Taxpayers, husband and wife, seek to review a Tax Court decision disallowing amortization deductions taken in 1952 and 1953 (in the respective amounts of $7,756.70 and $5,896.82) for premiums paid on the purchase of callable bonds.

The law applicable to this case is the Internal Revenue Code of 1939, Sections 23(v) and 125.1 The combined effect of these sections was to allow as a deduction from net income the difference between the purchase price of a bond (more accurately, the taxpayer's basis in the event of sale) and the amount payable if the bond were called.2

The taxpayer had the option of amortizing the full premium if he held bonds to the earliest date at which they were callable, whether they were called or not. There thus was created the opportunity to buy a bond at a premium and deduct the entire premium in the year of purchase if the bond were callable during that year.

Sophisticated taxpayers soon realized that by buying high premium bonds with early callable dates, borrowing from a bank up to the callable price, contributing their own funds to make up the premium, and pledging the bonds to the bank as collateral, they could convert their contribution into an instrument which could perform double duty. For, if the bonds, subject to the bank loan, were sold after six months, the net tax on capital gain (after deduction of bond amortization premium) would be effectively cut in half. If they were distributed to stockholders, a company would have both paid a dividend and, in effect, received a tax deduction for it. And if the bonds were given to charity, the taxpayer would have both a charitable deduction and a deduction for bond premium amortization.3

The gloss to the above text was written by the taxpayers in this case. To the elements of promptness in deduction and utilization of credit to buy bonds they added repetitiveness. In essence, they went through the above routine four times in each of the taxable years. They borrowed bank money, bought bonds from a broker, kept them for a period beyond the minimum 30 days required for a call, pledged the bonds to the bank, gave the bonds subject to the bank loan to a family charity, and deducted both the bond amortization premium and the charitable contributions. The charity then sold the bonds to the same broker, paid the bank, and kept essentially the amount of the premium. The broker immediately resold them to the taxpayers, who executed a new loan and pledge. After another minimum holding period, taxpayers gave them to the charity, subject to the bank loan, * * * and so on for four cycles.

The mechanics were simple. The bonds remained with the banks. Instructions at critical times followed forms devised by the broker. The same bonds were used, over and over again. In 1953 the procedure was simplified further, with the charity reselling the bonds through the broker directly to the taxpayers.

The Tax Court and respondent conceded that the basic pattern above described gave taxpayers, under the law applicable at the time, a double deduction (charitable and bond premium amortization) for each of the taxable years. But they draw the line at the first transaction and contend that in essence there was only one real transaction each year. The Tax Court reasoned that there was "a prearranged plan in which these round trips were contemplated from the start * * *.", and that one amortization deduction "will accurately reflect the true economic nature of the transaction before us."

The respondent frankly admits it has a hard row to hoe in view of Hanover Bank, et al. v. Commissioner of Internal Revenue, supra, and the other cases cited in note 3, supra. It states its basic position with plaintiff candor: "If any limit exists to the mere mechanical application of the statute, however, that line has assuredly been crossed in this case, where legal title was divested and revested in a matter of days (1952) or hours (1953)."

We agree with this statement of the issue. That drawing a line, however, is fraught with difficulty is underscored by the internal inconsistency of the respondent's arguments. These are four in number: (1) the successive transactions were really one, for only days or hours separated sale and repurchase, thus precluding any sufficient "hiatus" in taxpayers' ownership; (2) there was no more risk than would have been encountered if taxpayers had held the same bonds over the entire period; the situation would have been different if taxpayers had gone through the same number of transactions with different bonds; (3) the broker was a mere bookkeeper, the banks were not concerned, the charities were mere receptacles for "bare legal title", and all transactions were mere formalities; and (4) permitting deductions would not serve any "intelligible legislative purpose".

But if argument (1) were met by the intervening of a month, or six months between transactions, there would be a substantial hiatus but respondent could still make arguments (2) through (4). If taxpayers had taken the trouble to purchase four different issues of bonds each year and thus met argument (2) — and had kept to the same time schedule — they would still be vulnerable to arguments (3) and (4) (with the sole exception that the broker would have had to do more work in locating different early call, high premium securities). And if only one transaction a year had been executed — admittedly qualifying for a double deduction — arguments (3) and (4) are nevertheless applicable.

In short, we see no rationale in this admittedly technical and short-lived playing ground which gives us any guidance as to when to call "foul".

While, admittedly, taxpayers were shrewd and well advised, we cannot say, in the language of Knetsch v. United States, 1960, 364 U.S. 361, 366, 81 S. Ct. 132, 135, 5 L. Ed. 2d 128, that "there was nothing of substance to be realized * * from this transaction beyond a tax deduction." The husband taxpayer testified that, in addition to his basic objective of making as large a gift to his charity as he could, he had three other purposes in choosing to make several transactions rather than one:

"Q. * * * Why didn't you make one purchase of $200,000 instead [of four $50,000 purchases]?

A. Well, it was a matter of economics. First of all, to make one purchase would require a substantial outlay of which I didn't have available. It was naturally easier for me to do it at various times rather than all at one time. In addition, of course, there is the minimized possibility of a decline in the value of the bonds and a minimized chance of loss that would be sustained if the bonds were called."

Even were we to consider the approach of United States v. General Geophysical Co., 1961, 5 Cir., 296 F.2d 86, cert. denied, 1962, 369 U.S. 849, 82 S. Ct. 932, 8 L. Ed. 2d 8, as applicable to this case rather than that of Granite Trust Co. v. United States, 1956, 1 Cir., 238 F.2d 670, we could not say that six of the eight transactions in this case had an "unmistakably hollow sound" (296 F.2d at 86) while two did not.

We think pertinent the language of the court in Halle v. United States, supra, where the taxpayer had protected himself even more securely against the risks of loss by obtaining a guarantee of repurchase (put option) at a fixed price from his broker. The government tried to draw a line in that case by restricting the deduction to that part of the purchase price unprotected by the repurchase agreement. The court said:

"* * * we conclude that it would be unwise to attempt to formulate a rule apportioning to a taxpayer the benefit clearly provided by Congress under the then pertinent deduction statute by holding a transaction to be valid in part and `sham' in part." 346 F.2d at 551.

The mere fact that we face a series of transactions rather than one does not appear to us to change our basic rationale as set forth in Fabreeka Products Co. v. Commissioner of Internal Revenue, supra. This is a highly technical field. More often than not, the government profits by insisting on strict adherence to form. Cf. United States v. Rhode Island Hospital Trust Co., 1966, 1 Cir., 355 F.2d 7. We share respondent's feeling that the tax laws should not maintain avenues any wider than the traffic they are intended to carry. But we do not believe that adjustment of the traffic should be accomplished by changing the rules retrospectively, without clear rationale, and on an ad hoc basis to foil one taxpayer.

A judgment will be entered setting aside the decision of the Tax Court.