The principal question in this appeal from a decision of the Tax Court disallowing deductions for depreciation, interest, and payment of rent is whether the Tax Court's finding that the sale of a plant was a sham transaction that should be disregarded for federal income tax purposes was clearly erroneous, the standard applicable to such a finding, Thompson v. Commissioner, 631 F.2d 642, 646 (9th Cir.1980).

The case concerns the Bernstein family, which can for our purposes be divided into two groups: the brothers Ace and Marty, and everyone else. The taxpayers call the second group the "cousins" and we shall follow that terminology although Ace and Marty in fact are uncles rather than cousins of two of the other Bernsteins. Until 1966 the cousins, on the one hand, and Ace and Marty, on the other, had gone their separate ways. The cousins owned and operated Milbrew, Inc., which manufactured an animal food additive and a yeast used in the brewery business. By late 1966 most of Milbrew's manufacturing was being done on a toll (i.e., fee) basis by a plant in Juneau, Wisconsin, owned by a dairy cooperative. Milbrew's business was in very bad shape and in 1967 the cousins induced Ace, a successful businessman in an unrelated field, to take a hand in running the company. At the same time Northland Developers, a corporation wholly owned by Ace and Marty, bought the Juneau plant from the dairy cooperative for a shade under $525,000; and for the rest of 1967 and all of 1968, Milbrew paid Northland rent for the plant at an annual rate of about $100,000.

On January 1, 1969, the Bernstein family created a limited partnership called NVST Company. All the Bernsteins received shares of the partnership, and Ace and Marty the lion's share--40 percent. On the day of its creation NVST entered into a contract with Northland to buy the Juneau plant for $3 million. The contract provided that NVST would pay Northland interest at 4 percent, but no principal, for the first five years, $50,000 a year for the next five years, and $300,000 a year for the succeeding ten years; the remaining principal was due at the end of the 20 years. Between 1969 and 1973 NVST charged annual rent to Milbrew rising from $220,000 to $542,000, and charged off against this rental income annual depreciation rising from $219,000 to $307,000 and interest on its contract with Northland rising from $27,000 in 1970 to $58,000 in 1973. Milbrew's large rental expense for the Juneau plant virtually wiped out its taxable income, which rose from $1,500 in 1969 to $8,100 in 1973 on gross sales that rose from $1.5 million to $2.9 million. The Internal Revenue Service disallowed the depreciation and interest deductions claimed by NVST (and passed through to the partners in NVST, i.e., the Bernsteins), and disallowed Milbrew's annual deduction for rental expense to the extent it exceeded $120,000 ($150,000 for 1973), the Service's estimate of the plant's fair rental value. The premise for these adjustments is that Northland's sale of the Juneau plant to NVST in 1969 was a phony.

With Ace and Marty the sole owners of the seller of the plant (Northland) and major owners of the buyer (NVST), Ace the putative savior of the tenant (Milbrew), and everyone else in the picture a relative of Ace and Marty's, the sale was not an arm's length transaction, so there is no presumption that it reflected market values. The sale of the plant for almost six times as much as Northland had paid 20 months before is of course highly suspicious and the terms of the sale suggest that Northland was not serious about collecting the inflated purchase price. Also, even in 1969 4 percent was far below the interest rate on long-term, inherently risky loans, while the $50,000 due annually between the sixth and tenth years of the contract was less than 2 percent. The $300,000 a year due between the eleventh and twentieth years actually represented less than 10 percent of the principal, since not only would the principal balance not have been paid down at all by the eleventh year but it would have been increased by the interest arrears from the sixth through tenth years. If the stream of payments due on the contract, treating the $3 million principal as due in a lump at the end of the contract period, is discounted to present value at a 10 percent discount rate, the 1969 sale price was really only $1.7 million. Moreover, during the first five years of the contract, when NVST was nominally obligated to pay a total of $600,000, it actually paid only $137,000. This not only was far below any reasonable estimate of the interest that an arm's length installment seller would have charged for the $3 million that Northland in effect loaned to NVST by deferring receipt of principal, but it meant that NVST was in default from the beginning--a circumstance that seems not to have troubled Northland.

As soon as the deal was signed, NVST, Milbrew's new landlord for the Juneau plant, doubled the rent, and thereafter kept raising it without rhyme or apparent reason other than to minimize Milbrew's taxable income. What otherwise would have been taxable income to Milbrew was thus shifted to NVST, which was able to shield this rental income from tax by depreciating the Juneau plant on the basis of the $3 million that it had "paid" Northland for the plant. Northland also escaped having to pay substantial income tax on the sale, because NVST paid it little interest and no principal. So the tax on Milbrew's income was not merely shifted, but avoided altogether. The key to the scheme was the $3 million sale price: if the price had been the same as Northland had paid 20 months earlier NVST could not have taken large enough depreciation deductions to offset its income from the huge rentals that it charged Milbrew.

When persons who are not dealing with each other at arm's length enter into a transaction that gives them tremendous tax savings, the Internal Revenue Service is entitled to be suspicious of the genuineness of the transaction; and here the sale price, payment terms, subsequent payment history, and rental increases rightly reinforced the Service's suspicions. In addition, the parties treated the sale with extreme informality. It was not recorded, and Northland continued to represent the plant as being its own property, unencumbered by any sale contract. It is noteworthy, too, that the contract of sale originally provided for interest at the absurdly low rate of 1 percent and that Ace changed this to 4 percent, just before the contract was signed, without consulting any other Bernsteins and without changing the price or any other term in the contract--which is why the $50,000 payments required in the fifth through ninth years would not even cover the interest contractually due in those years.

The facts reviewed thus far make this case much like Narver v. Commissioner, 75 T.C. 53, 100-01 (1980), aff'd per curiam, 670 F.2d 855 (9th Cir.1982); but we have to consider two types of rebuttal evidence offered by the taxpayers. The first was that after Ace became the guiding light of Milbrew in 1967 it experienced a rapid improvement in its business prospects which enabled it to pay a higher rent and which therefore increased the fair market value of the Juneau plant. The second is that after 1973 (the last tax year in the case) NVST made very substantial payments on the contract--$1.1 million by the end of 1978, though all of this was interest--and that in May 1983, while this appeal was pending, a large corporation unrelated to any of the Bernsteins executed a letter of intent to purchase NVST's assets, consisting primarily of the Juneau plant, for $8.7 million.

The post-1973 evidence has little significance, and not only because nine years after the contract had been signed the purchaser had still not begun to repay principal. In having NVST make payments after 1973 the taxpayers may have been trying to shore up their case against the very substantial tax liability involved here (the deficiencies assessed by the IRS on the Bernsteins and on Milbrew exceed $800,000 for 1972 and 1973 alone). As for the expected sale of the plant in 1983, even if we can go outside the record before the Tax Court (more on this question later) and accept the proposition that the plant indeed is worth almost $9 million today, this does not show that it was worth at least $3 million 14 years ago. Inflation, improvements made to the plant, and unexpected increases in the demand for the plant's product could, singly or in combination, dramatically increase the plant's value over so long a period of time. The plant easily could have appreciated, because of inflation and other factors, at an annual rate of 10 percent; if so, and if the plant is worth $8.7 million today, it was worth only $2.3 million in 1969, and we shall see that it does these taxpayers little good to establish a fair market value of more than $525,000 but less than $3 million.

The evidence with regard to Milbrew's improving fortunes between 1967, when Northland bought the plant for $525,000, and 1969, when it sold it to NVST for $3 million, deserves more weight. The record shows that in 1967 the dairy business was in substantial decline in Wisconsin and that dairy processing plants such as the Juneau plant could be bought at their scrap value, between $400,000 and $600,000. True, the Juneau plant had a nondairy tenant, Milbrew, but Milbrew was, as we noted earlier, in a parlous state too, which is why it had called in Ace. The dairy cooperative that owned the plant presumably knew nothing of Ace's plans to revitalize Milbrew and was therefore happy to be able to sell the plant to Northland at a price approximately equal to its scrap value. By the end of 1968 things were looking up for Milbrew. Between 1967 and 1969 its gross sales almost doubled. Therefore, to determine the market value of the Juneau plant on January 1, 1969, we must imagine what the probable outcome of negotiations between NVST and the dairy cooperative, dealing at arm's length, would have been if NVST had tried to buy the plant from the cooperative on January 1, 1969, Northland never having entered the picture. On the one hand, NVST might have said it would not pay more than scrap value for the plant because it could buy an equivalent surplus dairy plant elsewhere in Wisconsin for scrap value and rent that to Milbrew. On the other hand, the dairy cooperative could have pointed out that Milbrew's business and the Juneau plant had made an exceptionally good match that had rescued Milbrew and put it on the road to prosperity, and that surely Milbrew would pay a hefty rental to continue enjoying the use of the plant; if so, the plant would be worth a lot to NVST, Milbrew's landlord if NVST bought the plant. The Service does not contend that the $100,000 rent that Northland had been charging Milbrew in 1968 was unrealistic and this implies that the plant must have been worth more than $525,000 then, for the Service argues that a landlord can expect to recover between 14 and 16 percent of the value of his property in the rental for it and this would imply that the fair market value of the plant was around $700,000 in 1969.

If the Juneau plant was the only one that Milbrew could have survived in, a proposition the record supports, this is a classic case of bilateral monopoly--a case where two persons (Milbrew and the owner of the Juneau plant) can, as a practical matter, deal only with each other. See, e.g., Chicago & North Western Transport Co. v. United States, 678 F.2d 665, 667-68 (7th Cir.1982). In such a setting, the price at which the parties will transact may be anywhere in the range bounded at the lower end by the minimum price that the seller is willing to accept and at the upper end by the maximum price that the buyer is willing to pay--a range that is broad because of the assumption that neither party can practicably turn elsewhere if his offer is refused. In this case the lower end of the range was about $400,000, the minimum scrap value for dairy plants in Wisconsin in this period. The upper end of the bargaining range was the maximum price that someone could pay who planned to rent the plant to Milbrew; this price would be limited by the rental payments that could be squeezed out of Milbrew without breaking the company. If Milbrew could pay $450,000 a year in rent, this would imply that the plant might be worth as much as $3 million to a purchaser, since at that price the purchaser would earn 15 percent a year on his investment.

On these assumptions there would be a bargaining range between $400,000 and $3 million for the Juneau plant, and the actual price might be anywhere in between. But this analysis cannot carry the day for the taxpayers in this case. To begin with, $3 million is too generous an estimate of the upper end of the bargaining range. In the first year of the sale, 1969, NVST charged Milbrew rent of $220,000, which at a 15 percent return on investment implies that the plant was worth less than $1.5 million. Moreover, Milbrew did not actually pay cash, but merely recorded the rental as a bookkeeping entry. This means that the nominal rental was more than Milbrew actually could pay, and hence that even $1.5 million is above the upper end of the bargaining range.

In any event, the dairy cooperative would not have dreamed of holding out for such a high price unless it had happened to discover by January 1, 1969, that Milbrew's fortunes were on the upswing, a fact Milbrew would try to conceal from it. Finally, a party to self-dealing--treating the Bernsteins as a single entity, as we must--will not be heard to argue that a deal should be valued on the basis of a hypothetical transaction the outcome of which is inherently indeterminate. Cf. Chicago & North Western Transport Co. v. United States, supra, 678 F.2d at 670; In re Valuation Proceedings, Etc., Regional Rail Reorganization, 445 F. Supp. 994, 1015 (Spec.Ct.R.R.R.A.1977) (Friendly, J.); Stigler, The Theory of Price 207 (3d ed. 1966). It is required to provide more tangible evidence of value. Although the Bernsteins did introduce evidence of the replacement cost of the Juneau plant, replacement cost is not a valid estimate of fair market value in a declining market, Fox River Paper Corp. v. United States, 165 F.2d 639 (7th Cir.1948), as the market for dairy plants in Wisconsin was in the 1960s. In such a market a buyer knows that he will not have to build a new plant if the sale falls through--that he can buy an old one at a distress price, which is what Northland had done in 1967--and armed with such knowledge no buyer would pay a price equal to replacement cost.

It is only a minor embarrassment to the Service that one of its own expert witnesses valued the Juneau plant at its replacement cost--and maybe no embarrassment, since the witness estimated that cost at only $1.6 million. It cannot carry the day for the Bernsteins merely to show that the plant was worth between what Northland had paid in 1967, $525,000, and what NVST purported to pay Northland in 1969, $3 million. NVST had to calculate depreciation on the basis of its cost, see 26 U.S.C. Secs. 167(g), 1011, 1012, and to establish its cost for the Juneau plant had to show that it had actually bought it. If the sale was a phony, the plant remained, in the contemplation of the law, in Northland's hands. The fair market value of the plant on January 1, 1969, is just one of the factors entering into a judgment of the genuineness of the sale. If that value were $3 million it would go some way to dispel the inference of sham created by the other suspicious circumstances that we have discussed, but on no reasonable reading of this record can the fair market value be pushed up that high. The unrealistic price in the contract, together with the payment terms and history, the family relationship, the presence of a strong tax-avoidance motive, and the informality of the arrangement persuade us that the Tax Court did not clearly err in concluding that the sale was a sham. And if the sale was a sham, the plant's fair market value in 1969 is immaterial. A taxpayer may not increase depreciation by revaluing the property being depreciated, Parsons v. United States, 227 F.2d 437, 439 (3d Cir.1955); he must use his original cost. Detroit Edison Co. v. Commissioner, 319 U.S. 98, 102, 63 S. Ct. 902, 904, 87 L. Ed. 1286 (1943). NVST had no original cost, because, as a matter of federal tax law, it never acquired the plant.

The fact that the plant may have been worth more than $525,000 in 1969 is relevant, however, to whether Milbrew's rental deductions were excessive, for 26 U.S.C. Sec. 162(a)(3) entitles Milbrew to a reasonable rental deduction whether NVST or Northland owned the plant. But obviously, in light of all we have said, the amount of rent that Milbrew agreed to pay NVST is no evidence of what a reasonable rent for the plant would have been. We have to imagine what rent Milbrew would have agreed in an arm's length transaction to pay. The higher the fair market value of the plant the greater that rent would have been. But the Service was sufficiently generous in its adjustments in allowing Milbrew a $150,000 deduction for rent in 1973, implying that the fair market value of the plant that year was about $1 million, a reasonable estimate. For 1972 the Service allowed a smaller deduction--$120,000, implying a fair market value of about $850,000, but this was still reasonable. (The previous years' rental deductions are not in issue in this appeal.) The higher rentals charged Milbrew by NVST in these years not only were unrealistic in relation to the fair rental value of the plant but exceeded Milbrew's ability to pay on a cash basis, and were properly disallowed. See Southeastern Canteen Co. v. Commissioner, 410 F.2d 615, 619 (6th Cir.1969); Baldwin Bros., Inc. v. Commissioner, 361 F.2d 668 (3d Cir.1966); S & S Meats, Inc. v. Commissioner, 38 T.C.M. (CCH) 36,036 (1979).

The taxpayers' remaining arguments are procedural. The Tax Court judge before whom the case was tried retired after the trial but before rendering his opinion. The taxpayers agreed that the case could be reassigned to another judge for decision on the record compiled before the first judge, and this was done. Having been willing to take their chances before the second judge they cannot complain because he decided the case against them. The first judge had made comments during the course of the trial that were very favorable to the taxpayers. The taxpayers were confident that his successor would be influenced by those comments and would decide for them. They were disappointed. Of course if he had decided for them they would be defending vigorously the procedure that was adopted. By consenting to the procedure they waived any objection.

Last, aware of our probable unwillingness to consider evidence that was not before the Tax Court the taxpayers ask that we order that court to give them a new trial on the basis of newly discovered evidence, namely the 1983 letter of intent to buy NVST's assets, including the Juneau plant, for $8.7 million. It would have to be an extraordinary case to warrant our ordering the Tax Court to set aside its decision and receive new evidence. Normally we would go ahead and decide the case and leave it to the Tax Court to decide afterward whether to reopen it. In any event, as we explained earlier, the newly discovered evidence has little probative value and cannot justify the extraordinary procedural relief that the taxpayers are asking from us.

AFFIRMED.