In the 1950’s Textron, Inc. had a wholly owned subsidiary, Hawaiian Textron, Inc. (Hawaiian). Hawaiian ran passenger ships between Hawaii and the West Coast and lost enormous sums of money in the process. In 1959, creditors foreclosed on its assets. Textron’s six million dollar investment in Hawaiian’s stock and debt became worthless, with one possible exception: Hawaiian’s huge losses could be used to reduce its taxable corporate income in future years. Because Hawaiian had no income prospects, a second, profitable corporation would have to be merged into Hawaiian’s empty shell to take advantage of its potential deductions. In 1959, the district court found, Textron had no specific plan to make use of this aspect of Hawaiian. In 1960, however, Hawaiian’s name was changed to Bell Aerospace Corp., and, using Textron’s funds, the renamed corporation acquired a successful business from Bell Aircraft.
Bell Aerospace made money, and it carried forward the old Hawaiian losses, amounting to some $6,745,000. At first the Internal Revenue Service (Service) disallowed any carry-over between two such different businesses. But in 1963 the Service reversed itself, ruling that a failing corporation may go into a new line of business and still carry forward losses from its earlier activity, so long as the owners of the corporation remain substantially the same. Rev. Ruling 63-40, 1963-1 C.B. 41. After the Service’s change of position, Bell Aerospace used Hawaiian’s losses to reduce its taxes for 1960, 1961, and 1962.
Just prior to the rehabilitation of Hawaiian, Textron had tried another way to make the best of its bad investment in Hawaiian. It took a six million dollar deduction in 1959, claiming that Hawaiian’s stock and debt became worthless to it in that year. The Service disallowed the deductions and assessed a deficiency. Textron paid and sued for a refund. In the district court and on appeal, the government has advanced only one argument — that Hawaiian’s stock was not worthless in 1959 because the shell had potential value as a source of carryover losses. The district court rejected this argument, as do we.
Ordinarily, bad debts and worthless stocks may be deducted only in the year in which they become wholly worthless. See 26 U.S.C. §§ 165(g)(3), 165(a), & 166(a)(1). See also 26 C.F.R. 1.165-4(a) (1976). By any ordinary definition, Hawaiian’s stock was quite worthless, as the district court found, in 1959. The shell together with the losses could not be marketed to others, 26 U.S.C. §§ 269 & 382, see infra.; and while it was within Textron’s power to rehabilitate the subsidiary, and, if profits were generated, deduct the tax losses on future tax returns, this contingency first required the infusion of brand new assets into what was a shell without assets — an initiative which created a wholly new ball game and certainly could not retroactively create value in 1959.
The government would have us cure what it perceives as an abuse of the tax laws by adopting its special definition of worthlessness.1 But Congress has already considered the abuse of high loss corporate shells. The remedy it chose does not call *1025into question the deduction Textron seeks.2 At one time, there was a large traffic in tax shells like Hawaiian, but Congress has now decreed that the buyer of such shells cannot take advantage of their tax attributes. 26 U.S.C. §§ 269 & 382. Thus, a failed company may have tax value today, but only if it has a large and healthy owner who is eager to acquire a second, more profitable business.
The Service has apparently made the argument it urges on us only once before. See Becker v. United States, 308 F.Supp. 555 (D.Neb.1970). There the taxpayers’ closely held corporation failed in 1956, and the taxpayers took a worthless stock deduction. In 1957, the corporation acquired a new and profitable business, which used the old business’s losses to offset its income. The Service argued that this showed the stock had not been worthless in 1956. The court rejected that claim:
“If this Court adopts the government’s theory in this case every taxpayer, at least every taxpayer who has the control of a closely held corporation, knowledge of the tax advantages and a desire to utilize these advantages, will be unable to take a deduction for worthless securities. As stated before, under this theory the tax advantages to his holdings would prevent his declaring those holdings as worthless. This Court does not believe, where the only evidence of some potential value remaining in a stock is carryfor-ward losses and the knowledge and desire to utilize them, that the taxpayer should be prevented from declaring the stock worthless. If taxpayer allowed 1956 to pass because of a desire to utilize the previous losses the only evidence during 1957 or any subsequent year of worthlessness would be futile attempts to make use of previous losses by forming a new corporation. As previously stated taxpayer must not only prove a stock is worthless but that it became worthless in the year in which the loss is taken. There would be no objective evidence of worthlessness for subsequent years as is present in 1956.” Id. at 557.
We are inclined to agree with the Becker court, for if we were to adopt the Service’s position we would introduce great uncertainty into this corner of tax law. A taxpayer who owned the bulk of a failed corporation’s stock would be hard-pressed to determine when his stock became worthless. If he guessed wrong, he might well forfeit his deduction. See, e. g., Keeney v. Commissioner of Internal Revenue, 116 F.2d 401 (2d Cir. 1940). The year in which the stock lost all value would depend on such vague and subjective factors as the taxpayer’s ability and desire to acquire another business. If the taxpayer lacked either desire or ability, the stock would be worthless when the corporation went under. Otherwise, it would continue to have worth for an uncertain period.3 Some taxpayers *1026might keep a loss shell, hoping to find a use for it, only to suffer unrelated losses that reduced their ability to obtain a second corporation. Under the Service’s approach, they would have to decide whether their setback was so severe that the loss shell had suddenly become truly worthless. Other taxpayers, hoping to take advantage of their loss shell, might acquire a second corporation that unexpectedly loses money. The shell’s carryover losses will be worthless to the taxpayers, but under the Service’s theory, they and the courts will have to decide just when the shell lost all value.
The Service might very well become the ultimate victim of the doctrine it now advocates. If a corporation has a bad year, primarily because one of its subsidiaries goes under, it will probably prefer to take its worthless stock deduction later. Under the Service’s approach, the corporation may do so, simply by going through the motions of seeking a second business for its shell. When the time for a deduction is more propitious, the “search” may be abandoned and the shell dissolved. The Service will have trouble proving that the corporation’s heart was not in the hunt. A rule with so much uncertainty and room for abuse should not be judicially created to close a loophole that is apparently used so seldom. Admittedly, Textron has turned its Hawaiian sow’s ear into a silk purse — and filled it at Treasury expense. But this is a matter that should be cured by statute or regulation, not by a far reaching retroactive court decision.4
The dissent, agreeing that the Service’s approach must fail, introduces a theory that the Service has advanced diffidently at best.5 The dissent would brand as a “double deduction” Textron’s worthless stock and debt claim and Bell Aerospace’s carryover loss deductions. We have grave doubts about the dissent’s casual eliding of the distinction between parent and subsidiary. They are separate taxpayers. In the absence of a consolidated return, cf. Ilfeld Co. v. Hernandez, 292 U.S. 62, 54 S.Ct. 596, 78 L.Ed. 1127 (1934), treating the two corporations as one may not be justified. But cf. Marwais Steel Co. v. Commissioner of Internal Revenue, 354 F.2d 997 (9th Cir. 1965). It is no answer to invoke the maxim that substance must prevail over form. Textron’s subsidiary was never a sham corporation lacking any substantial business purpose. In the first place, we are not inclined to adopt a policy of ignoring the distinction between parent and subsidiary in all tax cases. In the second place, corporate taxation is an area of careful planning, planning that will be seriously disrupted if courts simply ignore separate entities whenever it seems “fairer” to do so. We decline to inject so massive and unsettling a dose of “equity” into the tax laws without a clear *1027invitation from the Service and a careful exploration of the issue by both sides.
Moreover, attaching the label “double deduction” is not the end of analysis. We cannot decide this case by simply saying “No”. We have a duty to explain ourselves and to set out a rule of law to govern future cases. One rule at which the dissent hints would treat Bell Aerospace’s loss deductions like the recovery of a bad debt. This is not the case to apply such a rule. Textron’s worthless stock and debt deduction, which is at issue here, was taken in 1959, before Bell Aerospace began to take advantage of Hawaiian’s losses. Tex-tron’s deduction must have been proper at that time, for no “double” deduction had yet been sought. .Later recovery of a properly deducted bad debt does not void the first deduction. See 5 J. Mertens, Law of Federal Income Taxation § 30.37 (1975). Rather, the recovery becomes income in the year of recovery. See, e. g., West Seattle Nat’l Bank v. Commissioner of Internal Revenue, 288 F.2d 47, 49 (9th Cir. 1961). Applying the dissent’s “tax benefit” theory thus would lead us to question Bell Aerospace’s deductions in 1960 and later years, but not the original deductions taken by Textron in 1959. The Bell Aerospace deductions, however, were allowed many years ago, after strict government scrutiny. If the Service wished to advance the theory developed by the dissent, it had ample opportunity to do so in the early 1960’s when it examined Bell Aerospace’s deductions. We cannot cure in this case the Service’s decision not to attack those loss deductions.
At other points the dissent suggests that the first deduction should be disallowed because the second is no longer within reach. A rule invalidating worthless stock deductions because of later loss carryovers would be odd. Worthless stock deductions would linger in limbo for years before a taxpayer could know whether they were proper. The Service would be inclined to challenge every such deduction for fear that the shell’s losses would later prove valuable to the taxpayer. Even if we agreed that separate corporate identities could be easily ignored, we would be reluctant to approve such a rule.
Affirmed.
. We wonder, however, whether the government’s argument is different in kind from a line of reasoning that would ultimately destroy the worthless stock deduction. Stock that becomes worthless in a particular year represents a significant tax gain for its owner. The deduction that the stock permits has real economic value, a fact that could lead the government to make the same argument that it makes here: the stock has tax value, therefore it cannot be worthless. If accepted, such an argument would make it impossible ever to take a worthless stock deduction.
. This is not to say, given the policy disfavoring double deductions referred to in the dissent, that the Treasury might not legitimately have gone beyond the language of the Code to construct, by regulation, a mechanism for limiting all deductions in this situation to one. Such a regulation would be less an interpretation of a particular provision than a means of avoiding what there is ample reason to assume Congress did not mean to permit. Quite likely if the Treasury had anticipated this situation, its regulation would have limited the subsidiary’s later use of the carry forward losses in light of the parent’s early election to claim a worthless stock deduction — the converse of the present approach. But whatever the actual mechanics, the regulation would have signalled the Treasury’s intentions in an area where, otherwise, there is nothing to tell the taxpayer that he is lacking an open field. Textron, to be sure, as the dissent notes, had no inherent right to a double deduction. It did, however, have a right to have the government proceed in accordance with the Code and regulations. See Lan Jen Chu v. C.I.R., 486 F.2d 696, 704 (1st Cir. 1973) (concurring opinion). Loopholes cannot be repaired after the fact simply on the principle . that to apply the law as written leads to a bad result.
. The Treasury could, of course, determine worth in light of subsequent events, i. e., treating the stock here as worthless in 1959 only in light of the subsidiary’s use of the losses in later years. But a rule holding up the computation of taxes in one year for years to come would seem unworkable. On the other hand if a retrospective approach were not adopted, taxpayers and the Treasury would be burdened with the speculations we have mentioned. The *1026Treasury could, of course, disallow a worthless stock deduction unless the parent dissolved the subsidiary, but such an approach would go beyond the present code and regulations, and might well impose unwarranted burdens in the generality of cases.
. In any event, it is not clear that the government’s approach would justify disallowing Tex-tron’s worthless debt deduction as well as its worthless stock deduction. Textron’s stock carried with it the power to control Hawaiian, and thus to orchestrate acquisition of a new business. It is this power to control that the government believes valuable. No such power attends Textron’s status as a major creditor, so it is hard to see how Hawaiian’s debts had any continuing value to Textron after 1959, except to the extent that any worthless debt has value as a deduction. Cf. n. 1, supra. Because our decision rests on broader ground, we need not explore this difficult question further.
. At the close of its main brief, the Service virtually concedes that the double deduction cases will not carry the weight the dissent puts on them:
“Since the worthless stock and bad debt deductions here in question were the first to be claimed, and since the net operating losses were claimed by a separate corporate entity it may well be that the ‘double deduction’ rationale of Ilfeld Co. v. Marwais Steel would not provide a sufficient basis, standing alone, to disallow the worthless stock and debt deductions here in issue if the stock and debt could be deemed to have become wholly worthless by the end of the year in question.” Appellant’s brief, p. 30.
All three members of this panel agree that the stock and debt at issue here should be so deemed.