Full opinion text
JAMES LAWRENCE KING, District Judge. Taxpayers appeal the decision of the district court in which deficiencies were assessed against them, after a full trial, as follows: (i) against the Loftin and Woodard Corporation for fiscal years 1963, 1964 and 1966; (ii) against K. C. and Marie Loftin, as joint taxpayers, for calendar years 1962, 1963 and 1964; and (iii) against E. A. and Virginia Woodard, as joint taxpayers, for calendar years 1963 and 1964. In addition, the district court found Loftin liable for fraud as to tax years 1963 and 1964. At the same time, we are confronted with a cross-appeal lodged by the Government. In essence, the Government asserts that the district court’s failure to find fraud as to the Loftin & Woodard Corporation was incorrect. Further, the cross-appeal focuses on the district court’s decision not to utilize a late filing penalty to diminish the refund found to be owing to the corporation on the basis of an audit of its 1961 return. I. THE FACTUAL BACKGROUND: The factual context of this case is quite complex. We have before us three sets of taxpayers: K.C. and Marie Loftin [hereafter “Lof-tin”]; E.A. and Virginia Woodard [hereafter “Woodard”]; and the Loftin and Woodard Corporation [hereafter the “Corporation”]. The Loftins filed a joint return on a calendar year accounting basis, as did the Woodards. The Corporation filed its return on a fiscal year interval beginning on June 1 of each year. Loftin and Woodard were the sole stockholders of the Corporation. K.C. Loftin owned a 65% interest. The remaining 35% was under Woodard’s control. The Corporation was created primarily for the purpose of clearing land. It was this task to which it turned most of its efforts in the tax years at issue on this appeal — that is, 1961-1967. The central transaction in this appeal involves the purchase by the Corporation of over nine thousand acres of wooded land in Yazoo County, Mississippi on April 17,1963. Prior to the corporate purchase, Loftin and Woodard contemplated the formation of a partnership for the consummation of the transaction. However, because the partnership would have been unable to secure adequate funds in the requisite period of time, the would-be partners decided that the Corporation would have to serve as the purchasing entity. On April 17, the Corporation secured title to the first six thousand acres of the Yazoo property at a cost of $225,542.30. Prior to the closing, the Corporation adopted a resolution emphasizing its intention to sell the land to the partnership once the partnership secured the requisite funds. The transfer eventually occurred on November 1, 1963 — almost five months after the Corporation’s purchase. At this time, the partnership paid the Corporation the amount that the latter entity had paid on April 17. It is significant that prior to the November 1 transfer, the Corporation began its clearing operation on the property. These operations were performed pursuant to the authorization of the partnership. Although no actual clearing was done prior to November 1, the Corporation did repair several structures already on the land; clear boundary lines; and establish further drainage facilities, all in preparation for the eventual clearing operation. In fact, no land was cleared in fiscal year 1963.® By the end of fiscal year 1964, eight hundred acres of the Yazoo land had been cleared. By 1967, over nine thousand, six hundred acres were cleared. In arriving at a more complete understanding of the nature of this land clearing operation, one must understand the manner in which the taxpayers reflected it on paper. It is uncontested that the expenses incurred in the initial stages of the operation — that is, from April 17,1963 to May 31, 1964 (the end of fiscal 1964) — were not registered in a separately labeled category on the corporate ledger. Instead, these expenses were spread among some of the other corporate expense accounts in existence at that time. It was not until the end of the corporate fiscal year 1965, when 5,700 acres had been cleared, that the corporate ledger carried a separate category for Ya-zoo related expenses. The record of payments made by the partnership is equally ambiguous in the early stages of the clearing operation. The partnership ledger reflects a payment to the Corporation on the final day of the partnership’s tax year, December 31, 1964. Payments by the partnership to the Corporation were consummated on paper by the entry of credits to an account receivable due from the Corporation. The remainder of the 9,620 acres within the control of the partnership by May 31, 1967 was cleared in fiscal 1966 and 1967. Thus, payments for the land clearing operation progressed, on paper, as follows: DATE PARTNERSHIP CORPORATE ACCT. ACCT. RECEIVABLE PAYABLE 12/31/64 -$250,000 5/31/65 + $325,000 12/31/65 -$196,000 + $121,000 12/31/66 -$ 35,000 5/31/67 + $ 35,000 -$481,000 + $481,000 Credits entered in the partnership account receivable theoretically reflected payments made by the partnership to the Corporation. Receipt of those payments by the Corporation was reflected by a debit to its account payable. As is apparent on the face of these transactions, the Corporation habitually made its entries on the final day of its respective tax year and the partnership did likewise. The one exception to this habit, the corporate entry on December 31, 1965, was occasioned ostensibly by a desire to equalize the recordation of payments made and received up to and including that date. At the same time that the payments for these operations were being recorded, the Corporation incurred expenses that exceeded, in toto, the payments made. This, then, is the material from which the district court fashioned the constructive dividend contested herein. The district court, after an extensive trial, decided that a constructive dividend accrued to Loftin and to Woodard on the basis of the work performed by the Corporation for the partnership. The unreported income embodied in the constructive dividend finding formed the core of the deficiency asserted by the district court against both taxpayers for 1963, 1964 and 1966. At the same time, the lower court decided that several expenditures made by the Corporation on behalf of Loftin and Woodard were not business related and were, therefore, also in the nature of constructive dividends. These other expenses centered on trips taken by the taxpayers; advances received by them; and the payment of personal insurance premiums, political campaign expenses, and football tickets by the Corporation for Loftin and/or Woodard. The taxability of these items in terms of the constructive dividend question is not on appeal. However, their relationship to the finding of fraud made as to taxpayer Loftin is hotly contested and is the subject of Section III, infra. Finally, there is one further aspect of the land clearing operation with which we must concern ourselves on this appeal — the expenses incurred by the Corporation. While this issue is at the core of the constructive dividend question, it poses an additional problem for our consideration. The district court found that the corporate expenses incurred in the land clearing operation were, for the most part, non-deductible as ordinary and necessary business expenses. The ultimate resolution of this issue is highly dependent upon the determinations we make with regard to the constructive dividend issue. The only other issue with which we are confronted on this appeal does not stem from the land clearing operation. As will be discussed in greater detail in a later portion of this opinion, this final matter is related to an audit of the Corporate return filed for 1961. In sum, we are confronted with a factual context which may be viewed from a tripartite perspective. First, there is the relationship of the land clearing operation to the issue of a constructive dividend. Second, there is the relationship of the land clearing operation to the corporate decision to deduct expenses arising therefrom. And third, there is the relationship of (i) the constructive dividend, found to have arisen from the land clearing operation, along with the corporate payment of several expenses incurred by the taxpayers to (ii) the intent of the taxpayers in not declaring these payments as income. An additional issue, unrelated to the above, centers on the Corporation’s 1961 return. This is a complex case, admitting of no simple answers. An excellent trial record, along with lengthy and effective appellate briefs and argument have made our task much easier. The record clearly reflects that Judge Hunter diligently and relentlessly pursued the clarification of this factual ticket. In the end, we are “mere mortals unendowed with cosmic tax wisdom[. W]e have performed our task as well as our fallible mentalities and compositions will permit.” Cornelius v. Commissioner of Internal Revenue, 494 F.2d 465, 472 (5th Cir. 1974). 1107 pages of record; 264 pages of briefs; and almost two hours of oral argument later, these are our conclusions. II. THE CONSTRUCTIVE DIVIDEND ISSUE: When a corporation confers an economic benefit upon a shareholder, in his capacity as such, without an expectation of reimbursement, that economic benefit becomes a constructive dividend, taxable to the respective shareholder. This benefit is taxable to the shareholder whether or not the corporation intended to confer a benefit upon him. See, Crosby v. United States, 496 F.2d 1384 (5th Cir. 1974); Gibbs v. Tomlinson, 362 F.2d 394 (5th Cir. 1966). Section 61(a) of the Internal Revenue Code directs that a dividend shall be included in a taxpayer’s gross income. Section 316(a) indicates that a dividend encompasses “any distribution of property made by a corporation to its shareholders” from earnings and profits accumulated after 1913 or from earnings and profits of the taxable year. That the provision of services by a corporation is cognizable under the concept of “property” is assured by Regulation 1.317 — l. Simply stated, a taxpayer can be charged with disguised or constructive dividend income even though the corporation has not observed the formalities of dividend declaration, and has not made a pro rata distribution to the entire class of stockholders, and even though neither the corporation nor the shareholder intended a dividend and the corporation did not record the distribution as a dividend for bookkeeping purposes. Crosby, 496 F.2d at 1388. In the context of a closely held corporation, the potential presence of a constructive dividend is a perennial problem. Instances of ‘constructive’ or ‘disguised’ distributions are commonly encountered in the context of closely held corporations whose dealings with their stockholders are, more often than not, characterized by informality. Bittker & Eustice, Federal Income Taxation of. Corporations and Shareholders (3d ed.), 117.05, p. 7-24. In the ease sub judice, the district court determined that Loftin and Woodard received constructive dividends as a result of the land clearing operation undertaken by the Corporation on their behalf. The determination of a constructive dividend is a question of fact, reviewable by this court under the clear error criterion. Hardin v. United States, 461 F.2d 865 (5th Cir. 1972); Lengsfield v. Commissioner of Internal Revenue, 241 F.2d 508 (5th Cir. 1957). Under the clear error standard, we must decide whether the district court’s findings of fact are without substantial evidentiary support or are premised on that court’s misapprehension of the effect of the evidence before it. Rewis v. United States, 445 F.2d 1303, 1304 (5th Cir. 1971). See, Commissioner of Internal Revenue v. Duberstein, 363 U.S. 278, 291-92, 80 S.Ct. 1190, 4 L.Ed.2d 1218 (1960). As the Gibbs court noted, “[t]he determination of when the obligation is no longer owing or whether it was to be paid is, of course, a question of fact to be resolved by the trial judge. Such findings will not be set aside unless they are without substantial evidence to support them, or the court misapprehended the effect of the evidence, or whether on reviewing the entire evidence the court is left with the definite and firm conviction that a mistake has been committed.” 362 F.2d at 397. This is the standard by which we are guided in our review of the factual determination of whether an economic benefit was provided without expectation of repayment. A. THE FAILURE TO DETERMINE THE EXISTENCE OF A CONSTRUCTIVE DIVIDEND We begin our review of this finding by noting that the determination of a constructive dividend entails two separate inquiries. First, the trial court must determine whether a constructive dividend was conferred. Then, and only then, need the trial court focus on a computation of the amount of such a dividend. United Aniline Company v. Commissioner of Internal Revenue, 316 F.2d 701 (1st Cir. 1963); Challenge Manufacturing Co. v. Commissioner, 37 T.C. 650 (1962). While these determinations necessarily are entwined to a degree, it is impermissible to find the presence of a constructive dividend simply because the value of a benefit conferred exceeds the payments made in compensation. One of the underlying problems with the district court’s treatment of the constructive dividend issue in the present case is its apparent failure to demonstrate that it performed both segments of this two part operation. “Not every corporate expenditure incidentally conferring economic benefit on a shareholder is a constructive dividend.” Crosby, 496 F.2d at 1388. Before determining the amount of a constructive dividend, the court must determine whether “the distribution was primarily for shareholder benefit.” Sammons v. Commissioner of Internal Revenue, 472 F.2d 449 (5th Cir. 1972). Unfortunately, the line between “primarily for shareholder benefit” and “primarily for corporate benefit” is one not susceptible to easy delineation. Crosby, 496 F.2d at 1389. An example of the danger attendant to a failure to undertake both parts of this inquiry is provided by Gibbs, supra. There, the court dissected a transaction in which stockholders leased their property to their corporation. A portion of the lease required the corporation, at its own expense, to make certain improvements on the property. The corporation was to be paid from the rental payments due under another lease in which an unrelated corporation was the lessee. The court began by explicitly splitting the determination into two parts: the existence of a dividend and secondly, its amount. As to the first determination, the Gibbs court noted that [t]he crucial consideration with regard to whether a dividend resulted to taxpayer is whether it was intended that taxpayer was to pay the corporation for the work. Gibbs, 362 F.2d at 397. In accordance with this principle, the court decided that the transaction at issue embodied a constructive dividend since stockholder’s corporation only would receive payments for the construction work done if the unrelated corporation did not default on its lease payments. Because the linkage of lease payments with payment for improvements was not placed in the improvements contract until after the unrelated corporation had defaulted, the court found that there was no expectation of repayment sufficient to stave off a finding of a constructive dividend. Another example of the fundamental importance of the initial determination may be found in Benes v. Commissioner, 42 T.C. 358 (1964). Confronted with a stockholder whose corporation had undertaken a series of improvement on a piece of property eventually serving as his residence, the court first decided that these expenditures were primarily for the stockholder’s benefit. This presented a close question in that the property, at the time the improvements were made, ostensibly was not being utilized for his purposes. Had the court merely assumed the existence of a constructive dividend because of a difference between the value of the improvements made and the payments rendered by the stockholder on a piece of property connected with the stoekholder’s name, it would have overlooked the possibility that the improvements were not conferred for his benefit. Finally, in Crosby, the court dissected the transactions which allegedly served as the basis for constructive dividend treatment from the required two part perspective. Here, the corporation made improvements on its stockholder’s land after the stockholder had conveyed the land by deed to the corporation. In pursuit of the primary determination that must be made in a constructive dividend suit — whether a benefit was conferred primarily for the stockholder’s benefit — the court discovered that the only property right which the corporation had accrued through the transfer of the deed was the limited right to sell the property. An investigation of the substance of this transaction convinced the court that the underlying purpose of the improvements made was the provision of a stockholder benefit. Had the Crosby court merely accepted this transaction at face value, it would have overlooked the existence of an economic benefit conferred by the corporation without expectation of repayment and, as a result, overlooked a potential constructive dividend. It is clear from Benes, Gibbs and Crosby that the comparison of the value of the benefit conferred and the payments made in reimbursement cannot serve as the predicate for the finding of a constructive dividend. A prerequisite to the measurement of a dividend is the determination of its existence. The importance of this prerequisite is underscored by the apparent failure of the district court in the present case to have satisfied it. The court measured and determined the dividend simultaneously. It proceeded on a year by year basis, comparing payments made with costs incurred. Leaving aside the question of whether this was the proper method for computing the dividend herein, it is clear that such an approach will elicit the presence of a dividend whenever payment is made in a year other than the one in which the costs relating thereto are incurred. In a contract requiring the rendering of services over a period of several years, the use of this method of measurement as the means by which the court determines the preliminary issue of whether a benefit was conferred primarily for stockholder benefit without expectation of repayment can produce an ill-conceived result. In the case sub judice, the lower court reasoned that the absence of a payment (or sufficient payments) by the partnership in a particular year signaled that the corporation did not expect repayment. As a result, a constructive dividend was assessed. This is the reasoning which appears to have been the predicate, albeit a sub silentio predicate, for the district court’s determination of a dividend. We believe that this reasoning places the cart before the horse. By reasoning in this manner, the district court was measuring the dividend prior to determining that one had been conferred. The district court stated that it operated upon the premise that a “constructive dividend occurs when a corporation confers an economic benefit on its shareholders without expectation of repayment.” App. 180-181. On its face, this would appear to satisfy the obligation of the district court with regard to the primary determination entailed in a constructive dividend suit. However, it is clear from an examination of the record that this statement was made only in relation to the numerous cash withdrawals made by stockholders Loftin and Woodard — not in relation to the land clearing operation. As to this operation, which constituted the major part of the constructive dividend finding, the court was silent on the question of whether it performed the required first-part determination. And the classification of the cash withdrawals as constructive dividends — the subject of the court’s statement — is not before us on this appeal. Thus, in contrast to the clarity with which the court approached the dividend issue in relation to the cash withdrawals is the treatment it accorded to the dividend issue in relation to the land clearing operation. It does not appear from the record that the district court ever determined that the land clearing operation was a “distribution” conferred primarily for shareholder benefit. Instead, it appears that the district court assumed that in any year when the payments made by the partnership did not match the costs incurred by the corporation, a distribution was made primarily for the partners’ benefit. The danger of not maintaining a separation between the two parts of the constructive dividend determination cannot be overstated. It might be argued that the non-expectation of repayment was demonstrated with such clarity that there was no need for an explicit finding in this regard. While such an argument might succeed in a limited number of cases, we are not pointed to any facts which support such an approach here. For example, the initial spreading of costs by the Corporation among its other projects and the failure to make a payment until December 31,1964 are cited as significant in this regard. In addition, the Government emphasizes the failure of the stockholders to be guided by the advice of their own accountant with regard to the constructive dividend that would arise from the benefits accruing from the land clearing operation. As to the use of the “spreading of costs” as a fact indicating that the district court made the required first-part determination, this does not demonstrate that the Corporation did not expect repayment. Testimony at trial indicated that the initial spreading of the costs incurred at Yazoo among the general expenses of a corporation the size of this corporation was not highly unusual. We cannot disregard the fact that the Corporation did begin expensing the land clearing costs to a separately identifiable category once the project was in full swing. Indeed, this change occurred prior to the time that the IRS began its initial audit. Thus, the change was not undertaken in an effort to cover corporate tracks. As to tHe advice rendered by the Corporation’s accountant, Savage, it is true that he advised the stockholders not to authorize the Corporation to clear the land prior to the transfer of that land to the partnership because of the possibility of a constructive dividend arising therefrom. However, Savage was not concerned with the type of constructive dividend that arises when the value of the services rendered exceed the payments made for such services — that is, the constructive dividend at issue here. Rather, Savage was concerned over the possibility that the corporate clearing efforts would enhance the value of the property so much that when the Corporation transferred the property to the partnership, the incremental appreciation of the land’s value would be fully taxable at that time. In sum, it is unclear whether the district court determined that a constructive dividend was conferred prior to the court’s measurement of such a dividend. Given the guidance of Benes, Gibbs, and Crosby, this is a determination which must be made. The district court is instructed to make this determination consistent with this opinion. While this would be sufficient, alone, to support a remand for reconsideration of the deficiencies assessed against taxpayers on the basis of the Yazoo clearing, we believe that there are other problematic aspects of the district court’s approach to this issue. In the interest of-justice and judicial efficiency, we discuss them now. B. THE EQUATION OF VALUE WITH COSTS The second aspect of the district court’s treatment of the constructive dividend issue which we find troublesome centers on its theory of measurement. If we assume that a dividend is found to exist — the subject of discussion and remand in the preceding section — the district court must then compute the amount of the benefit conferred. Section 301 of the Code directs that the amount of a distribution of property to a noncorporate shareholder shall be measured as follows: “the amount of the money received, plus the fair market value of other property received.” (emphasis added). Thus, in determining the amount of taxable income embodied within a constructive dividend the court must diminish the value of the benefit conferred by the amount of any payment made by the shareholder in return. The sticking point, for present purposes, is the concept of fair market value. The district court measured the value of the benefit conferred by computing the costs incurred by the corporation in supplying the benefit. Leaving aside, for a moment, the types of expenses included by the district court in the computation of costs, we first focus on the correctness of the equation of value and cost. There is ample authority to support the use of costs as an equivalent of “value” for the purposes of a constructive dividend determination. See, for example, Walker v. Commissioner of Internal Revenue, 362 F.2d 140 (7th Cir. 1966), cert. den. 385 U.S. 865, 87 S.Ct. 124, 17 L.Ed.2d 92 (1966). However, a careful reading of this line of authority raises grave questions concerning the propriety of such an action in the present case. One of the primary motivations behind the district court’s use of costs as an equivalent of value was its apparent belief that the Corporation would not have undertaken the land clearing project unless such a project was profitable. In essence, this position provides that had the district court utilized the fair market value established by the taxpayers’ witnesses, such a value, multiplied by the number of acres cleared, would have been less than the costs incurred by the Corporation, as those costs were computed by the district court. There are two serious problems with the district court’s rationale. We agree that corporations are, for the most part, governed by the profit motive in their choice of undertakings. However, in the land clearing business, the price of a project is determined prior to the accrual of expenses. That is, in the land clearing business, a bid is made on a project on the basis of projected expenses, long before those expenses ultimately are incurred. Notwithstanding the fact that in the case sub judice, we are confronted with a closely held corporation, the structure of the land clearing industry suggests that a project might, upon completion, yield less revenue than initially had been anticipated. In fact, testimony indicated that this was not a unique occurrence. This might be particularly true in a project of several years duration. Secondly, the district court decided the project was unprofitable because it compared the fair market value figures testified to at trial with direct and indirect costs. Included in the category of indirect costs were a pro rata share of overhead and of the depreciation attributable to the machines involved in the Yazoo project. Without commenting on whether indirect costs belong in the computation of a constructive dividend, we suggest that their inclusion in the determination of the profitability of this operation may have produced a skewed resulting interpretation of profitability. The Yazoo project was particularly large. The Corporation’s indirect costs essentially were fixed. That is, the Corporation would incur certain costs whether or not the Ya-zoo project was undertaken. Salaries and depreciation would not be altered by the presence or absence of a further project unless the project was so large that it required the Corporation to hire further personnel and/or machines. Thus, the decision to undertake the Yazoo project for payments which exceeded the direct costs incurred, although less than the sum of indirect and direct costs, might have been premised on a concept of marginal profitability. Although the concept of marginal profitability necessarily entails the presumption that there were no other jobs of a nature similar to the Yazoo project but of greater profitability, we offer the concept as an indication of the potential inaccuracy of the district court’s premise. Given the above two criticisms of the district court’s premise, we suggest that its assumptions about profitability, without more, would not support an otherwise unsupportable decision to utilize costs as the measurement of the fair market value of a benefit conferred. By the above discussion, we do not intend to require a district court to undertake, in every case, the complex albeit elusive, determinations alluded to above. However, when a court assumes that a corporation was involved in an unprofitable venture without such further inquiry, that assumption may be incorrect. Without such further inquiry here, it was an inappropriate premise upon which to structure the treatment of the transaction in. the present case because this untested premise was the basis for the trial court’s decision to use costs as a measure of value — an equation which is the exception rather than the norm. We believe that the cases which have utilized “costs” as a measurement of “fair market value” in the constructive dividend context are distinguishable. The cases supporting the use of costs stand for the proposition that cost is an acceptable measure of value when there is no credible evidence of fair market value or when evidence of fair market value exists but is clearly rebutted. We analyze them in detail because they shed much light on the inexplicable quality of the district court’s decision to disregard the expert testimony provided in the present case. As the court stated in United Aniline, supra, it is permissible to measure the value of the benefit conferred as a constructive dividend through a computation of costs “in the absence of any direct evidence of fair value.” 316 F.2d at 705. And, in Walker, supra, the court justified the use of costs where there was sufficient expert testimony rebutting the taxpayer’s assertions of fair market value. 362 F.2d at 143. The other cases cited for their use of costs as a measure of value similarly are distinguishable. First, in Goldstein v. Commissioner of Internal Revenue, 298 F.2d 562 (9th Cir. 1962), the evidence of fair market value introduced by taxpayer was exceedingly weak. The selling price of the property at issue was in excess of even the most persuasive evidence of fair market value. Although there was some expert testimony supporting the stockholder’s fair market valuation, its validity was seriously undercut by the circumstances of the underlying transaction. The stockholder, himself, had paid far less for the property only three weeks prior to the sale to the corporation. In addition, the corporation, prior to the purchase, was leasing the land under a sale-leaseback arrangement which, in toto, would invest eventual ownership in the corporation at a price far below the price it decided to pay the stockholder. Finally, and of equal significance, the stockholder’s experts did not demonstrate that their opinions were predicated on the existence of the above recited circumstance. In essence, “appellant neglect[ed] to explain the lack of any direct evidence that either of these valuations took the effect of the lease into consideration.” 298 F.2d at 567. Second, the Tax Court case of Lester E. Dellinger, 32 T.C. 1178 (1959), further strengthens our suspicions concerning the district court’s treatment of valuation in the case sub judice. In Dellinger, costs were utilized because fair market value did not accurately reflect value in light of the fact that the shareholder himself sold the property in question for an amount greater than the value suggested by his evidence as to fair market value. 32 T.C. at 1183,1184. Moreover, taxpayer’s evidence of fair market value consisted of “only the vague testimony of the petitioner himself. Not only is it clear that he had no qualifications to express an acceptable opinion as to such values, but, in addition, his testimony was inherently improbable and unreasonable.” 32 T.C. at 1185-6. Compounding these problems from the court’s perspective was the failure of taxpayer to offer, as a witness in support of his valuation, the real estate appraiser best qualified to testify as to the value of the land in issue. 32 T.C. at 1186. Clearly, taxpayer had not met his burden of proof. See, also, Honigman v. Commissioner of Internal Revenue, 466 F.2d 69 (6th Cir. 1972). Finally, perhaps the classic example of the relationship of costs and valuation is found in Commissioner of Internal Revenue v. Riss, 374 F.2d 161 (8th Cir. 1967). There, the court was confronted with a constructive dividend comprised of a residence occupied by the stockholder’s family; insurance and operating costs of cars utilized by stockholder’s family; club bills; and a trip to Puerto Rico. As to the residence, the court, in opposition to the Tax Court’s decision in Riss, computed its value on the basis of the expenses and depreciation related thereto— that is, costs. Significantly, evidence of the fair market value of the residence was not provided to the Tax Court. Compounding this problem was the fact that a qualified real estate expert testified that the high cost of maintaining this property made it difficult to rent at any price. Further, rentals of comparable residences indicated that the value chosen by the Tax Court was wholly inconsistent with the little evidence of fair rental value that could be ascertained. In short, the Tax Court’s fair rental figure was decidedly unsupported by all credible evidence and that evidence of fair market value which did relate to this residence, albeit tangentially, supported a figure arrived at on the basis of cost. Thus, the Riss court determined that cost was a more reliable measurement of value when there was an absence of direct evidence of fair market value and a presence of evidence suggesting the inappropriateness of fair market value as the mode of measurement. This interpretation was borne out by the Riss court’s treatment of the benefit embodied in the personal car provided for the stockholder. Because evidence of fair market value was introduced as to this item and because “no evidence was introduced with respect to the total operating cost of the car,” the court’s decision to utilize fair market value instead of costs was entirely consistent with its overall approach to valuation in that case. 374 F.2d at 170 (emphasis added). It is noteworthy that the favored place of “fair market value” in the hierarchy of modes of valuation is bolstered by the Code itself. Regulation l-301.1(j) delineates the concept of a constructive dividend in terms of “fair market value”. Likewise, I.R.C. § 301(b)(1)(A) speaks of distributions of property to non-corporate distributees in terms of the fair market value of the property received. The equation of value with costs is, therefore, not the norm but the exception. With this in mind, we focus on the present case. In determining the current value of the economic benefit conferred, it is “clear that the taxpayer never loses the burden of proving the Commissioner’s determination erroneous.” Crosby, 496 F.2d at 1390. Loftin and Woodard do not argue for the adoption of a contrary proposition. Instead, appellants suggest that the evidence they introduced in support of the valuation of $50 per acre was sufficiently strong that the district court’s failure to utilize such a valuation can only be explained by a disregard for the appropriate legal standard — a standard favoring the use of fair market value over the use of cost for purposes of valuation. The evidence relating to value which was before the district court consisted of the testimony of three “expert” witnesses, all provided by the taxpayers. As a preliminary matter, we note that while the weight attributable to the opinions of these witnesses is hotly contested, the fact that they are experts is not. The district court permitted these witnesses to offer their opinions in the land clearing field after an objection to their qualifications to render such testimony. App. 483-85. While this action did not determine the weight which should attach to their testimony nor its credibility, it did establish that their testimony was admissible as the testimony of experts in the relevant area. F.R.Ev. 702; 3 Weinstein’s Evidence, 1702[01]. The three witnesses who testified at trial — Bill Moore, W. G. Boykin and W. B. Holloway — all rendered opinions as to the fair market value of the land clearing operation. No evidence was offered by the Government in rebuttal. Instead, it relied on cross-examination. The first witness, Moore, testified that he was familiar with the land in question before and after the clearing operation was completed. Moore was himself a land clearer whose operations were proximate to the Yazoo clearing in time and geography. His typical clearing operation consisted of shearing, piling, burning, cleaning and disc-ing. He explicitly indicated that stumps would not be removed in a clearing effort; instead, they would be cut down to the surface. When asked to estimate the fair market value of the work performed by the Corporation, he emphatically stated that $50 per acre would be the maximum compensation a land clearer could and would expect to receive for the Yazoo clearing. According to Moore, the economic environment surrounding the land clearing business in the years in question was intensely competitive. Thus, a typical clearer would be delighted to receive such a price for the efforts required at Yazoo at that time. The second witness, Boykin, was involved primarily in the land clearing business also in the period and geographic area present in this case. Like Moore, Boykin’s clearing experience included cutting, piling, sweeping, burning and ditching. He too emphasized that stumps were not removed. Although he noted that the performance of ditching and discing at a site was dependent on the party’s wishes, as reflected in a contract, he also clearly indicated that his fair market value estimate was predicated on an understanding of the condition of the land prior to and subsequent to the clearing operation. To infer that he may have taken into account any discing and ditching done would not be unwarranted. There is no evidence to suggest otherwise. Like Moore, Boykin placed the maximum value for such clearing at $50 per acre, noting, as had Moore, that the cost per acre would only decrease as the number of acres to be cleared increased. Taxpayers’ final witness was Holloway, the accountant whose clientele included several substantial land clearing entities that had cleared land near the Yazoo project during the years in question. When asked about his familiarity with the work actually performed on the Yazoo site, Holloway testified that he had seen the property and that the clearing performed amounted to a “normal” operation. Once again, he considered $50 per acre to be top dollar for that operation. Indeed, he expressly indicated that the only clearing jobs which would command a better price would be those involving a very small amount of acreage. We offer this synopsis of the expert testimony not in an effort to assess their credibility — a job exclusively within the province of the district court. Rather, we offer it because our review of its content leaves us puzzled as to the rationale underlying the district court’s decision to utilize costs as a measure of value instead of the fair market value. The Government did not introduce a shred of affirmative evidence to contradict the above testimony. Cross-examination revealed only that these “experts” were not wholly familiar with the precise work done in the clearing operation. While this could have been important, we believe that its edge was blunted by testimony in the record which suggested that the operation was not unusual in nature — that unrebutted testimony being provided by the above witnesses and the taxpayers themselves. Moreover, the Government did not introduce evidence suggesting that the operation was unusual. Indeed, two of the three witnesses had done a fair amount of clearing themselves in and around Yazoo and, therefore, were familiar with the types of operations performed in the course of a clearing project in that area. The court’s speculation that tree stumps may have been removed, thereby increasing the cost per acre, is without any support in the record. Significantly, the testimony of the three witnesses was that it would be highly unusual for such an action to be undertaken. It is difficult to understand the motivation of the Government and the court in suggesting that “something” special was done on the Yazoo property. Just because a closely held corporation was performing work for its stockholders does not mean, ipso facto, that the work was customized. The record indicates that the only work performed of an unusual nature was that which was done prior to the actual clearing. In essence, this work consisted of the repair of cabins already in existence on the land and the establishment of a better drainage network. The outside figure for this work was set at $22,000. This hardly would be a decisive piece of work in a project which cost, under the Government’s own estimate, $500,000. Furthermore, Woodard’s own testimony indicated that he wanted the corporation to clear the land “fairly well” so that he could sell it. This statement, which is asserted by the Government in support of its belief that special work was done on this land of which the expert witnesses were not aware, is equally susceptible to a contrary interpretation. This was not a piece of property which the buyers were seeking to have custom crafted for their own personal enjoyment. It was, instead, a piece of property which they eventually would seek to sell. There were other criticisms leveled at the “deficiency” of the expert testimony in this matter. It is true, as the Government contends, that none of the witnesses bid on this land. However, no one could bid on this land because it was never opened to competitive bidding. Indeed, if the Government’s criticism was accepted, no witness could have qualified as an expert in this matter. It is also true, as the Government asserted, that none of these witnesses had cleared property of such grand proportions. However, the applicability of this criticism also is highly suspect given the uncontradicted testimony in the record compelling the conclusion that the costs of clearing only decrease as the amount of acreage grows larger. Finally, perhaps the most incomprehensible criticism leveled by the Government at the expert testimony is that it was all irrelevant because the existence of a contract between the taxpayers and their corporation prior to 1965 was never proven. The testimony as to value was offered in order to determine the value of the benefit conferred, not the accuracy of any price agreed upon by the parties. A valuation had to be performed for constructive dividend purposes, irrespective of the existence of a contract. Indeed, the existence of a contract setting the price at $50 per acre— the figure taxpayers seek to establish— would not have been conclusive even on taxpayers’ behalf. At the same time that the Government was criticizing the taxpayers’ offerings, it inadequately explained its own shortcomings in this area. When questioned by taxpayers’ attorney as to its failure to introduce rebuttal evidence on the issue of value, the Government Agent responded that to have taken the fair market value, we would have had to call in an engineer agent and a forestry agent and have them come in and value it. We decided not to do it. App. 1051-52. Are we to conclude that the only reason the Government did not attempt to measure the value of the benefit conferred in fair market value terms was that it was inexpedient to do so? Surely, this would not pass as an ample justification for the court’s ultimate determination on the valuation issue given that taxpayers have offered substantial testimony on this issue. In sum, value is the overriding concept in the measurement of a constructive dividend. While the vagaries of any competitive situation often render elusive the measurement of value, there was a substantial amount of unshaken expert testimony produced at trial by the taxpayers. The testimony offered by the taxpayers produced a factual context far different from those cases in which courts correctly chose to equate value with costs. Taxpayers’ witnesses in the case sub judice were fairly knowledgeable with regard to the land clearing business and the actual land in question, unlike the witnesses offered in cases like Goldstein, supra. The district court’s decision to utilize costs to measure value in light of the presence of such testimony and the Government’s failure to contradict that testimony, suggests that the lower court was operating either under an improper construction of the pertinent legal standard or under a belief that the witnesses were not credible. Taxpayers had the burden of proof on the value issue. However, unlike United Aniline or Walker, supra, taxpayers did act to meet that burden. Whether the district court traversed the incorrect legal path or incorrectly traversed the correct legal path is a matter which is not susceptible to determination by this court at this time. Therefore, we direct the district court to reconsider the manner of measuring value that it employed should it decide first, on remand, that an economic benefit was conferred on a shareholder in his capacity as such without expectation of repayment. C. THE USE OF AN ANNUAL ACCOUNTING PERIOD The final problematic aspect of the district court’s treatment of the constructive dividend issue centers on that court’s decision to utilize a calendar year period for the measurement of the amount of the constructive dividend. That is, in reaching a figure for the value of the constructive dividend arising from the clearing operation, the lower court compared the costs incurred by the Corporation in a given calendar year with the payments received from the partnership in that same calendar year. There are two time frames within which the district court could have measured the constructive dividend, assuming one was properly found to exist — the time frame of annual units (fiscal or calendar) or the time frame beginning with the commencement of the project and ending with the project’s termination. It appears, from the district court’s findings, that it assumed that the only proper manner of computation was the calendar year basis. In support, the Government asserts that the use of any period but an annual period would cut against the notion of annual accounting ingrained in our tax laws. We disagree. 26 U.S.C. § 446 outlines the permissible methods of accounting under the tax law. One of the methods recognized as acceptable is the completed contract approach. When a taxpayer operates under this approach, he reports taxable income from a project upon its completion. Daley v. United States, 243 F.2d 466 (9th Cir. 1957), cert. den. 355 U.S. 832, 78 S.Ct. 46, 2 L.Ed.2d 44, Reg. 1.451-3; Reg. 118, § 39.42-4; and Reg. 111, § 29.42-4. The concept of the completed contract method of accounting was alluded to as far back as Burnet v. Sanford, 282 U.S. 359, 51 S.Ct. 150, 75 L.Ed. 383 (1931). While this method is recognized as an acceptable mode of accounting for tax purposes, it cannot be utilized capriciously by a taxpayer. A corporation must request permission from the Commissioner in order to utilize it. See, Reg. 1.451-3(e); Wright Contracting Co. v. Commissioner of Internal Revenue, 316 F.2d 249 (5th Cir. 1963), cert. den. 375 U.S. 879, 84 S.Ct. 147, 11 L.Ed.2d 110. Because taxpayers did not seek such consent here, they cannot argue that this method must be used to evaluate their clearing operation in terms of its tax consequences. However, the Commissioner is invested with the discretion to utilize such a method in rendering his computation of the taxpayers’ tax liability. 26 U.S.C. § 446(b). Under 26 U.S.C. § 446(b), the Commissioner is directed, in auditing a taxpayer’s return, to use the method chosen by the taxpayer to measure income. But, if the method used does not clearly reflect income, the computation of taxable income shall be made under such method as, in the opinion of the Secretary or his delegate, does clearly reflect income. 26 U.S.C. § 446(b). See, Commissioner of Internal Revenue v. Hansen, 360 U.S. 446, 79 S.Ct. 1270, 3 L.Ed.2d 1360 (1959); Burck v. Commissioner of Internal Revenue, 533 F.2d 768 (2d Cir. 1976); Stephens Marine, Inc. v. Commissioner of Internal Revenue, 430 F.2d 679 (9th Cir. 1970). Given the discretion with which the Commissioner is invested to choose a method of measurement most capable of accurately reflecting a transaction, blind adherence to the accounting method chosen by the taxpayers or the corporation performing the work would not be consistent with the overriding spirit of the law. “Consistency alone in the accounting method used by a taxpayer cannot satisfy the requirement that there be a clear reflection of income.” Photo Sonics Inc. v. Commissioner of Internal Revenue, 357 F.2d 656 (9th Cir. 1966). The discretion with which the Commissioner is invested may not be abused. Whitaker v. Commissioner of Internal Revenue, 259 F.2d 379 (5th Cir. 1968). Underlying the entire fabric of the tax structure is a desire to accurately portray the events that have transpired. Indeed, even a combination of accounting methods is permitted under the tax law if such a hybrid approach more clearly reflects income. Reg. 1.446-l(c)l(iv). While abuse of this discretion by the Commissioner must be proven by a clear showing, Wood v. Commissioner of Internal Revenue, 245 F.2d 888 (5th Cir. 1957), it would seem that such a showing could be predicated upon a decision to use an accounting method that is inaccurate under the circumstances. While we express no opinion as to the manner in which the district court ultimately should resolve this issue, we do note that in the present case, the measurement of the dividend on an annual basis appears to produce a patently unjust result. If most purchasers of land clearing services do not pay for these services until the project is completed, then the use of an annual accounting method to measure a constructive dividend arising from such services would work a severe hardship on the beneficiary of a clearing operation that spanned several years. The unjust nature of this choice of measurement method is underscored by the fact that it is the Corporation that chooses the date on which to declare its costs. If it has chosen the completed contract method, then the taxpayer is benefitted. But if it has not, then costs will be registered on an annual basis, in all likelihood, and payments will not be registered until the completion of the operation, years later. Given that the excess payment in one year is not transferable to offset a prior year’s costs, the taxpayer will not be able to recover from the corporation’s choice. There is much authority supporting an approach which breaks free from the moorings of annual accounting in the present case. That authority prescribes that the lodestar by which the court must be guided in its treatment of the measurement issue is the pursuit of substance over form. “The basic tenet of tax requires ... a substance-over-form practical construction —‘it is the underlying essence of a transaction that determines its taxability.’ ” Crosby, 496 F.2d at 1389 (cit. omit.). An application of this tenet may be found in Gibbs, supra. Gibbs provides an example of the pursuit of accuracy that reached fruition in a cumulative approach to the measurement issue. In this case, Judge Bell emphasized that the proper manner in which to measure the constructive dividend arising from property improvements was to focus on “the sale price of the alterations and improvements less payments made, if any.” 362 F.2d at 398. This formula, which rests on a cumulative approach to measurement, was tailored to the factual context confronting the Gibbs court — one in which the subject of the constructive dividend was leasehold improvements made over a period exceeding one year. It is, in short, a formula premised on accuracy; a formula which takes cognizance of a series of events which are not neatly confined to an annual span. An application of this tenet which produced a different result precisely because the factual circumstances were different may be found in Benes v. Commissioner, 42 T.C. 358, aff’d 355 F.2d 929 (6th Cir. 1966). In Benes, the focus on substance-over-form supported the measurement of a constructive dividend on an annual basis because the court found that the stockholder had no intention of ever paying his corporation for the cost of the improvements rendered by it on property found to have been built primarily for his benefit. Thus, the use of an annual accounting period in Benes was wholly consistent with the overriding judicial desire to choose a method of measurement most capable of accurately depicting the essence of the underlying transaction. A final noteworthy example of the continuing effort to procure a more accurate treatment of the measurement question is provided by the court’s management of the constructive dividend issue in Honigman, supra. In Honigman, a piece of property was sold to stockholder’s corporation by a corporation in which stockholder possessed a 35% interest. The court found a constructive dividend and measured it as the excess of the property’s fair market value over the purchase price, where the purchase price was calculated as a lump sum figure. This lump sum characteristic is particularly important because the payments comprising it were not made within one year’s time. The purchase was consummated via the assumption of a mortgage on the property; the assumption of a $21,000 tax liability; and the payment of $50,000 cash. Significantly, the fact that the taxpayer was not following a completed contract method of accounting did not prevent the court from explicitly recognizing the piecemeal quality of the transaction and from measuring the “payment” component of the constructive dividend in a cumulative manner. We urge the district court to reconsider the propriety of utilizing an annual accounting method in a situation where the parties do not intend to match costs and payments on an annual basis. In a judicial determination of tax liability, the value of consistency cannot be disregarded, but the value of accuracy must be foremost. “Whatever method is used must be followed in reporting and determining taxable income, unless some adjustment of the method used is required in the particular case in order to clearly reflect income.” 2 Mertens, Law of Federal Income Taxation (1974 rev.) § 12.02. In this regard, we cannot disregard the advice of an eminent tax authority who has unequivocally indicated that in the reporting of income from a long-term contract, the completed contract method is essential. Id., § 12-130-135. The court has discretion to utilize that method of accounting most apt to render an accurate narration of the tax events that transpired. The completed contract method, or a comparable cumulative approach, is best suited for a transaction spanning several years where payment is not expected concurrent with the accrual of costs and the constructive receipt of income by the taxpayer is established on the basis of the accounting scheme chosen by the corporate entity. Notwithstanding the failure of the parties to have utilized such a method in reflecting this transaction on paper, the district court must consider whether the determination of a deficiency in the context sub judice mandated an examination from a non-annual perspective. D. THE STIPULATION Assuming, on remand, that the district court finds that a constructive dividend was conferred; determines the proper measurement for value; and determines the proper interval in which to calculate the amount of the constructive dividend, one major problem with the district court’s approach still remains. This problem centers on the decision of the lower court to disregard, sub silentio, a stipulation entered into by the taxpayers and the Government. That stipulation stated that: Loftin and Woodard, Partnership, made certain payments to Loftin and Woodard, Inc. for clearing land. These payments are reflected by entries on the partnership books crediting accounts receivable from Loftin and Woodard, Inc. . The following entries reflect the payments made: DATE L & W Partnership *** (Credit Accts. Rec. L. & W. Inc.) 12/31/64 $250,000 *** We quote the above stipulation at length because we deem the language susceptible to no other interpretation than that which appears on its face. While we do not quarrel with the district court’s right to disregard any stipulation, we do not understand its decision to do so in the present case — especially in light of the court’s failure to explain its action. This stipulation is pivotal in the constructive dividend determination because, as noted above, the district court measured the constructive dividend by comparing costs incurred and payments received within a calendar year period. The calendar year was utilized rather than the fiscal year because the shareholders were the recipients of the dividend and they were on a calendar year accounting program. As to calendar 1964, the year to which the stipulation pertained, the district court approved a major deficiency, assessed against the taxpayers with regard to the land clearing operation, on the basis of a comparison of costs incurred by the Corporation and payments received in compensation in that year. Had the court followed the stipulation, the difference between payments made and costs incurred would have been minimal in that year, if not non-existent. The justification tendered by the Government for the court’s decision to disregard the stipulation is that the entries made on the Corporation’s books were inherently more reliable than those made on the partnership books. Therefore, the Government argues that the court was justified in marking the date of the first payment by referring to the date the Corporation used to enter income received rather than by referring to the partnership entry relating to the date on which a payment was made. We do not believe this is a tenable explanation. No showing was made to the district court of the “manifest injustice” which would have resulted had the stipulation been followed. Henry v. Commissioner of Internal Revenue, 362 F.2d 640 (5th Cir. 1966). Yet “stipulations entered into freely and fairly shall not be set aside except to avoid manifest injustice.” Henry, 362 F.2d at 640. Simply stated, the district court was entitled to disregard the stipulation if to accept it would have been manifestly unjust or if the evidence contrary to the stipulation was substantial. We disregard completely the Government’s argument that the terms of the stipulation can be read to support the position of the district court. To engage in such semantic soft shoe would serve to diminish the honored place accorded to stipulations reached by the solemn efforts of opposing parties. In addition, according to the testimony of the Corporation’s bookkeeper and accountant, the Corporation operated on a percentage of completion method in its recordation of income. In other words, it would make income entries on its books by tallying the number of acres cleared at the end of its fiscal year and then multiplying this by $50 per acre — the alleged contract price agreed upon by the partnership and the Corporation. The Government co