Palmer Ranch Holdings Ltd v. Commissioner of Internal Revenue Service

812 F.3d 982

• Court: Court of Appeals for the Eleventh Circuit
• Decided: February 5, 2016
• Docket: 14-14167
• Status: Published

Opinion

GOLDBERG, Judge:

As recently as 2007, Southwestern Florida served as a proud home to the American, or bald, eagle. One bald eagle nest, SA-010, was located on the eastern side of an 82.19 acre parcel of land known as B-10, in Sarasota County. To the west of B-10 lies Sarasota Bay, where the eagles would fly to feed. To allow the eagles to reach their feeding grounds safely, B-10 sported a nest-to-coast flyway in the form of a “wildlife corridor.” The wildlife corridor also provided a habitat to small urban animals of considerably less patriotic interest.

Freedom isn’t free. . Concern over the eagle nest, wildlife corridor, and wetlands on B-10 thwarted plans by the parcel’s owner, Palmer Ranch, Inc. (“Palmer Ranch”) to sell B-10 and the adjacent B-9 for residential development. Ever resourceful, Palmer Ranch turned around and donated a conservation easement on B-10 to Sarasota County in 2006, a strategy that allowed the corporation to deduct the easement’s value from that year’s tax returns. But Palmer Ranch’s backup plan fell prey to a sharp-eyed IRS, who disallowed the deduction on grounds that Palmer Ranch had overvalued B-10 in calculating the associated easement’s worth. Palmer Ranch had valued B-10 at $25,200,000 on the assumption that B-10’s highest and best use was residential development, with development of 360 dwelling units being reasonably probable. The IRS did not share Palmer Ranch’s optimism.

Palmer Ranch challenged the IRS’s decision in tax court. There, Palmer Ranch insisted that B-10’s highest and best use was residential development under a Moderate Density Residential (“MDR”) zoning designation, which would allow between two and five units per acre, or 164 to 410 units total. Based on this highest and best use, Palmer Ranch stuck to its initial $25,200,000 valuation. The IRS countered with a maximum highest and best use of 100 units and a corresponding valuation of $7,750,000.

The tax court held in favor of Palmer Ranch on B-10’s highest and best use, but nonetheless revalued B-10 at $21,005,278, instead of $25,200,000. The parties cross-appealed. The IRS argues that the tax court ballooned B-10’s highest and best use, and that the tax court’s valuation was therefore too rosy. Palmer Ranch likes the tax court’s highest-and-best-use analysis, but is not on board with the revaluation. For the reasons that follow, we affirm the tax court’s determination of B-10’s highest and best use, but reverse the tax court’s $21,005,278 valuation. We remand for further proceedings consistent with this opinion.

BACKGROUND

The reader might wonder, Why is the value of B-10 itself, or B-10’s highest and best use, at issue? After all, Palmer Ranch’s deduction was for the value of a conservation easement on B-10-not for the value of the underlying parcel-and it was that deduction that the IRS disallowed and the parties litigated in tax court.

The answer lies in the method Palmer Ranch used to determine the conservation easement’s value: the “before and after” method. Under 26 C.F.R. § 1.170A-14(h)(3)(i) (2015),

[T]he fair market value of a perpetual conservation restriction [in this case, a conservation easement] is equal to the difference between the fair market value of the property it encumbers before the granting of the restriction [the “before” value] and the fair market value of the encumbered property after the granting of the restriction [the “after” value].

So the reason that the value of B-10 itself (i.e. B-10’s before value) is at issue is because it determines the value of the conservation easement. And B-10’s highest and best use is in turn at issue because B-10’s before value is informed by the parcel’s highest and best use. 26 C.F.R. § 1.170A — 14(h)(3) (ii) mandates that

the [before value] take into account not only the current use of the property but also an objective assessment of how immediate or remote the likelihood is that the property, absent the restriction, would in fact be developed, as well as any effect- from zoning,, conservation, or historic preservation laws that already restrict the property’s potential highest and best use.

Or, in the words of tax court case law, B-10’s before value must take into account the parcel’s highest and best “reasonable and probable use that supports the highest present value.” Symington v. Comm’r, 87 T.C. 892, 897, 1986 WL 22044 (1986). The “focus [is] on ‘the highest and most profitable use for which the property is adaptable and needed or likely to be needed in the reasonably near future.’ ” Id. (quoting Olson v. United States, 292 U.S. 246, 255, 54 S.Ct. 704, 708-09, 78 L.Ed. 1286 (1934)). Hence the significance of B-10’s highest and best'use in resolving this case.

Another concept critical to disposing of this case is the “comparable sales” method for valuing land after the highest and best use has been determined. As 26 C.F.R. § 1.170A — 14 (h) (3)(ii) makes clear, simply knowing a parcel of land’s highest and best use does not automatically result in a value for the parcel: There remains the task of arriving at a figure that “take[s] into account” the parcel’s highest and best use. One way to get there is through the comparable-sales method, which entails considering sales of similarly situated parcels. See Wolfsen Land & Cattle Co. v. Comm’r, 72 T.C. 1, 19 (1979) (“The ‘comparable sales’ method functions by: (1) Locating [parcels] as physically similar (comparable) as possible to the subject [parcel] which (2) have been sold on the open market in noneollusive, nonforced sales for cash or cash equivalent, within (3) a reasonable time of the date for which a value of the subject property is desired.”). The relevance of the comparable sales approach will become apparent as we go along, but suffice to say for now that the approach bears on the propriety of the tax court’s determination that B-10 was worth $21,005,278, not $25,200,000.

Having explained the concepts of before value, highest and best use, and the comparable-sales approach, we will now turn to the factual exposition necessary to frame our discussion of the issues. We begin by recounting an event that informs B-10’s highest and best use: Palmer Ranch’s aforementioned attempt to sell B-10.

A. The Failed Sale

In 2003, Palmer Ranch agreed to sell 86.9 acres within B-10 and an adjacent parcel, B-9, to Pulte Homes (“Pulte”), a national home developer, for $8.25 million. Pulte aimed to develop 240 dwelling units across the two parcels,’with 84 to be located on B-10. But in order for Pulte to develop this number of units, both B-9 and B-10 would have to be rezoned: B-10 was zoned as Residential Estate-1, which capped development to one unit per two acres, for 41 total units. Pulte would therefore have to submit a rezoning application to the Sarasota Board of County Commissioners (“BOCC”), which would then have to approve the' application under the Sarasota County comprehensive plan. The comprehensive plan restricted B-10’s top future zoning to MDR.

Another potential hurdle to Pulte’s development was the fact that both B-9 and B-10 were part of the Palmer Ranch Development of Regional Impact (“DRI”). So, besides submitting a rezoning application to the BOCC, Pulte would also have to submit a separate development application. The BOCC would have to approve the development application under the DRI’s master development order, which protected the bald eagle nest and created the wildlife corridor.

In a word, at the time that Palmer Ranch contracted with Pulte, the odds that Pulte could successfully develop B-9 and B-10 were at least to some degree uncer tain. To hedge for this uncertainty, Palmer Ranch and Pulte agreed that their sale would be contingent on the BOCC approving development and rezoning applications.

The BOCC did not oblige, denying Pulte’s development application by a vote of three to two without addressing the rezoning application. Rather than giving up, though, Pulte and Palmer Ranch went back to the drawing board. They redrafted the sales contract to exclude any part of B-10. Pulte would purchase land on B-9 only, albeit at the same $8.25 million price. With the sales contract drawn up, Pulte submitted new development and rezoning applications to the BOCC, this time seeking to develop 184 dwelling units on B-9 only. Palmer Ranch and Pulte eagerly awaited the BOCC’s decision.

Unrelenting, the BOCC denied Pulte’s development and rezoning applications three to two. With the second denial, the BOCC issued an ordinance offering guidance on the changes that “would be necessary in order to allow for a favorable consideration of the proposal.” To be approved, an amended development application would have to “set[] forth development plans which include[] both Parcel B-9 and B-10 (totaling 86.35 acres+/-) ...; this application should include Parcel B-10 and should endeavor to keep Parcel B-10 [intact] as it relates to the Eagle Preservation Area, the wetlands, and the wildlife corridor.” Notwithstanding the BOCC’s hint, Pulte declined to submit a third set of applications, instead canceling the sales contract with- Palmer Ranch.

B. Palmer Ranch’s Altruistic Response

Once the sales contract with Pulte fell through, Palmer Ranch donated a conservation easement on B-10 to Sarasota County. 1 This response allowed Palmer Ranch to deduct the value of the easement from its 2006 tax returns. See 26 U.S.C. § 170(f)(3)(B)(iii), (h) (2012). To establish the value of the conservation easement, Palmer Ranch commissioned a land use analysis from WilsonMiller, a land use and planning firm, and an appraisal from Chad Durrance, a licensed appraiser. In the land use analysis, WilsonMiller considered whether and to what extent the BOCC would approve future development on B-10 in guiding light of the BOCC’s recent ordinance. According to WilsonMiller, although the ordinance required B-10 to remain intact as necessary to protect the eagle nest, wetlands, and wildlife corridor, the ordinance did not indicate that the BOCC had denied Pulte’s second development application for reasons “related to the zoning district that was being requested at the time, [ ]or the number of units that were [then] being proposed [184].” WilsonMiller concluded that a development of 360 units on B-10, as outlined in its analysis, “would be found to be consistent with ... Sarasota County’s Comprehensive Plan, the Sarasota County Zoning Ordinance and Land Development Regulations, and the Palmer Ranch Master Development Order.”

With WilsonMiller’s analysis in hand, Durrance appraised the value of the conservation easement on B-10 using the before and after method. Durrance concluded that the highest and best use of B-10 “before” the conservation easement was “residential development.” To support this conclusion, Durrance relied on Wilson-Miller’s land use analysis, which Durrance paraphrased as indicating that “it is reasonably probable that a total of 360 dwelling units could be developed on [B-10] with potential for even more” and that “[t]his type of development could be accommodated ... while still protecting and preserving the Eagle’s Nest and any environmentally sensitive areas.” Durrance then used the comparable-sales method to value B-10 in light of its highest and best use. Durrance identified seven sales of comparable properties whose per-acre prices ranged from $237,200 to $510,600. On the basis of these comparables, Durrance concluded that the per acre price for B-10 if put to the parcel’s highest and best use would be $307,000, or $25,200,000 for the whole shebang. Durrance explained that one reason his price fell on the lower end of the comparable-sales range was a “softening of the residential market in 2006,” when Palmer Ranch donated the conservation easement. “For instance, the executed contract and agreed upon sales price for Comparable 1 at over $110,000 per unit [and $510,600 per acre] occurred in September 2005, which is estimated to be near the top of the market.” 2 Durrance further concluded that B-10’s value “after” the easement was $1,260,000, meaning that the value of the easement itself was $23,940,000.

Palmer Ranch deducted Durrance’s $23,940,000 easement value from its 2006 tax returns. The IRS disallowed a portion of Palmer Ranch’s deduction. Palmer Ranch sued the IRS in tax court.

C. The Experts’ Take in Tax Court

In tax court, both Palmer Ranch and the IRS relied on experts to establish the before value of B-10. For its part, Palmer. Ranch looked to James Paulmann, who had been a senior vice president at Wilson-Miller when the firm completed its land use analysis for B-10, and Durrance. Paulmann prepared an expert report in advance of trial, in which he concluded that “it is reasonably probable that Palmer Ranch Parcel B-10 would be approved for development of 360 multifamily units.” Paulmann also explained why “the [BOCC’s] 2004 denial of the Pulte rezoning application for Parcel B-9 does not change [this] conclusion”: “[T]he issues and circumstances present in 2004 related only to the specific development proposed by Pulte for Parcel B-9” and do “not relate to whether, in 2006, B-10 could have been rezoned in a manner consistent with [Wil-sonMiller’s 360] multifamily unit development plan.” When asked at trial why the BOCC’s 2004 ordinance mentioned B-10 at all, Paulmann said “[it] didn’t make sense to me why [the BOCC] would put that in there.”

Durrance also prepared additional documentation in advance of trial, for him in the form of a second appraisal. In the new appraisal, Durrance stuck by his original before valuation of $307,000 per acre, $25,200,000 total. Citing Paulmann’s report, Durrance explained that “residential development of B-10 at a density ranging from a minimum of 164 units (2 units per acre) to a maximum of 410 units (5 units per acre)” was reasonably probable. Therefore, the highest and best use of B-10 was “residential development at a density of between two and five units per acre,” or MDR density. This version of B-10’s highest and best use constituted an update to what Durrance had offered in the first appraisal — that B-10’s highest and best use was simply “residential development.”

The second appraisal also gave Durranee an opportunity to frame his before valuation as consistent with an appraisal ■ the government had obtained in separate takings litigation. In' Rogers v. United States, 90 Fed.Cl. 418 (2009), the government hired appraiser John Underwood to value undeveloped land in the Palmer Ranch DRI, including B~9 and B-10, as of April 2004. Underwood testified that the land’s highest and best use was residential development at MDR density. According to Durrance, Underwood “concluded undeveloped residential land in Palmer Ranch was worth $200,000 per acre in April 2004.” 3

Durrance explained that the 2004 valuation of $200,000 was consistent with his 2006 $307,000 valuation because Underwood had also proposed a monthly appreciation rate of 1.5%, which covered the difference. In Durrance’s words, “[w]hen the $200,000 per acre value of Palmer Ranch land is adjusted for the increase in value of Palmer Ranch land occurring between April 2004 to December 2006 (using [Underwood’s 1.5% per month] rate ...) the value of Palmer Ranch land was $299,000 per acre.” Durrance continued,

[t]he 2.7% monthly appreciation rate Underwood uses for the 2004 — 2005 period is the more appropriate rate by which to adjust the increase in the value of Palmer Ranch during the 2004 — 2006 period. But, whether one chooses to adjust the increase in value at 2.7% per month over a 21-month period, or -adjust the value using an average 1.5% per month for 33 months (April 2004 — December 2006), the result is the same — an increase in the value of Palmer Ranch land between April 2004 and December 2006 of essentially 50%.

Therefore, Durrance concluded, the 2004 valuation,of $200,000 was consistent with his $307,000 valuation.

At trial, Durrance was asked whether he had made any “specific adjustments” to account for “the differences in ... the state of the market between the properties that closed in 2005 and B-10 in December of '06.” Durrance specified that he had not made any “quantitative adjustments,” explaining that there was no “notable nosedive” in property values in December 2006; instead, there was simply a “a pause in the market ... a softening ... of activity.” Durrance was also asked about the development calendar for Comparable 1, the comparable sale closest in time to Palmer Ranch’s donation of B-10. 4 Durrance acknowledged that the purchaser put development plans “on hold” due to unfavorable economic conditions, and that the land swapped in the sale lay vacant until “the last half of 2012 into 2013.” Durrance noted, however, that it was not clear exactly how much of the delay could be attributed to economic conditions, because the usual “site development and engineering process” always cost some amount of time. Durrance also conceded that, once development started, the purchaser “lowered the density” below the 303 units originally planned.

Finally, Durrance was asked what he thought of B-10’s rezoning history. Dur-rance stated that he gave “no weight” to the history because “B-10 has never been up for rezoning as it is described in this appraisal.” Durrance also stated that it did not matter that the BOCC had denied the rezoning of B-9 because that denial involved “[d]ifferent properties, different time periods.”

The IRS rebutted Palmer Ranch’s before valuation with a planning and engineering report prepared by Richard Doyle and an appraisal prepared by Bradley Page. According to Doyle, “Based on the rezoning history of Parcel B-10, it is reasonably probable to expect 72 to 84 units to be approved for development.” Doyle lifted the 84 unit cap from Pulte’s first development application, in which Pulte had proposed to build that number of units on B-10. Page, relying on Doyle’s projection, deemed the highest and best use of B-10 to be “residential development,” with “the most likely and reasonably probable development [being] 72 to 100 units.” Be cause a 100-unit development would fall well below the MDR minimum of 164 units (two units per each of B-10’s 82 acres), Page necessarily implied that MDR-level development was beyond B-10’s highest and best use. Based on his 100-unit highest and best use, Page concluded that B-10’s before value was $7,750,000, or approximately $94,000 per acre. Page reached this amount despite not only using the same valuation method as Durrance (comparable sales), but also working from four of the same benchmark sales — including “the only true comparable” sale of 78.34 acres for $40,000,000, $510,600 per acre, in January 2006.

D. The Tax Court’s Opinion

The tax court issued a memorandum opinion in which it adopted Palmer Ranch’s proposed highest and best use for B-10 — residential development at MDR density — but nonetheless revalued B-10 at $21,005,278, instead of $25,200,000. In siding with Palmer Ranch on the question of highest and best use, the tax court rejected the IRS’s view of B-10’s rezoning history. The IRS pointed to the BOCC’s 2004 ordinance in particular, maintaining that the ordinance was “a legally binding document that. show[ed] a historical rezoning and development denial for parcel B-10.” The implication, said the IRS, was that the BOCC would deny any future MDR-density development.

The tax court disagreed with the IRS’s reading of the ordinance, in part on grounds that the IRS had relied “on reams of hearsay evidence found in the BOCC meeting minutes.” Rather than resorting to this hearsay, the tax court instead looked at the BOCC ordinance in context. The tax court began by pointing out that the BOCC issued the ordinance in 2004, two years prior to the valuation date, and that the ordinance “applied to B-9 only” (because the BOCC had issued it after denying Pulte’s second, revised development application, which omitted - B-10). The only BOCC decision that applied to B-10 was the BOCC’s June 2004 denial of Pulte’s first development application, an application that included B-10 but “not ... the eagle nest zone.” Taking the June 2004 denial and the ordinance together, the tax court figured that the BOCC “was concerned that the [June 2004] development application did not consider the nearby eagle nest zone, wetlands, and wildlife corridor.” This concern was exacerbated, not allayed, when Pulte dropped B-10 from the second development application altogether. That was why the ordinance instructed that a new application should “include Parcel B-10 and should endeavor to keep Parcel B-10 [intact] as it relates to the Eagle Preservation Area, the wetlands, and the wildlife corridor.” Far from indicating that the BOCC would deny MDR-density development on B-10, the ordinance instead mandated that B-10 be included in development applications, albeit in such a way as to shield the eagles and the environment. Because WilsonMiller’s 360-unit (MDR-density) development plan “included all of parcel B-10 (including the eagle nest zone)” and addressed the BOCC’s environmental concerns, it was reasonably probable for the BOCC to approve that plan.

The tax court also pointed out two other reasons that the BOCC would, in reasonable probability, approve MDR-level development on B-10. First, the BOCC decided both the 2004 denial and the ordinance by a close three-two vote, “suggesting] that the BOCC’s decision could have changed over time.” Second, the BOCC approved development of the Hamptons, which “was built in a wildlife corridor on a tract of land adjoining parcel B-10, near the eagle nest.” On the basis of all, this, the tax court concluded it was reasonably probable for the BOCC to approve MDR-density development on B-10, rendering such development B-10’s highest and best use.

Partly because the tax court held that B-10’s highest and best use was MDR-level development, the tax court deemed Durrance’s MDR-predicated appraisal ($25,200,000) to be “more accurate” than Page’s lower-use figure ($7,750,000). The tax court “also f[ound] compelling that both [Durrance and Page] used four of the same comparables, yet only Mr. Dur-rance’s appraisal is within the selected comparables’ price-per-acre range.” But after finding Durrance’s appraisal more accurate, the tax court nonetheless held that the appraisal was “not without fault” and warranted reduction. In the tax court’s own words,

The judge in [the prior takings litigation] determined that undeveloped land in the DRI (including parcel B-10) was worth $200,000 per acre as of April 2004. Using that price as a starting point [in his second appraisal], Mr. Durrance states that a market condition (or time) adjustment must be made to account for the trend of increasing land values. He estimates that the value of parcel B-10 was about $300,000 per acre in December 2006. As support for his estimate, Mr. Durrance provides alternative calculations: (1) apply a 2.7% monthly appreciation rate to $200,000 for 21 months (April 2004 through December 2005) or (2) apply a 1.5% monthly appreciation rate to $200,000 for 33 months (April 2004 through December 2006).

Mr. Durrance borrowed these alternative time adjustment calculations from the Government’s appraiser (John Underwood) in the Rogers case. However, Mr. Durrance’s use of these calculations is incomplete. Mr. Underwood states in his Rogers appraisal that there was a 2.7% per month increase in values from 2004 to 2005. However, he said appreciation rates for raw land lag behind those for developed land, so he used a 1.5% per month adjustment for Palmer Ranch’s -undeveloped land. Parcel B-10 is raw and, so we believe the 1.5% per month time adjustment is appropriate.

Mr. Durrance applied the 1.5% appreciation factor over 33 months, ending in December 2006. Citing numerous newspaper articles reporting a 2006 third-quarter decline in housing prices, respondent argues the appreciation rate is too high. On cross-examination, Mr. Durrance stated that the real estate market peaked in late 2005. He stated there was a softening in demand, but no notable nosedive or decrease in value in 2006. According to his own admission, Mr. Durrance should not have applied a steady 1.5% appreciation factor to parcel B-10 for the months after the market hit its peak in 2005. Even if property values did not decline in 2006, we think they likely stagnated. Accordingly, we will reduce petitioner’s appraisal by the amount of 2006 appreciation Mr. Dur-rance calculated. This results in a “before value” of $21,005,278.

On the basis of this “before” value, and after finding an “after” value of $1,050,264, the tax court held that the fair market value of the conservation easement was $19,955,014.

The parties appeal. According to the IRS, the tax court overshot B-10’s highest and best use. Palmer Ranch does not think so, but quibbles with the tax court’s revaluation of B-10 at $21,005,278.

STANDARD OF REVIEW

We review the tax court’s legal conclusions de novo and its findings of fact for clear error. Long v. Comm’r, 772 F.3d 670, 675 (11th Cir.2014). “[W]hether the Tax Court used the correct standard to determine fair market value is a legal issue .... A determination of fair market value is a mixed question of fact and law: the factual premises are subject to a clearly erroneous standard while the legal conclusions are subject to de novo review.” Estate of Jelke v. Comm’r, 507 F.3d 1317, 1321 (11th Cir.2007).

DISCUSSION

A. The IRS is not Estopped from Appealing the Tax Court’s Highest- and-Best-Use Determination.

At the outset, we must address a motion to dismiss filed by Palmer Ranch in which the corporation argues that the IRS is barred from pursuing its half of this appeal. Palmer Ranch offers two theories in this refrain. First, Palmer Ranch invokes the doctrine of judicial estoppel, which prevents a party from asserting inconsistent claims across different legal proceedings. Burnes v. Pemco Aeroplex, Inc., 291 F.3d 1282, 1285-86 (11th Cir.2002). Palmer Ranch argues that the IRS cannot now claim á highest and best use below MDR-level development because in Rogers IRS expert John Underwood testified that B-10’s highest and best use was residential development at MDR density. Palmer Ranch also draws our attention to the IRS’s behavior after instituting its appeal. Evidently, in June 2015, the IRS sought to levy the personal assets of Hugh and Liza Culverhouse (Hugh Culverhouse having an interest in Palmer Ranch). According to Palmer Ranch, by doing so the IRS “embraced the [t]ax [c]ourt’s decision.” Palmer Ranch says the IRS’s present appeal is inconsistent with this embrace.

Second, Palmer Ranch conjures the acceptance of benefits doctrine, a kind of equitable estoppel. 5 In the Eleventh Circuit, “[w]here one in whose favor a judgment is rendered accepts the benefits, he is estopped from questioning the validity[ ] of the judgment in any subsequent litigation.” Zaklama v. Mount Sinai Med. Ctr., 906 F.2d 645, 650 (11th Cir.1990) (citation omitted). Palmer Ranch maintains that, by seeking to levy Hugh and Liza Culver-house’s assets, the IRS accepted the benefits of the tax court’s judgment.

Both of Palmer Ranch’s estoppel arguments fail. The hallmark of judicial estoppel is inconsistency: “Under this doctrine, a party is precluded from ‘asserting a claim in a legal proceeding that is inconsistent with a claim taken by that party in a previous proceeding.’ ” Bumes, 291 F.3d at 1285 (citation omitted). The Supreme Court has enumerated several nonexhaustive factors courts may consider in deciding whether to apply judicial estoppel:

(1) whether the present position is “clearly inconsistent” with the earlier position; (2) whether the party succeeded in persuading a tribunal to accept the earlier position, so that judicial acceptance of the inconsistent position in a later proceeding creates the perception that either court was misled; and (3) whether the party advancing the inconsistent position would derive an unfair advantage on the opposing party.

Id. (citing New Hampshire v. Maine, 532 U.S. 742, 750-51, 121 S.Ct. 1808, 1815, 149 L.Ed.2d 968 (2001)). The Eleventh Circuit has its own formulation of the test. To prevail on a judicial estoppel theory, the proponent must show (1) “that the allegedly inconsistent positions were made under oath in a prior proceeding [and (2) that] such inconsistencies ... have been calcu lated to make a mockery of the judicial system.” Id. (citation omitted).

There is at least one rudimentary problem with Palmer Ranch’s judicial estoppel argument: The IRS’s positions have not been inconsistent. In Rogers, the IRS’s appraiser John Underwood testified that B-9 and B-10’s highest and best use as of April 2004 was residential development at MDR density. In this case, the IRS argues that B-10 could not achieve that same highest and best use as of December 2006. The IRS bases its argument on circumstances unique to 2006, including B-10’s rezoning history and the development outlook going forward at that time. The IRS’s more optimistic 2004 appraisal therefore is not inconsistent with its position in this appeal.

The same goes for the IRS’s pursuit of a levy against Hugh and Liza Culverhouse’s personal assets. That route is consistent with the IRS’s position on appeal because, through both its pursuit of the levy and this appeal, the IRS maintains that Palmer Ranch owes taxes. Although the IRS advocates on appeal for a greater tax deficiency than the one the tax court found, the agency never undercuts its entitlement to at least the tax court’s figure. Besides, 26 U.S.C. § 7485 expressly authorizes the IRS to seek a levy simultaneous to an appeal. See 26 U.S.C. § 7485 (allowing such a course when the taxpayer appeals but fails to post an appeal bond). There is no inconsistency in the IRS following the course laid out in statute, and therefore no cause for judicial estoppel. 6

Palmer Ranch’s acceptance of benefits argument fails as well. In the Eleventh Circuit, a party cannot be deemed to have accepted the benefits of a judgment unless his counterparty has offered a benefit for him to accept — generally in the form of payment in the judgment amount. 7 Even payment does not always suffice: “[W]here a judgment is appealed on the ground that the damages awarded are inadequate, acceptance of payment of the amount of the unsatisfactory judgment does not, standing alone, amount to an accord and satisfaction of the entire claim.” United States v. Hougham, 364 U.S. 310, 312, 81 S.Ct. 13, 16, 5 L.Ed.2d 8 (1960). When a party has accepted payment on the judgment but still insists on his right to appeal, we look to “the parties’ objective manifestations of intent” to determine whether the appeal can go forward. Alvarez Perez v. Sanford-Orlando Kennel Club, Inc., 518 F.3d 1302, 1307 (11th Cir.2008).

Palmer Ranch flunks at the very first step. To date, Palmer Ranch has paid not one deficiency-devoted dime to the IRS, nor tendered any other benefit, so there can be no acceptance of benefits by the IRS. 8 There is therefore no basis for us to hold that the IRS is estopped from appealing the tax court’s judgment. Accordingly, Palmer Ranch’s motion to dismiss is denied. We now proceed to the merits.

B. The Tax Court Did Not Err in Determining B-10’s Highest and Best Use.

The IRS claims that the tax court erred in determining that B-10’s highest and best use was MDR-level development. The IRS argues that (1) contrary to the tax court’s reasoning, B-10’s rezoning history shows that the BOCC would not, in reasonable probability, approve a 360-unit, MDR-density development; (2) the tax court “failed to address whether [a] 360-unit development was needed or likely to be needed shortly after December 2006”; and (3) the tax court “failed to consider the effect of the zoning history on hypothetical willing buyers of B-10.” For the reasons that follow, we uphold the tax court’s determination of B-10’s highest and best use.

1. The Tax Court Did Not Clearly Err in its Reading of B-10’s Rezoning History.

We agree with the tax court that B-10’s rezoning history leaves open the reasonable probability that the BOCC would approve MDR-level development on B-10. To determine a property’s highest and best reasonably probable use, the court focuses on “[t]he highest and most profitable use for which the property is adaptable and needed or likely to be needed in the reasonably near future.” Symington, 87 T.C. at 897 (quoting Olson, 292 U.S. at 255, 54 S.Ct. at 708-09). 9 The IRS contends that it was not reasonably probable for the BOCC to approve MDR-density development on B-10 because “[t]he rezoning history, which culminated in the issuance of Ordinance No.2004-055, strongly suggests that the BOCC would not have permitted [MDR-level] development.” Besides the rezoning history considered by the tax court (including the BOCC’s two three-two votes and the language of the ordinance), the IRS invokes the BOCC meeting minutes and an information packet from the second meeting. According to the IRS, the meeting minutes, which the tax court rejected as inadmissible hearsay, illustrate BOCC members’ far-reaching “density and wildlife corridor concerns” with approving development on B-10. And the information packet confirms these concerns by stating that at the first meeting the BOCC “advised [Pulte] to revise [its] request to decrease the number of units and only propose development in the northern half of the increment [B-9] in order to lessen the impact to the Eagle Protection Zone.” The IRS says that the tax court was wrong to reject the meeting minutes as hearsay because the statements show the BOCC members’ states of mind regarding whether to approve development on B-10. Once considered, the state-' ments coalesce with the information packet and other rezoning history to show that the BOCC would not have approved MDR-level development. Therefore, the IRS argues, it was not reasonably probable for the BOCC to approve MDR-density development.

The IRS is wrong. We may assume that the tax court erred by rejecting the BOCC meeting minutes as inadmissible hearsay and that the minutes evidence considerable apprehension by individual BOCC members about approving MDR-level development. 10 Whatever was said in the meeting minutes, the best evidence of whether the BOCC would approve MDR-density development on B-10 is the BOCC’s actual voting record: what the BOCC set down as democratically vetted markers of its position. As the tax court noted, the BOCC voted to deny Pulte’s first development application, which, though it covered B-10, excluded the eagle nest zone. After the BOCC denied that development application, Pulte submitted a second application that left B-10 out altogether. The BOCC voted to deny this application as well, and issued an ordinance instructing that future development applications should “include Parcel B-10 and should endeavor to keep Parcel B-10 [intact] as it relates to the Eagle Preservation Area, the wetlands, and the wildlife corridor.” We agree with the tax court that the ordinance and the votes leading up to it suggest that the BOCC would approve any development that (1) included BIO and (2) addressed the BOCC’s environmental concerns. Because an MDR-level development (namely, a development under the WilsonMiller plan) could satisfy these criteria, the BOCC would, in reasonable probability, approve such development. The IRS tries to tell a different story by invoking the BOCC meeting minutes, but the minutes simply cannot compete with the final position that the BOCC memorialized in its votes and in the language of the ordinance. 11

2. Although the Tax Court Legally Erred By “Fail[ing] to Address Whether [Palmer Ranch’s] Proposed 360-unit Development Was Needed or Likely to Be Needed Shortly After December 2006,” the Error is Harmless.

The IRS argues that the tax court should have included in its highest-and-best use analysis an inquiry into whether development on B-10 “was needed or likely to be needed shortly after December 2006.” According to the IRS, such an inquiry was legally required. The IRS fur ther- argues that, had the tax court made this mandatory inquiry, it would have concluded that development of 360 units was not needed or likely to be needed. Dur-rance testified that there was “a pause in the market ... a softening ... of activity in 2006.” And the property whose 2006 sale the parties considered “the only true comparable” lay vacant until “the last half of 2012 into 2013” due to unfavorable economic conditions. Even when development started, the developer “lowered the density” below the 303 units originally planned. Therefore, the IRS concludes, development on B-10 was not needed or likely to be needed around December 2006.

The IRS is half right. It is true that the highest-and-best-use test requires an inquiry into whether the proposed use will, in reasonable probability, be “needed or likely to be needed in the reasonably near future” — in other words, into whether the market will demand the use. Symington, 87 T.C. at 897 (quoting Olson, 292 U.S. at 255, 54 S.Ct. at 708-09). 12 And no, the tax court did not undertake the requisite market-demand inquiry. The tax court limited its highest-and-best-use analysis to consideration of whether the BOCC would, in reasonable probability, approve MDR-density development on B-10. The tax court did not consider whether, were the BOCC to approve development, it was reasonably probable that a developer would perceive enough MDR-level housing demand to actually break ground on B-10. This was error.

But the error is harmless. The evidence clearly shows that, in 2006, the market for MDR-level development was bullish. Durrance reported that “multistory, attached residential housing developments within Palmer Ranch were effectively sold out” in 2006, thus increasing demand for B-10 as “one of a few large parcels that [wa]s available for residential development in the urban Sarasota area.” And although Durranee acknowledged a “softening” of demand for residential real estate in 2006, he noted that land values did not suffer a decrease. On the IRS’s side of the evidentiary ledger, Page agreed that in the beginning of 2006 there was “high demand for open land in Sarasota.” The demand for MDR-level development is also illustrated by the 2006 sale of B-10’s “only true comparable” property for $40 million. The property sold on the expectation that it would be developed at a density on par with MDR, with no unusual delay in the development timeline (that is, no delay beyond the typical “site development and engineering process”). In other words, at the time that B-10’s only true comparable changed hands, MDR-level development on the parcel looked “needed or likely to be needed in the reasonably near future.” It stands to reason that the same would be true for B-10. 13

As to the comparable sale, the IRS counters with retrospect. The underlying property lay vacant through “the last half of 2012 into 2018.” But the thing that matters is what appeared “needed or likely to be needed” as of 2006, not what turned out to be true. Besides, the IRS does not explain why a developer would not consider 2013 the “reasonably near future” from the vantage of 2006. As already noted, one reason development did not start until 2013 was “the site development and engineering process,” a process that is normal for any national development project and that developers would therefore expect.

In sum, the highest-and-best-use test includes within it an analysis of whether the proposed use is “needed or likely to be needed in the reasonably near future.” The tax court erred by omitting this analysis. But the error is harmless in this case because in 2006 the Sarasota market clearly demanded MDR-level development.

3. The Tax Court Did Not Clearly Err by “Failing to Consider the Effect of the Zoning History on Hypothetical Willing Buyers of B-10.”

The IRS next argues that the tax court should have considered the effect that B-10’s failed rezoning history would have on “hypothetical willing buyers.” According to the IRS, hypothetical buyers would glean from B-10’s rezoning history a risk that B-10 could not be developed to MDR density, and would respond by discounting the purchase price. The IRS says that the tax court should have calibrated B-10’s “before” value to the hypothetical buyers’ discounted price.

Although the IRS frames this as an argument against the tax court’s highest- and-best-use determination, it is really an attack on the tax court’s subsequent “before” valuation. That is, the tax court first determined that B-10’s highest and best use was MDR-level residential development, and subsequently calculated B-10’s before value. In calculating the before value, the tax court did not discount for the risk that B-10 could not be developed to MDR density. The IRS argues that this was error.

The IRS’s argument does not convince. In certain cases, it can be appropriate to discount a property’s before value to account for a risk that the property cannot be developed to the highest and best use. See, e.g., Frazee v. Comm’r, 98 T.C. 554, 578 (1992) (“[E]ven if it was reasonably foreseeable that an attempt to rezone the Carlsbad property would be favorable, its value would have to be discounted due to the uncertainty associated with the rezoning process....”) But such a discount will not always be suitable because the process for determining a property’s highest and best use itself eliminates proposed uses on the basis of development risk. In order to constitute the highest and best reasonably probable use, a proposed use must be one “for which the property is adaptable and needed or likely to be needed in the reasonably near future.” Symington, 87 T.C. at 897 (quoting Olson, 292 U.S. at 255, 54 S.Ct. at 708-09). If there is too high a chance that the property will not achieve the proposed use in the near future, then the use is too risky to qualify. Id. 14 So the test for highest and best use already bakes in some adjustment for development risk. In some cases, the test’s culling of proposed uses is enough, and no further accommodation to development risk needs to be made at the valuation stage.

This is just such a case. In this case, the tax court undertook a thoroughgoing analysis of the parcel’s rezoning history— in particular the BOCC’s vote on Pulte’s first development application and the language of the ordinance. On the basis of this evidence, the tax court concluded that it was reasonably probable for the BOCC to approve development at MDR density. And the evidence clearly shows that, had the BOCC approved MDR-level development on B-10, the 2006 market was poised to demand that development as “needed or likely to be needed.” Put another way, from the vantage of 2006, there was substantially no risk that B-10 would not be developed at MDR density in the near future. No risk means no reason for a risk-based discount at the valuation stage. The tax court’s failure to provide such a discount therefore was not clear error.

Nor are the cases cited by the IRS to the contrary. None of the cases suggest that courts must always discount properties’ before values to account for . a risk of nondevelopment, even when no such risk exists. See Frazee, 98 T.C. at 578-79; Mathis v. Comm’r, 57 T.C.M. (CCH) 519 (1989), No. 37395-85, 1989 WL 53828; Hay v. Comm’r, 64 T.C.M. (CCH) 228 (1992), Nos. 27785-89, 27786-89, 1992 WL 166154. Also, Mathis is distinguishable because it involved a tax-deductible contribution of property, not of a conservation easement on property. Because the issue was the value of the property itself (not the easement), there was no highest-and-best-use inquiry to capture some of the risk of nondevelopment. And Hay did not discuss risk discounts at all, but rather discounts for “the cost of removing the [zoning] restriction and for the time needed to accomplish such removal.” 64 T.C.M. (CCH) 228. Because these cases do not counsel a different result, we conclude that the tax court was not required to discount its before valuation of B10 to account for nondevelopment risk. 15 That is the last issue raised by the IRS, so we now move to Palmer Ranch’s appeal.

C. The Tax Court Erred By Reducing Durrance’s Proposed “Before” Value from $25,200,000 to $21,005,278 to Account for a Stagnation of Property Values in 2006.

Palmer Ranch claims error in the tax court’s decision to reduce Durrance’s proposed before value from $25,200,000 to $21,005,278. The tax court tailored the reduction to fit “the amount of 2006 appreciation Mr. Durrance calculated” in his second appraisal of B-10. In the second appraisal, Durrance had framed his valuation of B-10 as consistent with John Underwood’s IRS-side appraisal in the Rogers takings litigation. Durrance did so by factoring a $200,000 per-acre valuation with two “monthly appreciation rate[s]” that stemmed from Underwood’s work: 1.5% over 33 months (April 2004 through December 2006) or the 2.7% rate over 21 months (April 2004 through December 2005). Whichever rate Durrance applied, he wound up with an “increase in the value of [B-10] between April 2004 and December 2006 of essentially 50%.” This result pretty much matched Durrance’s 2006 valuation of $307,000 per acre, $25,200,000 total. But the tax court was critical of Durrance’s use of the appreciation rates, calling the work “incomplete.” According to the tax court, although Underwood “stated in his Rogers appraisal that there was a 2.7% per month increase in values from 2004 to 2005,” he also “said appreciation rates for raw land [like B-10] lag[ged] behind” the 2.7% rate. Underwood pegged the raw-land rate at 1.5%, and B-10 was raw land, so Durrance should have limited his calculations to the 1.5% rate. And Durrance should not have applied the 1.5% appreciation rate across 33 months, including 2006, because Durrance had conceded on cross-examination “that the real estate market peaked in late 2005.” The tax court sought to cure the issue by removing all 1.5% appreciation that Dur-rance had included for 2006. This left the tax court with a before value of $21,005,278.

Palmer Ranch is not happy. According to Palmer Ranch, once the tax court accepted Durrance’s appraisal as “more accurate” than Page’s, the burden shifted to the IRS to prove Palmer Ranch’s tax liability. But Page’s appraisal could not serve to discharge this burden because that appraisal was predicated on a lowball highest-and-best use that the tax court had already rejected. Therefore, says Palmer Ranch, the tax court was bound to adopt Durrance’s valuation dollar for dollar without entertaining any reduction. Palmer Ranch further argues that the tax court legally erred by “concoct[ing] sua sponte a valuation method [that was (1) ] advanced by neither party” and (2) “contrary to established [comparable-sales] appraisal principles.” Next, Palmer Ranch argues that the tax court impermissibly ventured beyond the evidentiary record in its discussion of Underwood’s monthly appreciation rates. Finally, Palmer Ranch argues that the tax court clearly erred “in finding that there was no appreciation throughout 2006,” and in any case incorrectly calculated the amount of pre-2006 appreciation.

We disagree with Palmer Ranch’s first contention, that the tax court erred by entertaining any reduction to Durrance’s valuation of B-10. Palmer Ranch marshals three cases in support of this argument, but none of them provide the necessary backing. To begin with, both Estate of Elkins v. Commissioner, 767 F.3d 443 (5th Cir.2014), and Carson v. United States, 560 F.2d 693 (5th Cir.1977), are different from this case because in those cases the IRS presented no evidence whatsoever to substantiate the position it was taking. The Fifth Circuit held that, regardless of whether the IRS or the taxpayer bore the burden of proof, the tax court was bound to accept the taxpayer’s stance when the IRS countered with nothing at all. Elkins, 767 F.3d at 450; Carson, 560 F.2d at 696. That rule does not apply to this case because the IRS presented a suite of evidence on the B-10’s value, including but not limited to Page’s thoroughgoing appraisal. Therefore, neither Elkins nor Carson effectuate Palmer Ranch’s argument that the tax court was obligated, to accept Durrance’s valuation of B-10 point blank.

The same goes for Palmer Ranch’s last case purportedly on point, Caracci v. Commissioner, 456 F.3d 444 (5th Cir.2006). Caracci is an excise tax case in which the Fifth Circuit held that the burden of proof shifted to the IRS when the agency conceded that 'its tax deficiency notices were excessive and erroneous. 456 F.3d at 457. This case is different from Caracci because it is a deduction case: Palmer Ranch claimed a deduction on its 2006 tax returns, and the IRS disputes the amount of the deduction claimed. In deduction cases, there is no burden shifting. Instead, the taxpayer retains the burden of proving the amount of his deduction “[a]t all times.” Gatlin v. Comm’r, 754 F.2d 921, 923-24 (11th Cir.1985). At least in these kinds of cases, the tax court is not bound to accept the taxpayer’s valuation in full merely because the IRS’s valuation is unsatisfactory: The tax court may determine its own valuation based on the whole of the evidentiary record. Id. 16 It necessarily follows that the tax court is free to treat the work of the taxpayer’s expert as a starting point, which is what the tax court did here. Because Caracci does not mandate a different rule, we hold that the tax court did not err merely by looking past Durrance’s face valuation of B-10. 17

However, we still find fault in the tax court’s valuation, insofar as we agree with Palmer Ranch that the tax court’s choice of valuation method was problematic. As the tax court recognized, “[b]oth parties’ appraisers used the comparable • sales method as a basis for their before and after valuations.” The tax court never gave any indication that it considered the comparable-sales method inappropriate. To the contrary, the tax court endorsed Durrance’s appraisal as “more accurate” than Page’s in part because Durrance’s valuation fell within the “price-per-acre range” of both appraisers’ comparable sales. Yet — without even acknowledging that it was doing so or explaining why— the tax court departed from the comparable-sales method in the course of reducing Durrance’s valuation from $25,200,000 to $21,005,278.

Specifically, to reach a value of $21,005,278, the tax court looked not to the comparable sales both Durrance and Page had vetted, but instead to a 2004 valuation of B-10. Durrance had included the 2004 valuation in his second appraisal only to frame his own valuation as consistent with Underwood’s work in Rogers. Doing so required carrying the $200,000 2004 valuation forward to 2006, which Durrance accomplished using “monthly appreciation rate[s]” that he had gotten from Underwood. The tax court valued B-10 by wiping all 2006 interest from Durrance’s carry-forward calculation on grounds that “Durrance stated that the real estate market peaked in late 2005.” Therefore, the tax court’s $21,005,278 valuation was derived entirely from Durrarice’s cross-check reconciliation exercise — not from comparable sales. In other words, the tax court’s valuation was premised on an old appraisal as modified by monthly appreciation rates, instead of on comparable sales. Without being based on comparable sales, the valuation cannot have been a comparable-sales valuation. Because the parties’ appraisers both used the comparable-sales method, and because the tax court neither voiced disapproval nor acknowledged (much less explained) its departure from the method, that departure was error. The tax court must at minimum explain why it departed from the comparable-sales method in valuing B-10. 18

Even were the tax court’s unexplained departure from the comparable-sales method proper, the tax court still erred by relying on evidence outside the record to value B-10. To execute its chosen valuation method, the tax court had to decide between Underwood’s 1.5% monthly appreciation rate and his 2.7% rate. The tax court settled on the 1.5% rate on the basis of information from Underwood’s appraisal report. Underwood stated in his appraisal report that the 1.5% rate pertained to raw land like B-10, whereas the 2.7% rate did not, so the tax court went with the 1.5% rate. But Underwood’s appraisal report was not in the evidentiary record before the tax court, so the tax court erred by relying on the report in selecting an appreciation rate. Cf. Johnson v. United States, 780 F.2d 902, 910 (11th Cir.1986) (“The trial judge may not ... undertake an inde pendent mission of finding facts ‘outside the record of a bench trial over which he [presides].’ ” (citation omitted)). 19

Finally and relatedly, the tax court may have erred in how it applied the 1.5% monthly appreciation rate. The tax court reached a valuation of $21,005,278 by reducing Durrance’s proposed before value by “the amount of 2006 appreciation Mr. Durrance calculated” at the 1.5% rate. At least on the limited record before this court (which does not include month-to-month interest calculations by Durrance— only a narrative description of Durrance’s formula), we do not discern how removing 1.5% appreciation per each of the twelve months in 2006 brought the valuation to $21,005,278. To avoid muddying the waters on remand, we will leave our mathematical critique at that.

On remand, then, the tax court must either stick with the comparable-sales analysis or explain its departure. Whatever the tax court chooses to do, the court must keep its sights set strictly on the evidentiary record for purposes of selecting an appreciation rate, and ensure that it crunches the numbers correctly. 20

CONCLUSION

With regard for B-10’s proudest avian residents and for the enduring American principles they have come to symbolize, it is time now to close. We affirm the tax court’s determination of B-10’s highest and best use, but reverse the ensuing valuation. We remand for the tax court to revalue B-10 in a manner not inconsistent with this opinion.

AFFIRMED IN PART, REVERSED IN PART, AND REMANDED FOR FURTHER PROCEEDINGS.

1 . One might say Palmer Ranch confected.a "donut of intent” regarding B-10. R. at 1439 (“[We] included those exhibits when we believed the donut of intent and the charitable nature of the contribution was at issue.”) This presumably should not be confused with the more vanilla "donative intent.”

2 . Comparable 1 did not actually switch hands until January 2006, although the parties (evidently) agreed on a price and executed the contract the preceding fall. The total price was $40,000,000 for 78.34 acres.

3 . Durrance’s claim that Underwood valued Palmer Ranch land at $200,000 conflicts with the Court of Federal Claims’ opinion in Child-ers v. United States, 112 Fed.Cl. 617 (2013) {Childers being the opinion that actually recounts the takings litigation appraisals). The Childers court said Underwood valued B-9 and B-10 at $165,000 per acre. 112 Fed.Cl. at 693-94.

In his second appraisal, Durrance said Underwood’s $165,000 per-acre valuation was consistent with a $200,000 valuation because the $165,000 valuation featured a built-in “large tract discount” of 20%, which would not apply to the 82-acre B-10 standing alone (as opposed to B-0 and B-10 together). In Durrance’s words,

Mr. Underwood generally concludes the land value of property in Palmer Ranch is worth $200,000 per gross acre for 10 to 25 acre tracts, including property that is not in the DRI. (Property in the DRI would be more valuable than property not in the Palmer Ranch DRI.) ... But, for tracts of land larger than 25 acres, Mr. Underwood applied a size discount. In other words, Mr. Underwood concluded that simply because the owner held title to a larger tract of land, the property — by virtue of that fact alone — was less valuable. On this basis, Mr. Underwood discounted his $200,000 per acre value for the combined Parcel B-9 and Parcel B10 by 20% for a value of $165,000 per acre in April 2004. But the Underwood size discount is not relevant to the appraisal of Parcel B-10, as it is only 82 acres.

This summary at best oversimplifies the mechanics of Underwood’s large-lot discount. Underwood applied a large-lot discount to accommodate his observation that larger tracts of land sold at lower per-acre prices than smaller tracts of land. Childers, 112 Fed.Cl. at 631-32. Based on this observation, Underwood attempted to neutralize size-based price discrepancies when using comparable sales to value B-9 and B-10. Specifically, B-9 and B-10 being of a determinate size, Underwood adjusted the prices of higher acreage sales upwards to approximate the higher per-acre price that B-9 and B-10 would themselves fetch. Id. at 692-93. And Underwood did the same with lower acreage sales, only he adjusted downwards.

In valuing B-9 and B-10, Underwood gave "the most weight” to a comparable sale of 199.4 acres for $140,393 per acre. 112 Fed. Cl. at 693-94. Because the comparable sale covered a larger acreage area than B-9 and B-10, Underwood adjusted the comparable’s per-acre price upwards 15%. Id. at 692-93. Thus, Underwood deemed the closest comparable’s functional per-acre price (that is, the per-acre price the comparable would have had if the same size as B-9 and B-10 together) to be $161,452.

Underwood also constructed functional prices for the other comparable sales he considered. Because some of the sales involved parcels smaller than B-9 and B-10, the functional per-acre prices were lower, not higher, than the actual prices. The largest downward adjustment Underwood made was 20%. On the basis of all these comparable sales, including the functional $161,452 sale to which he gave “the most weight,” Underwood valued B-9 and B-10 at $165,000 per acre.

What Underwood did not do was reach a valuation of $200,000 per acre and then lop off 20% simply because B-9 and B-10 together covered a large acreage. While Underwood's methodology might have led him to estimate a higher per-acre value than $165,000 were he to value B-10 alone, nothing in his methodology suggests that he would have arrived at $200,000. Durrance appears to have gotten the 20% figure (which he uses to alchemize $165,000 into $200,000) from the largest downward adjustment that Underwood made. But, in applying this adjustment directly to an imaginary valuation of $200,000, Durrance robbed Underwood’s appraisal of context and reason.

4 . According to the tax court’s eventual opinion, "the parties agree[d] that comparable 1 [wa]s the only trae comparable.”

5 . This doctrine has also been coined the accord and satisfaction doctrine. See Crawford v. Andrew Sys., Inc., 39 F.3d 1151, 1153 (11th Cir.1994).

6 . We additionally note that judicial estoppel arguments are fueled by "inconsistent positions ... [taken] under oath in a prior proceeding.” Bumes, 291 F.3d at 1285. It is unclear to us how the IRS’s pursuit of a levy amounts to making a statement under oath in a proceeding.

7 . See, e.g., Alvarez Perez v. Sanford-Orlando Kennel Club, Inc., 518 F.3d 1302, 1305-06 (11th Cir.2008) (considering whether "defendants paying the plaintiff not only the original amount of the judgment ... but also trial level attorney’s fees ... plus interest” and signing and filing "satisfaction of judgment'” documents "rendered [defendants' appeal for inadequate damages] moot”); Crawford, 39 F.3d at 1153 (considering whether plaintiffs' withdrawal of funds associated with one undisputed judgment barred pursuit of appeal of second judgment); Fidelcor Mortg. Corp. v. Ins. Co. of North America, 820 F.2d 367 (11th Cir.1987) (considering whether full payment of judgment and execution and filing of a standard satisfaction of judgment form barred appeal).

8 . We leave for another day the question of whether, had Palmer Ranch paid the IRS, "the parties’ objective manifestations of intent” would bar appeal. This strikes us as a thorny question in light of 26 U.S.C. § 7485, which authorizes the IRS to collect on a deficiency simultaneous to an appeal.

9 . See also 26 C.F.R. § 1.170A-14(h)(3)(ii) (conservation restriction’s fair market value must “take into account ... an objective assessment of how immediate or remote the likelihood is that the property, absent the restriction, .would in fact be developed, as well as any effect from zoning, conservation, or historic preservation laws that already restrict the property’s potential highest and best use”).

10 . We note, however, that some of the statements cannot possibly be construed as indicating BOCC members' states of mind because the speakers are not BOCC members, but instead representatives of Pulte. See Fed. R Evid. 803(3) (excepting from the rule against hearsay any "statement of the declar-ant's then-existing state of mind”). It also is not evident from the record on appeal that the IRS ever argued before the tax court that the meeting minutes were admissible on a state of mind theory, as opposed to the business-records exception to the hearsay rule. And even if the Pulte representatives' statements are admissible as nonhearsay, they are of limited probative value in determining whether the BOCC itself would or would not approve MDR-level development.

11 . The IRS also argues that the tax court "largely assumed” that it was reasonably probable for the BOCC to approve MDR-level development, and improperly beset the IRS with the burden of proving otherwise. We are not persuaded. Concededly, some language in the tax court’s opinion can be read as imposing on the IRS a burden to disprove the reasonable probability of BOCC approval. {E.g., "[T]his rezoning history does not eliminate the reasonable probability on the valuation date of a successful rezoning.'') But at the end of the day the tax court rested its reasonable-probability-of-BOCC-approval determination on a thoroughgoing interpretation of the BOCC’s voting record and the ordinance as well as other supporting evidence — not on assumption.

12 . See also 26 C.F.R. § 1.170A-14(h)(3)(h) (conservation restriction's fair market value must "take into account ... an objective assessment of how immediate or remote the likelihood is that the property, absent the restriction, would in fact be developed” at the proposed level).

Esgar v. Commissioner, 744 F.3d 648 (10th Cir.2014), is also informative. In Esgar, the Tenth Circuit considered whether certain properties’ highest and best use was the current use, agriculture, or was instead gravel mining. 744 F.3d at 658. In tax court, the parties had “stipulated that, absent the conservation easements, it was likely that the necessary permits to mine gravel could be obtained.” Id. at 652 n. 4. Despite this stipulation, the tax court had selected agriculture (not gravel mining) as the properties! highest and best use. The tax court had explained that "there was no unfilled demand [for gravel] and ... no unmet market,” plus "no evidence” that "this was to change in the reasonably foreseeable future.” Esgar v. Comm’r, 103 T.C.M. (CCH) 1185 (2012), Nos. 23676-08, 23688-08, 23689-08, 2012 WL 371809, at *19. On review, the Tenth Circuit held that the tax court had not clearly erred in its application of the highest-and-best-use test. 744 F.3d at 658. According to the Tenth Circuit, “the [t]ax [c]ourt was not bound to accept gravel mining as the properties’ highest and best use if that use was not reasonably probable to manifest in the reasonably near future.” Id. That is, just because the properties could be rezoned for gravel mining did not mean that gravel mining was the highest and best use. Determining the properties’ highest and best use necessitated a look at market demand, so the tax court could not be faulted for taking that look.

13 . Comparable l’s $40 million price tag also independently suggests that B-1Ó could be valued at (at least) $25.2 million. Of course, as Durrance acknowledged; there are other facts that indicate that B-10's value should be at least somewhat lower than Comparable l's.

14 . See also 26 C.F.R. § 1.170A-14(h)(3)(ii) (conservation restriction's fair market value must "take into account ... an objective assessment of how immediate or remote the likelihood is that the property, absent the restriction, would in fact be developed" at the proposed level). The principle can also be articulated in terms of willingness to pay. If a proposed use is too risky for "a hypothetical willing buyer [to] consider [the use] in deciding how much to pay for the property,” then the use should not be deemed the highest and best available. Whitehouse Hotel Ltd. P'ship v. Comm’r, 615 F.3d 321, 335 (5th Cir.2010).

15 . In a footnote, the IRS argues that

the [t]ax [c]ourt gave no consideration to [i.e. provided no discount for] the fact that an actual willing buyer (Pulte) was willing to pay only $8.25 million for B-9 and B-10 together before any of these zoning proceedings made development less attractive.... Notably, the parties assigned virtually no value to B-10 [when they dropped the parcel from the sale in 2004], as they did not reduce the $8.25 million purchase price.

The IRS’s argument fails because it is premised on sales contracts from 2003 and 2004, while the tax court was valuing B-10 in 2006. As evidenced by "the only true comparable” sale — a sale of 78.34 acres for $40 million completed in January 2006 — prices from 2003 and 2004 have little bearing on prices two years down the road.

16 . Cf. Lukens v. Comm’r, 945 F.2d 92, 96-97 (5th Cir.1991) (deduction case in which the Fifth Circuit affirmed tax court’s rejection of three expert valuations in favor of "its own [lower] valuation [based on] what credible testimony was proffered”); Orth v. Comm’r, 813 F.2d 837, 841-42 (7th Cir.1987) (26 U.S.C. § 170 charitable deduction case in which the Seventh Circuit affirmed the tax court’s calculation of its own valuation, notwithstanding two expert appraisals on record).

17 . Actually, Elkins and Carson are both distinguishable from this case on grounds that they are not deduction cases and that the burden of proof therefore operated in a different manner. Elkins, 767 F.3d at 449-50; Carson, 560 F.2d at 696. But in both cases the Fifth Circuit declined to rely on the burden of proof in reaching its holding, so the more honest distinction is the one offered above.

18 . To be as clear as the Sarasota sky, we do not hold that the tax court is obligated to use the comparable-sales method to value B-10. According to Palmer Ranch, the tax court is forbidden from adopting (or, in Palmer Ranch's ostensibly outraged parlance, “sua sponte concocting”) an appraisal method proposed by neither party. But we recognize the tax court’s expertise in adjudicating tax disputes, and therefore will not lay down a hardline rule confining the tax court to the parties’ preferred appraisal method(s) in every case. The tax court has discretion to adopt a valuation method befitting the matter before it— even if the parties have not proposed that method. The cases relied upon by Palmer Ranch are not to the contrary because in each there was some identifiable flaw with the tax court’s method besides that it had not been suggested by the parties. Elkins, 767 F.3d at 445 (tax court erred by discounting valuation ten percent “despite the absence of any record evidence whatsoever on which to base the quantum of its self-labeled nominal discount”); Succession of McCord v. Comm’r, 461 F.3d 614, 626 (5th Cir.2006) (tax court erred by engaging in the "long-prohibited practice of relying on post-gift events” for valuation), Caracci, 456 F.3d at 458 (tax court should have rejected valuation method that had "no legal support” and that both parties’ experts had recognized as "inferior” and "less rigorous”).

In sum, then, we hold only that, when the parties both use one method to value a parcel of land, the tax court errs by departing from that method without acknowledgment and explanation. We do additionally note that it might be difficult to explain why valuing B-10 using data from a two-year-old appraisal in a different case makes more sense than relying on the valuations actually presented to the tax court.

19 . Palmer Ranch also argues that the tax court erred by considering hearsay newspaper articles in order to reach the conclusion that Sarasota property values did not appreciate during 2006. No the tax court did not. The tax court referenced the newspaper articles only while summarizing the IRS’s no-appreciation argument. In its analysis, however, the tax court relied exclusively on Dur-rance's testimony during cross-examination.

20 . Our holding in no way precludes the tax court from valuing B-10 below $25,200,000. For example, suppose the tax court applies the comparable-sales method to value B-10. Both appraisers agreed on four comparable sales ranging from $237,200 to $510,600 in per-acre price. Durrance’s appraisal, which the tax court found "more accurate” than Page's equates to $307,000 per acre. In his appraisal reports, Durrance noted that the market stagnated in 2006. However, Dur-ranee admitted at trial that he made no "quantitative adjustments” to the appraisal value to account for the state of the market in 2006. The tax court could certainly make its own quantitative adjustment, particularly since Durrance’s appraisal falls well within the range of comparables. Any quantitative adjustment, however, would have to account for the qualitative adjustment that Durrance did make. See R. at 111 (reaching $70,000 per-acre valuation — at low end of comparable sales range — in light of the "softening of the residential market in 2006”). Alternatively, the tax court could simply perform its own comparable-sales analysis on the basis of the agreed-upon comparables. .(Incidentally, the tax court’s original appraisal worked out to $255,570 per acre, also within the range of comparables.) Whatever route the tax court pursues, it retains its obligation to base its valuation on the evidence.