absence of a statute and/or regulation clearly authorizing the disputed deductions. At issue was whether Code § 312(1) and Regulation § 1.1502-32 authorized the taxpayer to include discharge of indebtedness income in the calculation of its subsidiaries’ earnings and profits. The Court distinguished Woods on the ground that the latter’s holding “was based in large part on § 312(k), which specifically require[d]” the taxpayer to compute its basis as it had, whereas “there exist[ed] no comparable statutory provision that require[d] inclusion of discharge of indebtedness income in earnings and profits.” Id. at 219. As no specific provision of any regulation, including § 1.1502-32, supported the taxpayer’s position, the Court disallowed the deductions as an impermissible duplication of loss. Id. at 224. Though there was thus no on-point regulation at issue in Wyman-Gordon, it nonetheless indicates that Woods’ instruction to apply the Code and regulations as written requires a statute and/or regulation affirmatively permitting the taxpayer to act as it did.¹¹

Section 165, however, did not specifically authorize Duquesne to claim a duplicative deduction of the 2002-2003 losses. As discussed above, its subsection (a) broadly allows a single deduction for any loss, but it does not contemplate the possibility of a double deduction. Moreover, as a materially identical statute was at issue in Ilfeld and Pacific Lumber, those decisions send a particularly strong signal that § 165 does not authorize a double deduction. Thus it is out of play for coupling it to potentially applicable regulations.

4. Deduction-of the 2002-2003 Losses Under Solely the Regulations

Duquesne counters that, even absent § 165, its regulatory compliance with §§ 1.1502-32 and 1.337(d)~2T alone was sufficient to overcome any effect of Ilfeld.12 Yet we are not persuaded that these regulations are sufficiently clear to preclude application of the Ilfeld doctrine. At the outset, § 1.1502-32 does not support Du-quesne’s position because its basis adjustments address the prospect of a duplicated loss in a transaction where the subsidiary recognizes its loss before the parent engages in a stock sale. Consolidated Return Regulations; Special Rules Relating to Dispositions and Deconsolidations of Subsidiary Stock, 55 Fed. Reg. 9426-01, 9427 (Mar. 14, 1990). Where, as here, the parent’s stock sale occurs first, the prospect of a double deduction has been addressed by other regulations since § 1.1502-20 was issued in the early 1990s. Id.

We thus turn to the other regulation on which Duquesne relies, § 1.337(d)-2T. As noted above, subsection (a)(1) generally forbids any deduction for losses incurred on sales of subsidiary stock, and subsection (c)(2) provides that “[l]oss is not disallowed under paragraph (a)(1) ... to the extent the taxpayer establishes that the loss ... is not attributable to the recognition of built-in gain on the disposition of an asset (including stock and securities).” From this double negative (“not disallowed”) on a negative (“not attributable”), Duquesne asks us to infer that any stock losses not reflecting a built-in gain, including duplica-tive losses, are deductible.

We do not believe that the structure and purpose of the broader regulatory regime support Duquesne’s interpretation of § 1.337(d)-2T. Consolidated taxpayers pay based on the consolidated taxable income (CTI) of the group. United Dominion Industries, Inc. v. United States, 532 U.S. 822, 826, 121 S.Ct. 1934, 150 L.Ed.2d 45 (2001). CTI is calculated by combining the separate taxable income (STI) of each member of the group and then incorporating certain adjustments calculated on a consolidated basis. Id. Adjustments on a consolidated basis are an example of the consolidated return regime adopting the “single entity” approach to prevent distortion of tax liability. See American Standard, Inc. v. United States, 602 F.2d 256, 261-62 (Ct. Cl. 1979). This treats the entire consolidated group as a single taxpayer and “reduce[s] the significance of each member’s separate existence.” Don Leatherman, Why Rite-Aid Is Wrong, 52 Am. L. Rev. 811, 815-16 (2003). The Ilfeld doctrine also reflects a single-entity approach by preventing the group as a whole from claiming duplicative deductions that the separate existence of parent and subsidiary would otherwise allow. See Axelrod, supra, § 18:2 (4th ed. 2015).

This structure for consolidated returns indicates the intent that the 2002-2003 losses be dealt with on a consolidated basis, and thus it reflects the same single-entity approach as the Ilfeld doctrine. Section 1.1502-32, on which Duquesne relies, has the express purpose of treating the group as “a single entity.” Treas. Reg. § 1.1502-32(a)(l). And Treas. Reg. §§ 1.1502-ll(a)(3) and 1.1502-22 provide that net capital gains and losses — that is, the overall amount of capital gains or losses in a particular year — are those items calculated on a consolidated basis.

The calculation of individual stock losses also occurs on a consolidated basis, see Treas. Reg. 1.1502-12(j), and for nearly thirty years these calculations have followed detailed rules designed to prevent tax evasion. Before Rite Aid, § 1.1502-20 limited stock losses básed on various loss disallowance factors. See Corporations; Consolidated Returns — Special Rules Relating to Dispositions and Deconsolida-tions of Subsidiary Stock, 56 Fed. Reg. 47379-01, 47379 (Sept. 19, 1991). Those factors primarily disallowed stock losses reflecting a built-in gain in the subsidiary’s assets. See Consolidated Return Regulations; Special Rules Relating to Dispositions and Deconsolidations of Subsidiary Stock, 55 Fed. Reg. 9426-01, 9427 (Mar. 14, 1990). Though the sale of a capital asset with a built-in gain generally results in tax liability, see 26 U.S.C. § 1231, a formerly common type of transaction allowed consolidated taxpayers to arrange a stock loss to offset that gain and thereby avoid tax liability.¹³ Section 1.1502-20 also had the distinct purpose of preventing double deductions, see Consolidated Return Regulations, 55 Fed. Reg. at 9427, which was reflected in the duplicated loss factors. Specifically, it prevented double deductions when, as here, the parent’s stock loss occurred before the subsidiary recognized a loss. Id.

But because Rite Aid invalidated § 1.1502-20, the IRS had to issue new regulations limiting stock losses. In doing so, it separated the rules for stock losses occurring on and after March 7, 2002, primarily into two temporary regulations: §§ 1.337(d)-2T and 1.1502-35T.¹⁴ As noted, § 1.337(d)-2T(e)(2)’s exception to the general- ban on losses for subsidiary stock sales in consolidated tax regimes requires the taxpayer to prove that a loss is not the result of a built-in gain.

Though the IRS acknowledged that § 1.337(d)-2T “d[id] not disallow stock loss[es] that reflect! ] ... built-in asset losses of a subsidiary member,” it informed taxpayers that it “and Treasury believe that a consolidated group should not be able to benefit more than once from one economic loss” and would issue another regulation to prevent double deductions. I.R.S. ' Notice 2002-18, 2002-1 C.B. 644 (Mar. 9, 2002). The IRS thus had the prospect of issuing another regulation to bar double deductions when it issued § 1.337(d)-2T, and doing so was consistent with the consolidated return practice of imposing multiple restrictions on stock losses (as in the former § 1.1502-20).

■ Section 1.1502-35T was the pertinent regulation, and it was expressly intended “to prevent a group from obtaining more than one tax benefit from a single economic loss.” Treas. Reg. § 1.1502-35T(a). It accomplished this through a loss suspension rule:

Any loss recognized by a member of a consolidated group with respect to the disposition of a share of subsidiary member stock shall be suspended to the extent of the duplicated loss with respect to such share of stock if, immediately after the disposition, the subsidiary is a member of the consolidated group of which it was a member immediately pri- or to the disposition (or any successor group).

Treas. Reg. § 1.1502-35T(e)(l) (as issued in 2003 retroactive to March 7, 2002). Yet this regulation failed to prevent a double deduction here because of how the Du-quesne group structured the relevant transactions in the wake of Rite Aid. The loss suspension rule of paragraph 1.1502-35T(c) applies only when a member of a consolidated group, such as AquaSource, sells stock in its subsidiaries and those subsidiaries remain members of the group after the sale. But because AquaSource sold all of the stock in its subsidiaries to third parties, they, were no longer members of the group after the 2002-2003 losses. Paragraph 1.1502-35T(c) thus did not prevent AquaSource from deducting the 2002-2003 losses. Though Duquesne may nonetheless claim that it complied with § 1.1502-35T in the sense that it did not violate it, we agree with the Tax Court that this is not enough to meet the clear authorization test of Ilfeld.

So this case comes down to whether, under the interpretive principle of Ilfeld, § 1.337(d)-2T clearly authorizes the losses in 2002-03 for the Duquesne consolidated group that duplicates the loss it took for 2001. Its argument is that, even if Ilfeld applies, while paragraph (a)(1) bars a deduction for any losses by any member of a consolidated group with respect to the disposition of stock by a subsidiary (here the sale by AquaSource of the entire interest in eight of its direct and indirect subsidiaries), paragraph (c)(2) does not disallow those losses if AquaSource shows that they do not result from built-in gain on that sale, and here they do not.

If literally applied to the exclusion of all other provisions of the Internal Revenue Code, Duquesne (and our dissenting colleague) make a strong argument. By regulation there is excluded —that is, disallowed — any deduction for a loss that occurs when a member of a consolidated group sells stock in a subsidiary. 26 C.F.R. § 1.337(d)-2T(a)(l). Yet that deduction disallowance “is not disallowed” under paragraph (c)(2) when the sale does not result in the recognition of built-in gain (an assertion we accept here).

So is the ease closed? No. That is because the partial excision in paragraph (c)(2) is one-time only, for there is no mention in the regulation of approval for a loss deduction that duplicates another already taken for the underlying economic loss. We know that it has nothing to do with loss duplication because, at the same time paragraph (c)(2) was issued notwithstanding its exception to loss disallowance, the Notice accompanying it warned that the “IRS and Treasury believe that a consolidated group should not be able to benefit more than once from one economic loss.” IRS Notice 2002-18, 2002-1 C.B. 644 (Mar. 9, 2002). In this context, the blinkered approach of Duquesne does not stand so long as there is Ilfeld. It remains a valid interpretative principle for consolidated returns until the Supreme Court tells us otherwise.¹⁵

We therefore confront circumstances fundamentally unlike those before the Tax Court in Woods. This is not a. case of the IRS failing to act when its own regulations and related statute specifically authorize the result sought by the taxpayer. Here, unlike Woods, the IRS has never had a regulatory scheme in place that would authorize Duquesne. to take' both deductions it has claimed for the same economic loss. And the text of § 1.337(d)-2T does not clearly allow in the future losses already taken on consolidated returns. Put another way, future losses cannot be added to past losses already deducted for the same group of assets. The dissent would have us apply Woods and hold that the IRS implicitly authorized Du-quesne’s double deductions when they were not explicitly banned. For consolidated-return taxpayers, implied authorizations are not enough; they must be clear statements that can fairly be read to allow double deductions for the same economic loss, and here that did not occur.¹⁶

C. Statute of Limitations

Finally, Duquesne contends that the IRS is time-barred from ordering' repayment of at least some of its tentative refunds resulting from the carryback of the 2002 and 2003 losses to tax year 2000. Though we agree with the Tax Court that it is difficult to understand what Duquesne is attempting to argue on this issue, its briefing on appeal enables us to discern the basic contours of the argument. Du-quesne asserts that, to the extent losses were carried back from 2002 to 2000 and it received tentative refunds for tax year 2000 as a result, 26 U.S.C. § 6501(k) places those refunds beyond the statute of limitations. As Duquesne agreed to extend the statute of limitations for losses carried back from 2003 to 2000, however, well-established law provides that the Government could demand repayment of any refunds issued for tax year 2000. As we see nothing in § 6501(k) that overcomes the effect of Duquesne’s agreement with the IRS, we reject the former’s argument that the tentative refunds here are outside the limitations period.

Section 6501 provides the framework for determining the statute of limitations. When a tax year is within the statute of limitations, it is said to be “open;” when the limitations period expires, it is “closed.” Under § 6501(a), a tax year generally remains open for three years after the taxpayer files its return.

But when a corporate taxpayer carries losses back to an earlier year, determining the statute of limitations is more complex. A loss carryback uses capital losses incurred in one year to offset capital gains in an earlier year. See 26 U.S.C. §§ 1211-12. Though the statute of limitations is typically measured from the earlier year, § 6501(h) provides that the limitations period is extended to include the period applicable to the later year from which losses were carried back. For exam-pie, if losses were carried back from 1992 to 1989, § 6501(h) would extend the limitations period for 1989 until at least 1995 (that is, three years after 1992). Subsection 6501(k), on which Duquesne relies, provides in essence that the extended period provided by § 6501(h) also applies when the taxpayer carries back losses in order to claim a tentative refund.¹⁷

Regardless whether the taxpayer carries back any losses at all, it may agree under § 6501(c)(4) to extend the limitations period for a given tax year. If the taxpayer does so, the IRS may demand payment of additional taxes for that year on any ground. See Calumet Industries, Inc. v. Comm’r of Internal Revenue, 95 T.C. 257, 278 (1990) (“So long as the period for assessment is open under some provision in the year under consideration, we have jurisdiction to consider all items that may affect that taxable year”). This rule has been applied repeatedly and specifically to allow the IRS to disallow losses carried back from a closed year. See id.; see also Pacific Transport Co. v. Comm’r of Internal Revenue, 483 F.2d 209, 214-15 (9th Cir. 1973); Mecom v. Comm’r of Internal Revenue, 101 T.C. 374, 392-93 (1993) (collecting cases), aff'd, 40 F.3d 385 (5th Cir. 1994) (table). As a result, when the taxpayer agrees to extend the statute of limitations for a particular year, the IRS may disallow within the extended period any losses carried back to that year and demand repayment of the resulting refunds.

In this case, Duquesne agreed to extend the statute of limitations for tax year 2000. After Duquesne carried back losses from 2003 to 2000 and received a tentative refund, the statute of limitations for tax year 2000 was extended until at least 2006 under § 6501(k). It is undisputed that, before that period lapsed, Duquesne agreed to extend the statute of limitations and the IRS timely served Duquesne with a notice of deficiency based on losses carried back to 2000. Though Duquesne contends that § 6501(k) somehow requires that tax year 2002 be within the statute of limitations as well, we see nothing in the subsection’s text to disturb well-settled law on the effect of a taxpayer’s agreement with the IRS. As 2000 was open by agreement with respect to the 2003 losses and the 2002 losses were carried back to 2000 as well, we conclude that the IRS is not time-barred from demanding repayment of the refunds resulting from any losses carried back to 2000.

⅜ ⅜ ⅜ ⅜ ⅜

Though it is tempting to dismiss this case as merely further proof of the “difficult and torturous path” revealed when “we are constrained to enter the labyrinthine structured tax laws,” Ambac Industries, Inc. v. Comm’r of Internal Revenue, 487 F.2d 463, 464 (2d Cir. 1973), it also serves to remind us that double deductions for consolidated taxpayers are treated differently from other aspects of tax law. Rite Aid created a gap in the regulations preventing them, and thus Duquesne concluded correctly that a door had opened in the consolidated-return regime. It took a deduction for losses it incurred in 2001. That is a ticket for only one ride; Duquesne cannot do so again for the same economic loss. It especially cannot do so when the IRS told it, when Reg. 1.337(d)-2T went effective, that a second ride on the same-loss train is closed to consolidated taxpayers. Yet that is what Duquesne attempted in the hope that not paying for that ride would go unchallenged or, if challenged, its position would be upheld by a court. Du-quesne and our dissenting colleague believe that the authority for a loss given in this temporary regulation looks forward only, with no accounting for the same loss already taken. We believe Ilfeld counsels otherwise and thus affirm the Tax Court’s judgment.

1 . The Tax Court had jurisdiction to hear this case under 26 U.S.C. § 6213 and we have jurisdiction to review its judgment per 26-U.S.C. § 7482(a)(1).

2 . The IRS also issued Treas. Reg. § 1.1502-20T, which applied retroactively to stock losses occurring between the date of the Rite Aid decision and March 7, 2002. See Loss Limitation Rules, 67 Fed. Reg. 11034-01, 11037 (Mar. 12, 2002).

3 . At oral argument before a prior appellate panel, Duquesne's counsel represented that AquaSouree had different subsidiaries in 2001 versus 2002-2003 in support of its claim that the case should be remanded for further development of the factual record. Oral Arg. Trans, at 18 (Nov. 18, 2015). When opposing counsel pointed out that the record demonstrated that AquaSource’s subsidiaries had in fact remained the same, see Oral Arg. Trans, at 26-27 (July 12, 2016), Duquesne abandoned this claim.

4 . That AquaSouree incurred stock losses on the sale of its assets was the source of much terminological confusion in the litigation of this case. As Duquesne did not make this clear in its individual briefing, and though the record remained unclear whether all of Aqua-Source’s losses were incurred on the sale of stock, the IRS acknowledged that at least 90% of the losses were stock losses. IRS Supp. Br. at 4. As this issue is not ultimately relevant to our analysis, we accept Duquesne's representation that all of AquaSource’s losses were stock losses for purposes of this appeal.

5 .The IRS also asserted that a $228 million excess loss account should be triggered for tax year 2005, which would have been taxable as income for that year. As this claim is no longer at issue, we need not decide it.

6 . Our decision in Nat’l State Bank v. Federal Reserve Bank of N.Y., 979 F.2d 1579 (3d Cir. 1992), is not to the contrary. There National State Bank ("NSB”) bore the burden of proof as plaintiff to show that the Federal Reserve Bank of New York had negligently handled certain checks that were part of a check-kiting scheme. The District Court granted summary judgment because the Federal Reserve had taken too long to inform NSB that it had lost the checks in dispute. We reversed, as there was no evidence on when the checks were lost or by whom and the Federal Reserve likely had the relevant records. This is not our case; Duquesne had the relevant records and did not produce documents to counter the IRS’s allegations of duplicate losses.

7 . In its brief to the Tax Court, Duquesne suggested that the IRS should have used a different valuation for the AquaSource subsidiaries as of December 31, 2001, which Du-quesne asserted would have yielded a different amount of duplicated loss. J.A. at 660. This argument is waived, as Duquesne failed to raise it in its opening brief on appeal. See United States v. Hoffecker, 530 F.3d 137, 162 (3d Cir. 2008). Moreover, the issue is not material, as the amount of the 2002-2003 losses must be determined by comparing AquaSource's adjusted basis in its subsidiaries’ stock to the sale price. See 26 U.S.C. § 1001(a). If this calculation uses Duquesne’s preferred valuation, it does not reduce the amount of duplicated loss. See IRS Br. at 28 n.5.

8 . To the extent Duquesne contends that tracing is still required as a matter of law, we reject it as unsupported by any authority. Though it relies on Edward Katzinger Co. v. Commissioner, 44 B.T.A. 533 (1941), aff'd, 129 F.2d 74 (7th Cir. 1942), that case held it was the taxpayer who failed to show that its deductions reflected different losses. See id. at 76. The IRS's 2007 comments on a proposal to prevent double deductions by disallowing asset losses are also of little-to-no use. See Unified Rule for Loss on Subsidiary Stock, 72 Fed. Reg. 2964-01, 2976 (Jan. 23, 2007). While the IRS rejected a proposal that would "present considerable tracing issues,” id., it was never adopted and hence is irrelevant to its burden at summary judgment.

9 . Duquesne also argues that the Tax Court erred in how it allocated the $199 million in disallowed loss between the 2002 loss and the 2003 loss. The Tax Court allocated $59 million to the 2002 loss and the remaining $140 million to the 2003 loss. J.A. at 48-49. This allocation was proposed by Duquesne, however, and agreed to by the IRS. Id. at 47-48. Moreover, as we reject below the statute-of-limitations argument made by Duquesne, the allocation of losses between 2002 and 2003 is irrelevant.

10 . In its supplemental briefing, Duquesne also points to the Supreme Court’s decision in United Dominion Industries, Inc. v. United States, 532 U.S. 822, 121 S.Ct. 1934, 150 L.Ed.2d 45 (2001), as an indication of Ilfeld’s diminished role. Though United Dominion is a consolidated-return case in which the Court declined to rule in the IRS's favor with respect to its double-deduction argument, it does not apply here because the calculation at issue “[was] not in and of itself the basis for any tax event.” Id. at 834, 121 S.Ct. 1934. What this means is that the Court rejected the analogy to Ilfeld because there was no loss deduction at issue.

11 . The Tax Court’s decision in CSI Hydrostatic Testers, Inc. v. Commissioner, 103 T.C. 398 (1994), is not to the contrary. In that case, the Court applied Woods when legislative history indicated clearly that Congress intended the statute in dispute to require including relevant income in the calculation of its subsidiaries’ earnings and profits, and thus no more specific statement was needed for the taxpayer to adjust its basis under § 1.1502-32. See id. at 405.

12 . Duquesne also posits that Treas. Reg. § 1.1502-80(a) precludes application of the Ilfeld doctrine because it requires separate treatment of parent and subsidiary in the absence of a consolidated return regulation to the contrary. Though the Tax Court once interpreted the regulation in this manner without reasoned analysis, see Gottesman & Co., Inc. v. Commissioner, 77 T.C. 1149, 1156 (1981), this interpretation finds no support in its text. See Treas. Reg. § 1.1502-80(a) (as amended March 2003) ("The Internal Revenue Code, or other law, shall be applicable to the group to the extent the regulations do not exclude its application”). We therefore reject it and simply look to other law, including the Ilfeld doctrine, to determine the proper treatment of stock losses for entity groups filing consolidated returns.

13 . In what was known as a “son of mirrors” transaction, a consolidated taxpayer would purchase a new subsidiary that had assets with built-in gains. See Leatherman, supra, at 846-47 & n.172. It would then trigger a gain by transferring those assets to another member of the group, which increased the parent’s basis in the new subsidiary’s stock under § 1.1502-32. Id. But because the transfer of the assets also reduced the value of the new subsidiary's stock, the parent could then sell its stock and claim a loss to offset the gain realized on the assets. Id. .

14 . Though § 1.1502-35T was not issued in final form until 2003, it applied retroactively to stock sales starting on March 7, 2002 (with certain exceptions not relevant here). See Treas. Reg. § 1.1502-35T(j).

15 . What our dissenting colleague characterizes as our deference to the IRS’s "position regarding the status or strength of judicial precedent," Dissent at 423, is no deference at all, but our independent assessment that possibly relevant regulations did not provide the required authorization under Ilfeld for Du-quesne to take a double deduction. Because we happen to agree with the IRS as to whether Ilfeld's requirements were satisfied in this circumstance does not mean we defer in any way to its reading of judicial precedent.

16 . .Duquesne also makes the argument that even if its interpretation of the consolidated return regulations proved incorrect, it is still entitled to claim a double deduction because its interpretation of those regulations was reasonable. Though the Tax Court allowed a reasonableness defense in Gottesman, this was because the IRS sought to impose a penalty tax that is strictly construed against the IRS. See 77 T.C. at 1156. Indeed, the Court repeated that its reasoning rested on the penalty tax at issue "for emphasis.” Id. Duquesne is unable to cite any case in which reasonableness was allowed as a defense for non-payment of ordinary corporate taxes such as those before us today (despite its insistence that Applied Research Assocs., Inc. v. Commissioner, 143 T.C. 310 (2014), somehow implies this).

Though Duquesne also argues that Gottes-man teaches courts not to fill gaps in the consolidated return regulations, Gottesman’s language to that effect is also off point. See 77 T.C. at 1158. It concerned not only a penalty tax, but also, as in Woods, a situation in which the IRS had failed to amend its regulations' despite having years to do so. See id.

17 . Subsection 6501 (k) reads in full:

Tentative carryback adjustment assessment period. — In a case where an amount has been applied, credited, or refunded under section 6411 (relating to tentative carryback and refund adjustments) by reason of a net operating loss carryback, a capital loss car-ryback, or a credit carryback (as defined in section 6511(d)(4)(C)) to a prior taxable year, the period described in subsection (a) of this section for assessing a deficiency for such prior taxable year shall be extended to include the period described in subsection (h) or Q), whichever is applicable; except that the amount which may be assessed solely by reason of this subsection shall not exceed the amount so applied, credited, or refunded under section 6411, reduced by any amount which may be assessed solely by reason of subsection (h) or (j), as the case may be.

26 U.S.C. § 6501 (k).