Muskat v. United States

SELYA, Circuit Judge. This case turns on the appropriate tax treatment of a contractual payment initially reported as ordinary income but later recharacterized as a capital gain. The Internal Revenue Service (IRS) denied a requested refund and, following a bench trial, the district court upheld that action. This appeal touts four claims of error, the most significant of which require us to elaborate upon the use and meaning of, *187 and then to apply, the “strong proof’ rule. After careful consideration of all four claims, we affirm the judgment below. I. BACKGROUND Irwin Muskat worked for years in a family business, Jac Pac Foods, Ltd., based in Manchester, New Hampshire. 1 The firm’s signature line of business was the processing and distribution of meat products to restaurant chains and other commercial entities. In 1968, Muskat assumed operating control of Jac Pac. Under his stewardship, Jac Pac’s annual revenues soared to nearly $130,000,000. Along the way, Muskat developed valuable relationships with customers, suppliers, and distributors. This litigation grew out of the acquisition of Jac Pac by Manchester Acquisition Corporation, a subsidiary of Corporate Brand Foods America, Inc. (collectively, CBFA). In 1997, George Gillett, CBFA’s chairman, contacted Muskat about a possible deal (at the time, Muskat was Jac Pac’s chief executive officer and the owner of 37% of its outstanding stock). Negotiations ensued. The negotiations touched in part on whether Muskat would receive remuneration over and beyond his share of the sale price for Jac Pac’s assets. Following lengthy deliberations, the parties agreed that Muskat would continue to run the business after CBFA acquired the assets, and that he would receive incremental payments under both an employment agreement and a noncompetition agreement. On March 31, 1998, representatives of CBFA and Jac Pac executed an asset purchase agreement, which provided that CBFA would buy all of Jac Pac’s assets (save for certain real estate) for $34,000,000 in cash and CBFA’s assumption of enumerated liabilities. The asset purchase agreement contained several conditions precedent, three of which pertained to Muskat’s execution of specific contracts, namely, a subscription agreement, an employment agreement, and a noncompetition agreement. Muskat signed the required agreements on May 7, 1998. The noncompetition agreement is of pivotal importance here. In it, CBFA pledged to pay Muskat $3,955,599 in return for a covenant not to compete over a thirteen-year period. The first installment — $1,000,000—was to be paid immediately, with other installments to be paid in varying amounts and at varying intervals over the next thirteen years. These payment obligations were to survive Muskat’s death. Muskat received the first installment in 1998. On his 1998 federal income tax return he listed the payment as ordinary income and paid income and self-employment taxes accordingly. In 2002, however, Muskat reversed his field; he filed an amended return for 1998, recharacterizing the payment as a capital gain and seeking a tax refund in the amount of $203,434. 2 The IRS denied the requested refund. Muskat then sued in New Hampshire’s federal district court, alleging that the payment was compensation for the trans *188 fer of his personal goodwill and, thus, was taxable as a capital gain. In a pretrial ruling, the district court declared that, in order to prevail, Muskat would have to show by “strong proof’ that he and CBFA intended the payment to be compensation for personal goodwill. Muskat v. United States (Muskat I), No. 06 Cv. 30, 2008 WL 138052, at *2 (D.N.H. Jan.10, 2008). Following a bench trial, the court determined that Muskat had failed to adduce the requisite strong proof. Muskat v. United States (Muskat II), No. 06 Cv. 30, 2008 WL 1733598, at *7-8 (D.N.H. Apr.2, 2008). At the same time, the court concluded that it lacked jurisdiction over Muskat’s claimed entitlement to a return of self-employment tax. Id. at *2-3. Consequently, the court entered judgment in favor of the government. This timely appeal ensued. II. ANALYSIS This case plays out against two background principles of tax law: The first principle holds that payments received in exchange for a covenant not to compete are usually taxable as ordinary income, whereas payments received for the sale of goodwill are usually taxable as capital gains. Compare, e.g., Baker v. Comm’r, 338 F.3d 789, 794 (7th Cir.2003) (holding payments under covenant not to compete taxable as ordinary income), with, e.g., Patterson v. Comm’r, 810 F.2d 562, 569 (6th Cir.1987) (holding amounts received for goodwill taxable at capital gain rates). The second principle holds that the tax rates applicable to ordinary income are normally higher than those applicable to capital gains. See 26 U.S.C. § 1 (tax tables). We proceed against the backdrop of these principles. In this venue, Muskat advances four assignments of error. These include: (i) the applicability of the “strong proof’ rule; (ii) the weight of the evidence as to whether the challenged payment constituted compensation for personal goodwill (and, thus, should have been taxed at capital gain rates); (iii) the exclusion of expert testimony; and (iv) the lower court’s refusal to consider the claim for a refund of self-employment tax. We discuss these issues sequentially. A. Strong Proof. It is beyond hope of contradiction that, in a tax refund suit, the complaining taxpayer bears the burden of proving the incorrectness of the challenged tax treatment. See Webb v. IRS, 15 F.3d 203, 205 (1st Cir.1994). Here, however, the parties disagree as to the quantum of proof required to satisfy that burden. Appellate courts review abstract legal questions de novo, and a level-of-proof question comes within that purview. See United States v. Goad, 44 F.3d 580, 585 (7th Cir.1995); N. Am. Rayon Corp. v. Comm’r, 12 F.3d 583, 586-87 (6th Cir.1993); see also Putnam Res. v. Pateman, 958 F.2d 448, 468-71 (1st Cir.1992). Accordingly, we review de novo the district court’s determination that Muskat had to adduce strong proof to prevail on his refund claim. The strong proof rule is peculiar to tax cases. 3 It applies when the parties to a transaction have executed a written instrument allocating sums of money for particular items, and one party thereafter seeks to alter the written allocation for tax purposes on the basis that the sums were, in reality, intended as compensation for some other item. The rule provides that, *189 in order to effect such an alteration, the proponent must adduce “strong proof’ that, at the time of execution of the instrument, the contracting parties actually intended the payments to compensate for something different. See Harvey Radio Labs., Inc. v. Comm’r, 470 F.2d 118, 119— 20 (1st Cir.1972); Leslie S. Ray Ins. Agency, Inc. v. United States, 463 F.2d 210, 212 (1st Cir.1972). Phrased another way, the party seeking to alter a written allocation must demonstrate an actual meeting of the minds with respect to some other allocation. 4 The heightened standard strikes the appropriate balance between predictability in taxation and the desirability of respecting the contracting parties’ real intentions. See Harvey Radio, 470 F.2d at 120. In applying it, evidence that a written allocation lacks independent economic reality, though likely relevant, is not sufficient to satisfy the strong proof test. 5 Id. at 119-20. Muskat vigorously contests the deployment of the strong proof rule in the circumstances of this case. He starts with the bald proposition that Harvey Radio is a relic of a bygone era and should not be perpetuated. We reject this assault on the continued vitality of Harvey Radio. “We have held, time and again, that in a multi-panel circuit, prior panel decisions are binding upon newly constituted panels in the absence of supervening authority sufficient to warrant disregard of established precedent.” United States v. Wogan, 938 F.2d 1446, 1449 (1st Cir.1991). Such “supervening authority” may take the form of a congressional enactment, a new Supreme Court opinion, or an en banc decision of our own. See United States v. Allen, 469 F.3d 11, 18 (1st Cir.2006); Williams v. Ashland Eng’g Co., 45 F.3d 588, 592 (1st Cir.1995). The second and third of these escape routes plainly do not apply here: Muskat has not cited to any overriding judicial decision that would call into question the durability of Harvey Radio. This leaves only the escape route paved by statutory enactments. In this vein, Muskat has argued that changes in the tax code have rendered lifeless the rationales undergirding Harvey Radio. But the strong proof rule is generic; it applies to the entire universe of written allocations, not just to those where changes in tax treatment have occurred. More importantly, the modifications highlighted by Muskat make no mention of the strong proof rule, nor do they necessarily imply that a different rule is desirable. Accordingly, Harvey Radio remains good law in this circuit. Muskat’s fallback position is that the strong proof rule, even if velivolant, does not apply in the circumstances at hand *190 because Muskat was not a party to the written allocation. The factual predicate on which this privity argument rests is faulty. We need not tarry. The record shows with conspicuous clarity that Muskat was a party to the allocation of funds to the noncompetition agreement. For one thing, that agreement bears Muskat’s signature in his personal capacity. For another thing, the testimony makes pellucid that, both individually and through his representatives, he negotiated the overall CBFA/Jac Pae transaction. That Muskat signed the asset purchase agreement on Jac Pac’s behalf was not a mere formality but, rather, an indicium of his deep involvement in the structuring of the deal. Finally, Muskat contends that the written allocation is ambiguous and that this ambiguity renders the strong proof rule inapposite. The premise behind this argument is sound: the strong proof rule does not apply to ambiguous contractual allocations. See, e.g., Kreider v. Comm’r, 762 F.2d 580, 586 (7th Cir.1985); Peterson Mach. Tool, Inc. v. Comm’r, 79 T.C. 72, 81-82, 1982 WL 11124 (1982). But Muskat’s attempt to rely upon this premise is an effort to force a square peg into a round hole. Whether a contract is ambiguous is a question of law. Torres Vargas v. Santiago Cummings, 149 F.3d 29, 33 (1st Cir.1998); Allen v. Adage, Inc., 967 F.2d 695, 698 (1st Cir.1992). “But a contract is not ambiguous merely because a party to it, often with a rearward glance colored by self-interest, disputes an interpretation that is logically compelled.” Blackie v. Maine, 75 F.3d 716, 721 (1st Cir.1996). Rather, a contract “is ambiguous only if the language is susceptible to more than one meaning and reasonable persons could differ as to which meaning was intended.” Uncle Henry’s Inc. v. Plant Consulting Co., 399 F.3d 33, 47 (1st Cir.2005). In this instance, the noncompetition agreement hardly could be clearer. It expressly states that the sums specified therein will be paid to Muskat in order to protect Jac Pac’s goodwill and in consideration of his serial promises not to participate in rival businesses, not to solicit employees to leave CBFA, and not to divert business opportunities from CBFA. The specified payments are clearly allocated to this covenant not to compete. In short, the noncompetition agreement unequivocally reads — as its title suggests — like a garden-variety agreement not to compete. See Black’s Law Dictionary 392 (8th ed.2008); see also LDDS Commc’ns, Inc. v. Automated Commc’ns, Inc., 35 F.3d 198, 200-01 (5th Cir.1994) (identifying agreement with similar restrictions as a non-competition agreement); Heritage Auto Ctr., Inc. v. Comm’r, 71 T.C.M. (CCH) 1839, 1996 WL 22405, at *5, *10 (1996) (treating agreement with similar provisions as covenant not to compete for tax purposes). In an endeavor to blunt the force of this reasoning, Muskat notes that the preamble to the noncompetition agreement recites that it was executed in consideration of the substantial benefits accruing to Muskat under the asset purchase agreement — an agreement that is itself ambiguous because it purports to allocate a $59,000,000 purchase price even though Jac Pac received only $45,000,000 from the sale. We fail to see how this arguable discrepancy, most likely explicable in terms of assumption of liabilities and other considerations, introduces an ambiguity into the allocation set forth in the noncompetition agreement. Whatever ambiguities might permeate the asset purchase agreement, there are none in the noncompetition agreement itself (to *191 which the asset purchase agreement unambiguously allocates $3,955,599). 6 To sum up, none of Muskat’s counterarguments is persuasive. Accordingly, we agree with the district court that Muskat had to adduce strong proof that the contracting parties intended, at the time of the transaction, that the challenged payment would be compensation for Muskat’s personal goodwill. It is to that issue that we now proceed. B. Weight of the Evidence. The court below found that Muskat had failed to adduce strong proof that the contracting parties intended the challenged payment to be compensation for personal goodwill. Muskat II, 2008 WL 1733598, at *8. This is essentially a factual finding, and we review it only for clear error. See Cumpiano v. Banco Santander P.R., 902 F.2d 148, 152 (1st Cir.1990) (“Findings concerning an actor’s intent fit neatly within the integument of the ‘clearly erroneous’ rule.”). Consequently, we will not disturb the district court’s determination “unless, on the whole of the record, we form a strong, unyielding belief that a mistake has been made.” Id. The phrase “strong proof’ is not self-elucidating. While the utilization of an enhanced level of proof is consistent with the spirit of our earlier cases, see, e.g., Leslie S. Ray, 463 F.2d at 212, the precise import of the strong proof rule is arguably best worked out on a case by case basis, see, e.g., United States v. Tex. Educ. Agency, 467 F.2d 848, 864 (5th Cir.1972). The tax court cases leave the matter pretty much up in the air; that court has construed the strong proof rule to require proof “beyond a mere preponderance of the evidence.” Major v. Comm’r, 76 T.C. 239, 247, 1981 WL 11362 (1981). Despite the advantages of a loose articulation, we think that a benchmark would be helpful. In our view, to constitute “strong proof’ a taxpayer’s evidence must have persuasive power closely resembling the “clear and convincing” evidence required to reform a written contract on the ground of mutual mistake. See, e.g., Ind. Ins. Co. v. Pana Cmty. Unit Sch. Dist. No. 8, 314 F.3d 895, 904 (7th Cir.2002); Berezin v. Regency Sav. Bank, 234 F.3d 68, 72 (1st Cir.2000); see also Nat’l Austl. Bank v. United States, 452 F.3d 1321, 1329 (Fed.Cir.2006) (collecting cases). This analogy seems compelling because, under the strong proof rule, a party seeking to vary a written allocation for tax purposes must show a meeting of the minds different from that professed in the written instrument—a showing that bears a family resemblance to the showing required for the reformation of a contract. See, e.g., Ind. Ins. Co., 314 F.3d at 904 (requiring party that seeks contract reformation to show “a meeting of the minds resulting in an actual agreement between the parties” different from that embodied in their written contract). The sources of “strong proof’ are case-specific. For that reason, an inquiring court, in determining whether there is strong proof that the parties to a transaction intended the allocation set forth in a written agreement to serve as a proxy for some other (more genuine) allocation, should closely examine the course of negotiations leading up to the agreement. See *192 Critz v. Comm’r, 54 T.C.M. (CCH) 947, 1987 WL 48884 (1987); Feller v. Comm’r, 45 T.C.M. (CCH) 902, 1983 WL 14102 (1983). In this ease, the trial testimony revealed no discussion of Muskat’s personal goodwill during the negotiations. By the same token, none of the transaction documents (including the early drafts of those documents) mentioned Muskat’s personal goodwill. Muskat had ample opportunity to introduce the concept of personal goodwill into the noncompetition agreement (which went through at least five iterations), but he did not do so. And although there is a reference to goodwill in the preamble to the noncompetition agreement, that reference is to an avowed purpose to protect Jac Pac’s goodwill. This is telling evidence. In our judgment, the absence of any reference to Muskat’s separate goodwill, combined with this express reference to Jac Pac’s goodwill, makes it extremely unlikely that the contracting parties intended the payments limned in the noncompetition agreement to serve as de facto compensation for Mus-kat’s personal goodwill. This intuition is reinforced by other pieces of evidence. CBFA’s president, Benjamin Warren, testified that he could not imagine that any goodwill other than Jac Pac’s was material to the transaction. The asset purchase agreement allocated almost $16,000,000 of the sale price to Jac Pac’s goodwill, lending credence to Warren’s testimony and making a focus on Muskat’s separate goodwill implausible. Seen in this light and with due deference to the district court’s prerogative to judge the credibility of the witnesses, the clear weight of the evidence supports the conclusion that the challenged payment was, as stated, compensation for the covenant not to compete, not compensation for Muskat’s personal goodwill. Muskat musters two arguments designed to undermine this conclusion. First, he asserts that the noncompetition agreement’s survivability provision is a dead giveaway that the payments called for by the agreement are for something other than refraining from competition (after all, Muskat hardly could compete subsequent to his demise). Second, he asserts that the terms of his employment and subscription agreements were so lucrative as to eliminate any realistic possibility that, at a relatively advanced age, he would cross swords with CBFA. The common thread that binds these arguments together is that they are in service of Muskat’s attempt to cast doubt upon the economic reality of CBFA’s need for a noncompetition agreement. That game may not be worth the candle; proof that a written allocation lacks economic reality does not, in and of itself, constitute strong proof that the contracting parties intended some other allocation. See Harvey Radio, 470 F.2d at 119-20. In any event, as we explain below, the noncompet-ition agreement possessed an adequate basis in economic reality. Muskat, by his own admission, had a considerable following in the trade (that is the core element of the “personal goodwill” that he touts). While the presence of a survivability provision is in some tension with the categorization of an agreement as a covenant not to compete, see, e.g., In re Johnson, 178 B.R. 216, 220 (9th Cir. BAP 1995); Brinson Co.-Midwest v. Comm’r, 71 T.C.M. (CCH) 1891, 1996 WL 27664, at *6 (1996); Ackerman v. Comm’r, 27 T.C.M. (CCH) 1342, 1968 WL 1339 (1968), it is not wholly antithetic to that taxonomy. After all, a person who has the wherewithal (knowledge, contacts, and the like) to compete effectively is in a strong position to drive a hard bargain in exchange for his agreement to eschew competition. A sur *193 vivability provision may be part of that hard bargain. Thus, courts frequently have classified agreements that contain survivability provisions as valid noncom-petition agreements for tax purposes. See, e.g., Thompson v. Comm’r, 73 T.C.M. (CCH) 3169, 1997 WL 344737, at *7-8 (1997); Buchner v. Comm’r, 60 T.C.M. (CCH) 429, 1990 WL 110212 (1990); Wager v. Comm’r, 52 T.C. 416, 419, 1969 WL 1736 (1969). We think that classification is apt here, notwithstanding the survivability provision contained in Muskat’s noncom-petition agreement. So too Muskat’s argument that it was unlikely that he would try to compete. It is true that competing would have had an up-front cost. After the sale, Muskat continued to work for CBFA in an executive capacity. He possessed a powerful financial incentive to remain with CBFA: he had invested $2,000,000 in the company through the subscription agreement, and his employment agreement paid him a base salary of $273,000 per year, together with a comprehensive benefits package and the prospect of sizable bonuses. On the other side of the balance, however, Muskat had been enormously successful prior to the sale. There is no indication that he was committed to retirement, infirm, or otherwise situated so as to render his promise not to compete of little value. Cf Welch v. Comm’r, 73 T.C.M. (CCH) 2256, 1997 WL 102431, at *6-8 (1997) (finding that covenant not to compete lacked economic reality when taxpayer was terminally ill at time of sale). Both Gillett and Warren testified that they valued Muskat’s relationships with customers, suppliers, and distributors, and there is every reason to suppose that a man with Muskat’s business acumen could have earned as much or more money by turning his back on CBFA and pursuing other (competitive) opportunities. Indeed, Gil-lett testified that the noncompetition agreement was meant to prevent that very possibility. The district court, as the arbiter of witness credibility, see Fed. Refin. Co. v. Klock, 352 F.3d 16, 29 (1st Cir.2003), was entitled to credit that testimony. The noncompetition agreement was, therefore, sufficiently grounded in economic reality. The short of it is that the weight of the evidence is completely consistent with the district court’s conclusion that CBFA sensibly protected its substantial investment in Jac Pac’s assets and goodwill by its contractual arrangement with Muskat. It follows inexorably that the court did not clearly err in holding that Muskat failed to adduce strong proof that the contracting parties intended the payments delineated in the noncompetition agreement to be compensation for Muskat’s personal goodwill. The payment received in 1998 was, therefore, taxable as ordinary income. See Baker, 338 F.3d at 794 (holding that payment under covenant not to compete is taxable as ordinary income). C. Expert Testimony. We review rulings admitting or excluding expert testimony for abuse of discretion. United States v. Sebaggala, 256 F.3d 59, 66 (1st Cir.2001). We will reverse such a ruling only if the trial court “committed a meaningful error in judgment.” Ruiz-Troche v. Pepsi Cola of P.R. Bottling Co., 161 F.3d 77, 83 (1st Cir.1998) (citation and internal quotation marks omitted). Such a bevue occurs when the court commits an “error of law, ... considers improper criteria, ignores criteria that deserve significant weight, or gauges only the appropriate criteria but makes a clear error of judgment in assaying them.” Rosario-Urdaz v. Rivera-Hernández, 350 F.3d 219, 221 (1st Cir.2003). At trial, Muskat sought to offer the opinion testimony of George O’Brien, a *194 certified public accountant, in order to establish that most of the goodwill (73%) acquired by CBFA was attributable to Muskat individually, not Jac Pac corporately. The court excluded the proffered testimony on relevancy grounds. Muskat appeals, maintaining that O’Brien’s opinion tended to show that he possessed a valuable asset (personal goodwill) that the contracting parties probably would have taken into account. The government’s rejoinder begins with the suggestion that Muskat neglected to preserve this objection. The record tells a different tale: when the district court questioned the relevance of the proffered testimony, Muskat’s counsel summarized what O’Brien would say and explained how the testimony would support Muskat’s position. No more was exigible to preserve the point for appeal. See Fed.R.Evid. 103(a)(2); see also Curreri v. Int’l Bhd. of Teamsters, 722 F.2d 6, 13 (1st Cir.1983). The government’s defense of the ruling on the merits is sturdier. The principal issue at trial was whether the contracting parties intended payments under the non-competition agreement to represent compensation for the transfer of personal goodwill. If O’Brien planned to testify that Muskat possessed personal goodwill separate from Jac Pac’s goodwill, his testimony arguably may have been relevant to that issue. But according to the proffer, O’Brien would not have testified to that effect; rather, he would have testified that a large slice of Jac Pac’s goodwill was attributable to Muskat. This is a significant distraction. All of Jac Pac’s goodwill, including any portion attributable to Mus-kat, was sold under the asset purchase agreement. Thus, O’Brien’s testimony would have shed no light on the meaning of the noncompetition agreement. That ends this aspect of the matter. It is black-letter law that “district courts enjoy wide latitude in determining the relevancy vel non of evidence.” Morales Feliciano v. Rullán, 378 F.3d 42, 58 (1st Cir.2004). Given the nature of the proffered testimony, the district court operated within the realm of its discretion in excluding it. D. Self-Employment Tax. Muskat’s final plaint concerns his separate claim for a refund of self-employment tax. He maintains, in the alternative, that the lower court erred in concluding that it lacked subject matter jurisdiction over this claim; that the court abused its discretion in denying his motion for leave to amend his complaint; and that, in all events, he should have prevailed on a theory of judicial estoppel. We review the district court’s determination that it lacked subject matter jurisdiction de novo. Dominion Energy Brayton Point, LLC v. Johnson, 443 F.3d 12, 16 (1st Cir.2006). The Internal Revenue Code recognizes the value of an orderly refund process, requiring the exhaustion of remedies available through administrative channels prior to opening the courthouse doors. Under that scheme, a district court has jurisdiction to adjudicate only those refund claims that have first been “duly filed” with the Secretary of the Treasury. 26 U.S.C. § 7422(a). Relevant regulations provide that a filed claim “must set forth in detail each ground upon which a ... refund is claimed and facts sufficient to apprise the Commissioner of the exact basis thereof.” 26 C.F.R. § 301.6402-2(b)(l). Taken together, these provisions bar “a taxpayer from presenting claims in a tax refund suit that ‘substantially vary1 the legal theories and factual bases set forth in the tax refund claim presented to the IRS.” Lockheed Martin *195 Corp. v. United States, 210 F.3d 1366, 1371 (Fed.Cir.2000); accord Charter Co. v. United States, 971 F.2d 1576, 1579 (11th Cir.1992). It follows that a claim or theory not explicitly or implicitly set forth in the taxpayer’s administrative refund application cannot be broached for the first time in a court in which a subsequent refund suit is brought. See Lockheed Martin, 210 F.3d at 1371. Here, the administrative refund claim filed on Muskat’s behalf stated in full: Taxpayers are amending their tax return to properly record the allocation between the sale of goodwill and a covenant not to compete. This change results in the reclassification of income erroneously reported as fully ordinary income to the correct allocation between ordinary income and capital gain. Fairly read, this statement indicates that the sole purpose of the refund claim is to change the allocation of the 1998 payment between goodwill and noncompetition, emphasizing the former and deemphasizing the latter, to the end of taxing at the (lower) capital gain rate monies previously taxed at the (higher) ordinary income rate. The IRS addressed that claim head-on, and Muskat’s ensuing judicial complaint neither mentioned an alternative claim for self-employment tax nor raised any other new issues. At trial, Muskat shifted gears. He sought to argue, in part, that he was entitled to a refund of the self-employment tax remitted with respect to the reported payment whether or not the payment constituted ordinary income. In support, he pointed to a line of cases holding that sums paid in consideration of covenants not to compete are not deemed to have been earned in the conduct of a trade or business and, thus, are not subject to self-employment tax. See, e.g., Milligan v. Comm’r, 38 F.3d 1094, 1098 n. 6 (9th Cir.1994); Barrett v. Comm’r, 58 T.C. 284, 289, 1972 WL 2450 (1972). This was an entirely new theory, neither mentioned in nor adumbrated by the administrative claim. The district court did not reach the question of whether payments made under a noncompetition agreement are subject to self-employment tax. Instead, the court noted that the administrative refund claim did not raise the self-employment tax issue and, therefore, the court lacked jurisdiction over it. See Muskat II, 2008 WL 1733598, at *2-3. We concur with the district court that this theory, voiced for the first time in the district court, worked a substantial variance from the administrative refund claim. Regardless of whether the IRS might have deduced from the general parameters of the refund claim that Muskat was eligible for a refund of self-employment tax even if the reported payment was attributed to the noncompetition agreement, the district court lacked jurisdiction. A taxpayer is the master of his refund claim, and it is not the IRS’s responsibility to make a case for the taxpayer that the taxpayer himself has opted not to make. See, e.g., IA 80 Group, Inc. v. United States, 347 F.3d 1067, 1075 n. 9 (8th Cir.2003); Charter Co., 971 F.2d at 1579. Because Muskat failed to put the IRS on notice during the administrative phase of the basic nature of his present theory, the district court lacked subject matter jurisdiction over that claim. See Lockheed Martin, 210 F.3d at 1371; Charter Co., 971 F.2d at 1580. That determination is dispositive of Muskat’s further contention that the trial court abused its discretion in refusing to allow the filing of an amended complaint. The law is settled that futility is a sufficient basis for denying leave to file an *196 amended complaint. See Foman v. Davis, 371 U.S. 178, 182, 83 S.Ct. 227, 9 L.Ed.2d 222 (1962); Universal Commc’n Sys., Inc. v. Lycos, Inc., 478 F.3d 413, 418 (1st Cir.2007). Since the district court lacked jurisdiction over Muskat’s claim for a refund of self-employment tax, Muskat’s proposed amendment would have been utterly futile. In a last-ditch effort, Muskat suggests that the IRS has conceded that no self-employment tax was owed on payments made under the noncompetition agreement in 2001 and 2002. 7 On that basis, he implores us that the government should be judicially estopped from contesting the impropriety of the 1998 tax in this proceeding. We reject his importunings. “As a general matter, the doctrine of judicial estoppel prevents a litigant from pressing a claim that is inconsistent with a position taken by that litigant either in a prior legal proceeding or in an earlier phase of the same legal proceeding.” InterGen N.V. v. Grina, 344 F.3d 134, 144 (1st Cir.2003). There are at least two preconditions to a successful claim of judicial estoppel. “First, the estopping position and the estopped position must be directly inconsistent, that is, mutually exclusive. Second, the responsible party must have succeeded in persuading a court to accept its prior position.” Alternative Sys. Concepts, Inc. v. Synopsys, Inc., 374 F.3d 23, 34 (1st Cir.2004) (internal citations omitted). Even if we assume, for argument’s sake, that (i) this judicial estoppel theory somehow eludes the jurisdictional bar and (ii) the IRS has conceded that payments under the noncompetition agreement are not subject to self-employment tax, Muskat’s judicial estoppel claim falters at the first step. The 2001-2002 position that Muskat attributes to the government (“no self-employment tax on payments received pursuant to noncompetition agreements”) is not inconsistent -with the basic position that the government urges in this litigation: that a taxpayer who has failed to exhaust his administrative remedies may not litigate a self-employment tax issue in a refund suit. Given this circumstance, the doctrine of judicial estoppel is not implicated. III. CONCLUSION We need go no further. For the reasons elucidated above, we uphold the judgment of the district court. Affirmed. Footnotes [1] Muskat and his wife, Margery, filed joint income tax returns for the relevant years. They jointly appear as plaintiffs and appellants in this case. For ease in exposition, we refer throughout to Muskat as if he were the sole party in interest. Nevertheless, our decision binds both taxpayers. [2] That amount comprised $176,652, which corresponded to the lower tax rate on capital gains, and $26,782, which corresponded to the elimination of self-employment tax. The parties now agree that any overpayment of self-employment tax would be in the amount of $21,479, rather than $26,782. [3] In the area of taxation, the strong proof rule has been thought to promote important values, such as administrative ease, certainty, predictability in taxation, and general notions of fairness. See Harvey Radio Labs., Inc. v. Comm’r, 470 F.2d 118, 120 (1st Cir.1972). [4] A few courts of appeals employ a more rigorous formulation of the strong proof rule. See, e.g., Comm’r v. Danielson, 378 F.2d 771, 775 (3d Cir.1967). That variation allows alteration of a written allocation only through "proof which in an action between the parties to the agreement would be admissible to alter that construction or to show its unenforce-ability because of mistake, undue influence, fraud, duress, etc." Id. The tax court generally applies the strong proof rule as we have articulated it unless an appeal would lie in a circuit that has adopted the alternative formulation. See, e.g., Pinson v. Comm’r, T.C.M. (RIA) 2000-208, 2000 WL 949390, at *11 (2000). [5] The strong proof rule applies only when a contracting party, or someone claiming by, through, or under a contracting party, attempts to vary a written allocation. When the IRS seeks to secure the reallocation of funds expressly earmarked for a given purpose, it may do so by showing that the original allocation does not comport with economic reality. See, e.g., Sullivan v. United States, 618 F.2d 1001, 1007 (3d Cir.1980); Harvey Radio, 470 F.2d at 120. [6] Muskat claims that the survivability provision renders the noncompetition agreement ambiguous. We do not agree. In all events, the question is not whether that provision— with which we deal infra—-is a typical feature of a noncompetition agreement but, rather, whether it lends an element of uncertainty to the bargain struck by the contracting parties. In this case, it does not. [7] The record reflects that, in a protest letter to the IRS dated October 15, 2004, Muskat flagged the self-employment tax issue with respect to tax years 1999 through 2002. The tax year at issue in the instant case is 1998.