concurring in part and dissenting in part.
I join the portions of the majority’s opinion affirming Hoskins’s conviction, the district court’s ultimate finding regarding the amount of the tax loss, and the district court’s application of the U.S.S.G. § 2Tl.l(b)(l) enhancement. I respectfully dissent from the portion of the opinion in which the majority takes the unnecessary step in announcing a rule permitting defendants in future cases to offer deductions they did not actually claim in order to establish “a more accurate determination of the tax loss” under U.S.S.G. § 2Tl.l(a). *1100We are not called upon in this case to reach this issue and, as a consequence, I would not reach it. Further, if required to reach it, I would side with our own prior precedent and the vast majority of our sister circuits who have addressed the issue.
I agree with the majority’s conclusion that the district court did not err when it accepted the government’s evidence regarding tax loss, and that this determination is sufficient to uphold the calculation of Hoskins’s base offense level.1 The majority then discusses the hypothetical case in which a defendant somehow offers “convincing proof’ of the tax return she would have filed had she known that she would later be caught and prosecuted for tax evasion. See Op. at 1094-95. Because the answer to the more abstract question presented “makes no difference to the outcome of the case before us ... we need not and [should] not decide it.” Valley Forge Ins. Co. v. Health Care Mgmt. Partners, Ltd,., 616 F.3d 1086, 1095 (10th Cir.2010). “Judicial restraint, after all, usually means answering only the questions we must, not those we can.” Id. at 1094. However, because the majority chooses to discuss it, I write to express my disagreement with a rule that will, in my view, greatly and improperly complicate sentencing in tax cases. I would adhere to our prior statements in United States v. Spencer, 178 F.3d 1365 (10th Cir.1999), and reiterate that a defendant may not attempt to adjust the amount of the tax loss by proposing deductions that she did not actually claim at the time the fraudulent tax return was filed.
The majority views our discussion in Spencer, where we reject a defendant’s entitlement to retroactive deductions when computing § 2Tl.l(a) tax loss, as dictum. I do not think the discussion in Spencer can be so readily ignored. In Spencer, we noted that the Second Circuit had recently held that defendants could “employ ‘legitimate but unclaimed deductions’ in calculating tax loss,” but “[w]e questioned] this conclusion.” Id. at 1368 (quoting United States v. Martinez-Rios, 143 F.3d 662, 671 (2d Cir.1998)). We then specifically rejected the interpretation of the phrase “a more accurate determination of the tax loss” which is now adopted by the majority:
The sentencing guidelines simply authorize a court to avoid the presumptive tax rates if a “more accurate determination of the tax loss can be made.” We do not interpret [U.S.S.G. § 2Tl.l(c) Note (A)] as giving taxpayers a second opportunity to claim deductions after having been convicted of tax fraud. It must be remembered that, in tax loss calculations under the sentencing guidelines, we are not computing an individual’s tax liability as is done in a traditional audit. Rather, we are merely assessing the tax loss resulting from the manner in which the defendant chose to complete his income tax returns.
Id. (citation omitted). Although this statement was an alternative basis upon which *1101to affirm the district court’s ruling because there was no evidence to support the defendant’s proposed deductions, id. at 1369, the statement is a clear rejection of the rule the majority now advances.2 “Alternative rationales such as this, providing as they do further grounds for the Court’s disposition, ordinarily cannot be written off as dicta.” Surefoot LC v. Sure Foot Corp., 531 F.3d 1236, 1243 (10th Cir.2008).
The goal of a tax loss calculation is to assess the tax loss “resulting from the manner in which the defendant chose to complete his tax returns.” Spencer, 178 F.3d at 1368. The scope of a defendant’s tax evasion is determined at the point at which the return is filed, not after the defendant is charged and convicted. The majority’s statement that “the government is not supposed to reap windfall gains as a result of tax evasion,” Op. at 1095, has no bearing on the instant issue, where we are asked to determine tax loss based on the tax return the defendant actually filed. The tax loss embodied in the fraudulent return is not necessarily the amount that the government actually lost in revenue or the amount that the defendant could ultimately be ordered to pay, because “[t]he tax loss is not reduced by any payment of the tax subsequent to the commission of the offense.” U.S.S.G. § 2Tl.l(c)(5). Surely if § 2T1.1 tax loss cannot be reduced by the defendant’s subsequent payment of taxes, § 2T1.1 tax loss cannot be reduced by unclaimed deductions proffered in an unfiled return after conviction.
Five other circuits have also concluded that a defendant cannot reduce the U.S.S.G. § 2T1.1 tax loss with unclaimed deductions. See United States v. Yip, 592 F.3d 1035, 1041 (9th Cir.2010) (“We hold that § 2T1.1 does not entitle a defendant to reduce the tax loss charged to him by the amount of potentially legitimate, but unclaimed, deductions even if those deductions are related to the offense.”); United States v. Phelps, 478 F.3d 680, 682 (5th Cir.2007) (per curiam) (holding that the defendant could not reduce the tax loss by taking a social security tax deduction that he did not claim on the false return); United States v. Delfino, 510 F.3d 468, 473 (4th Cir.2007) (“The law simply does not require the district court to engage in [speculation as to what deductions would have been allowed], nor does it entitle the Delfinos to the benefit of deductions they might have claimed now that they stand convicted of tax evasion.”); United States v. Chavin, 316 F.3d 666, 678 (7th Cir.2002) (holding that the definition of tax loss “excludes consideration of unclaimed deduc*1102tions”); see also United States v. Clarke, 562 F.3d 1158, 1164 (11th Cir.2009) (holding that the defendant was not entitled to have the tax loss calculated based on a filing status other than the one he actually-used and that “[t]he district court did not err in computing the tax loss based on the fraudulent return Clarke actually filed, and not on the tax return Clarke could have filed but did not”).
The reasoning of these cases from our sister circuits is sound. The majority’s ruling essentially allows the defendant a “do over” by permitting the defendant, after conviction, to prepare a hypothetical, substitute return which minimizes the defendant’s hypothetical tax liability. The fact that the defendant might have done things differently had she known she would be caught does not alter what she actually did, which was file a return without the deductions now proposed.3 Thus, the unclaimed deductions were not part of “the manner in which the defendant chose to complete [her] tax returns.” Spencer, 178 F.3d at 1368; see also Chavin, 316 F.3d at 678 (“[T]he defendants’ intention is embodied in the tax return that was filed with the IRS.”). When affirming Hos-kins’s conviction for tax evasion in violation of 26 U.S.C. § 7201, the majority concludes that Hoskins’s signature on the false return was the affirmative act of tax evasion for which “she cannot now escape criminal liability.” Op. at 1091. However, when reviewing the sentence imposed for this criminal act, the majority shifts gears and permits an after-the-fact “do over,” by concluding we should consider for sentencing purposes a hypothetical tax return that did not serve as the basis for her criminal conviction. The majority’s new rule would allow a defendant to escape the full consequences of the return the defendant chose to file.
The majority’s decision is based in part on the idea that, at the time the false return was filed, the defendant legally owed a certain amount of taxes which included unclaimed deductions. This is a fiction; deductions do not reduce one’s tax liability unless they are actually claimed. Equally as troubling, the majority’s rule invites defendants to turn a sentencing hearing into a tax audit and the district court into a tax court tasked with determining whether the deductions proposed at sentencing would have been viable when the defendant’s return was actually filed. See Martinez-Rios, 143 F.3d at 670 (criticizing a regime requiring courts to consider unclaimed deductions because it would oblige “a sentencing judge ... to make a precise determination of tax liabilities, resolving issues normally determined in administrative proceedings of the Internal Revenue Service, sometimes subject to civil litigation”).
The majority also bases its ruling on the evolution of § 2Tl.l’s language. My reading of that evolution does not comport with the majority’s. The previous version of the guideline defined tax loss as “the greater of: (A) the total amount of tax that the taxpayer evaded or attempted to evade; and (B) the ‘tax loss’ defined in § 2T1.3.” In turn, § 2T1.3 defined tax loss as “28 percent of the amount by which the greater of gross income and taxable income was understated, plus 100 percent of the total amount of any false credits claimed against tax.” The gross income-based calculation was an alternative meth*1103od of calculating tax loss. The application notes to this previous version of § 2T1.1 stated: “The guideline refers to § 2T1.3 to provide an alternative minimum standard for the tax loss, which is based on a percentage of the dollar amounts of certain misstatements made in returns filed by the taxpayer. This alternative standard may be easier to determine, and should make irrelevant the issue of whether the taxpayer was entitled to offsetting adjustments that he failed to claim.”
In 1993, the Sentencing Commission changed the definition of tax loss to “the tax loss is the total amount of loss that was the object of the offense (i.e., the loss that would have resulted had the offense been successfully completed).” The Commission also deleted the application note discussed above. The note was deleted because it was no longer relevant, not in order to “delete[ the] rule explicitly foreclosing consideration of unclaimed offsetting adjustments,” Op. at 1096. See Yip, 592 F.3d at 1041. I would conclude that the change in language supports a determination that the tax loss that was the “object of the offense” is something other than the minimum amount of tax that the defendant possibly could have owed at the time. The old language, “[t]he total amount of tax that the taxpayer evaded or attempted to evade,” sounds more like actual government revenue loss than “the amount of loss that was the object of the offense.” See Chavin, 316 F.3d at 678 (finding it plausible that the application note was deleted “because the new tax-loss definition specifically excludes consideration of unclaimed deductions on its face by defining tax loss as ‘the object of the offense.’ ”). Further, I note that in the 1993 amendments the Commission also adopted a revised tax loss table “to provide increased deterrence for tax offenses.” U.S.S.G.App. C, Amendment 492. The majority’s interpretation cuts against the goal of greater deterrence for tax offenses because it will greatly reduce the sentencing range for defendants who could have taken deductions if they had filed truthful tax returns.
Finally, I am puzzled by footnote nine of the majority’s opinion, which “emphasized ... that § 2T1.1 does not permit a defendant to benefit from deductions unrelated to the offense at issue.” Op. at 1094 n. 9. I fail to see why it should matter whether the unclaimed deductions are related to the offense or not.4 In fact, it might make more sense to permit unrelated deductions precisely because they are unrelated to the offense and, thus, not part of the tax evasion scheme to be addressed at sentencing. Cf. Clark v. United States, 211 F.2d 100, 103 (8th Cir.1954) (“Some times the failure to claim deductions in a return may well be part of the taxpayer’s scheme to cover up his unreported income as a matter of not creating suspicion on the face of his return.”). Once a district court begins entertaining hypothetical unclaimed deductions, it must inevitably attempt to calculate the government’s actual revenue loss. Why stop at unclaimed deductions relating to the specific offense? If the goal is to determine what an honest, tax-minimizing taxpayer would have done to determine what tax was legally owed (which becomes the goal when the court entertains unclaimed deductions), I would think that courts would be compelled to consider all possible exemptions, deductions and tax credits.
For these reasons, I cannot join in the portion of the-majority’s opinion that permits defendants to proffer unclaimed de*1104duetions in order to reduce tax loss under U.S.S.G. § 2T1.1.
. The majority addresses the issue because the government argues that the district court ■ erred in considering Hoskins's unclaimed deductions. See Op. at 1092-93 n. 4. The fact that a party raises an alternative argument does not mean we are bound to address it. In this case, the government argued both that a defendant cannot point to unclaimed deductions and that "even if it were legally possible, under some circumstances, for a criminal defendant to take advantage of hypothetical, unclaimed deductions at sentencing, the district court in this case did not err — much less clearly err — when it evaluated the evidence and found that Hoskins's tax loss was more than $400,000.” Aple. Br. at 15; see also Aple. Br. at 20-23.
. The majority relies on the government’s concession at oral argument in Roy Hoskins’s case. See Op. at 1094 n. 8 (discussing the government’s concession that the language in Spencer was dicta). First, I note that a party’s interpretation of our prior rulings is not determinative. Cf. U.S. Nat’l Bank of Or. v. Indep. Ins. Agents of Am., Inc., 508 U.S. 439, 447, 113 S.Ct. 2173, 124 L.Ed.2d 402 (1993) (''[T]he Court of Appeals acted without any impropriety in refusing to accept what in effect was a stipulation on a question of law.’’); United States v. Charles, 576 F.3d 1060, 1066 (10th Cir.2009) ("[T]he government’s concession is not dispositive because a party's position in a case ... does not dictate the meaning of a federal law.” (internal quotation omitted)). Second, I think the majority misstates the government's position. The government in its brief repeatedly cites Spencer and treats our rejection of the very tax-loss argument Hoskins is now making, not as dictum, but as a holding of the case. See Aple. Br. at 8, 16 ("In United States v. Spencer, this Court expressly rejected the argument Hoskins now makes.”), 17 ("Hoskins attempts to escape Spencer's holding____”), 17-18 (“Spencer holds that 'the total amount that was the object of the offense' must be calculated based on what the defendant ... actually did, not what they could have done or might have done.”), 18 ("[Tjhis Court's holding in Spencer was not limited to its statement that there was no 'competent evidence' in the record that would allow a 'more accurate determination of the tax loss to be made.’ ”).
. Further, the majority’s approach would permit a defendant to claim deductions to which she is no longer entitled because the time period for amending her tax return has expired. The majority states that Hoskins offered hypothetical tax returns in this case because "it was too late to submit amended returns to the IRS.” Op. at 1088. It was too late because the Hoskinses did not choose to file timely amended or truthful returns.
. Further, if the offense is willfully filing a false tax return, I fail to see how some unclaimed deductions would be related to the offense and some deductions would not be. All the deductions relate to the tax return.