delivered the opinion of the Court.

Decedent, Milliken C. Byrum, created in 1958 an irrevocable trust to which he transferred shares of stock in three closely held corporations. Prior to transfer, he owned at least 71% of the outstanding stock of each corporation. The beneficiaries were his children or, in the event of their death before the termination of the trust, their surviving children. The trust instrument specified that there be a corporate trustee. Byrum designated as sole trustee an independent corporation, Huntington National Bank. The trust agreement vested *127in the trustee broad and detailed powers with respect to the control and management of the trust property. These powers were exercisable in the trustee’s sole discretion, subject to certain rights reserved by Byrum: (i) to vote the shares of unlisted stock held in the trust estate; (ii) to disapprove the sale or transfer of any trust assets, including the shares transferred to the trust; (iii) to approve investments and reinvestments; and (iv) to remove the trustee and “designate another corporate Trustee to serve as successor.” Until the youngest living child reached age 21, the trustee was authorized in its “absolute and sole discretion” to pay the income and principal of the trust to or for the benefit of the beneficiaries, “with due regard to their individual needs for education, care, maintenance and support.” After the youngest child reached 21, the trust was to be divided into a separate trust for each child, to terminate when the beneficiaries reached 35. The trustee was authorized in its discretion to pay income and principal from these trusts to the beneficiaries for emergency or other “worthy need,” including education.1

*128When he died in 1964, Byrum owned less than 60% of the common stock in two of the corporations and 59% in the third. The trust had retained the shares *129transferred to it, with the result that Byrum had continued to have the right to vote not less than 71% of the common stock in each of the three corpora*130tions.2 There were minority stockholders, unrelated to Byrum, in each corporation.

Following Byrum’s death, the Commissioner of Internal Revenue determined that the transferred stock was properly included within Byrum’s gross estate under § 2036 (a) of the Internal Revenue Code of 1954, 26 U. S. C. §2036 (a). That section provides for the inclusion in a decedent’s gross estate of all property which the decedent has transferred by inter vivos transaction, if he has retained for his lifetime “(1) the possession or enjoyment of, or the right to the income from, the property” transferred, or “(2) the right, either alone or in conjunction with any person, to designate the persons who shall possess or enjoy the property or the income *131therefrom.”3 The Commissioner determined that the stock transferred into the trust should be included in Byrum’s gross estate because of the rights reserved by him in the trust agreement. It was asserted that his right to vote the transferred shares and to veto any sale thereof by the trustee, together with the ownership of other shares, enabled Byrum to retain the “enjoyment of . . . the property,” and also allowed him to determine the flow of income to the trust and thereby “designate the persons who shall . . . enjoy . . . the income.”

The executrix of Byrum’s estate paid an additional tax of $13,202.45, and thereafter brought this refund action in District Court. The facts not being in dispute, the court ruled for the executrix on cross motions for summary judgment. 311 F. Supp. 892 (SD Ohio 1970). The Court of Appeals affirmed, one judge dissenting. 440 F. 2d 949 (CA6 1971). We granted the Government’s petition for certiorari. 404 U. S. 937 (1971).

I

The Government relies primarily on its claim, made under §2036 (a)(2), that Byrum retained the right to *132designate the persons who shall enjoy the income from the transferred property. The argument is a complicated one. By retaining voting control over the corporations whose stock was transferred, Byrum was in a position to select the corporate directors. He could retain this position by not selling the shares he owned and by vetoing any sale by the trustee of the transferred shares. These rights, it is said, gave him control over corporate dividend policy. By increasing, decreasing, or stopping dividends completely, it is argued that Byrum could “regulate the flow of income to the trust” and thereby shift or defer the beneficial enjoyment of trust income between the present beneficiaries and the re-maindermen. The sum of this retained power is said to be tantamount to a grantor-trustee’s power to accumulate income in the trust, which this Court has recognized constitutes the power to designate the persons who shall enjoy the income from transferred property.4

At the outset we observe that this Court has never held that trust property must be included in a settlor’s gross estate solely because the settlor retained the power *133to manage trust assets. On the contrary, since our decision in Reinecke v. Northern Trust Co., 278 U. S. 339 (1929), it has been recognized that a settlor’s retention of broad powers of management does not necessarily subject an inter vivos trust to the federal estate tax.5 Although there was no statutory analogue to § 2036 (a)(2) when Northern Trust was decided, several lower court decisions decided after the enactment of the predecessor of § 2036 (a)(2) have upheld the settlor’s right to exercise managerial powers without incurring estate-tax liability.6 In Estate of King v. Commissioner, 37 T. C. 973 (1962), a settlor reserved the power to direct the trustee in the management and investment of trust assets. The Government argued that the settlor was thereby empowered to cause investments to be made in such a manner as to control significantly the flow of income into the trust. The Tax Court rejected this argument, and held for the taxpayer. Although the court recognized that the settlor had reserved “wide latitude in the exercise of his discretion as to the types of investments to be made,” id., at 980, it did not find this control over the flow of income to be equivalent *134to the power to designate who shall enjoy the income from the transferred property.

Essentially the power retained by Byrum is the same managerial power retained by the settlors in Northern Trust and in King. Although neither case controls this one — Northern Trust, because it was not decided under §2036 (a)(2) or a predecessor; and King, because it is a lower court opinion — the existence of such precedents carries weight.7 The holding of Northern Trust, that the settlor of a trust may retain broad powers of management without adverse estate-tax consequences, may have been relied upon in the drafting of hundreds of inter vivos trusts.8 The modification of this principle now sought by the Government could have a seriously adverse impact, especially upon settlors (and their estates) who happen to have been “controlling” stock*135holders of a closely held corporation. Courts properly have been reluctant to depart from an interpretation of tax law which has been generally accepted when the departure could have potentially far-reaching consequences. When a principle of taxation requires reexamination, Congress is better equipped than a court to define precisely the type of conduct which results in tax consequences. When courts readily undertake such tasks, taxpayers may not rely with assurance on what appear to be established rules lest they be subsequently overturned. Legislative enactments, on the other hand, although not always free from ambiguity, at least afford the taxpayers advance warning.

The Government argues, however, that our opinion in United States v. O’Malley, 383 U. S. 627 (1966), compels the inclusion in Byrum’s estate of the stock owned by the trust. In O’Malley, the settlor of an inter vivos trust named himself as one of the three trustees. The trust agreement authorized the trustees to pay income to the life beneficiary or to accumulate it as a part of the principal of the trust in their “sole discretion.” The agreement further provided that net income retained by the trustees, and not distributed in any calendar year, “ 'shall become a part of the principal of the Trust Estate.’ ” Id., at 629 n. 2. The Court characterized the effect of the trust as follows:

“Here Fabrice [the settlor] was empowered, with the other trustees, to distribute the trust income to the income beneficiaries or to accumulate it and add it to the principal, thereby denying to the beneficiaries the privilege of immediate enjoyment and conditioning their eventual enjoyment upon surviving the termination of the trust.” Id., at 631.

As the retention of this legal right by the settlor, acting as a trustee “in conjunction” with the other trustees, *136came squarely within the language and intent of the predecessor of § 2036 (a)(2), the taxpayer conceded that the original assets transferred into the trust were in-cludable in the decedent’s gross estate. Id., at 632. The issue before the Court was whether the accumulated income, which had been added to the principal pursuant to the reservation of right in that respect, was also in-cludable in decedent’s estate for tax purposes. The Court held that it was.

In our view, and for the purposes of this case, O’Malley adds nothing to the statute itself. The facts in that case were clearly within the ambit of what is now § 2036 (a). That section requires that the settlor must have “retained for his life ... (2) the right... to designate the persons who shall possess or enjoy the property or the income therefrom.” O’Malley was covered precisely by the statute for two reasons: (1) there the settlor had reserved a legal right, set forth in the trust instrument; and (2) this right expressly authorized the settlor, “in conjunction” with others, to accumulate income and thereby “to designate” the persons to enjoy it.

It must be conceded that Byrum reserved no such “right” in the trust instrument or otherwise. The term “right,” certainly when used in a tax statute, must be given its normal and customary meaning. It connotes an ascertainable and legally enforceable power, such as that involved in O’Malley.9 Here, the right ascribed to Byrum was the power to use his majority position and influence over the corporate directors to “regulate the flow of dividends” to the trust. That “right” was *137neither ascertainable nor legally enforceable and hence was not a right in any normal sense of that term.10

Byrum did retain the legal right to vote shares held by the trust and to veto investments and reinvestments. But the corporate trustee alone, not Byrum, had the right to pay out or withhold income and thereby to designate who among the beneficiaries enjoyed such income. Whatever power Byrum may have possessed with respect to the flow of income into the trust was derived not from an enforceable legal right specified in the trust instrument, but from the fact that he could elect a majority of the directors of the three corporations. The power to elect the directors conferred no legal right to command them to pay or not to pay dividends. A majority shareholder has a fiduciary duty not to misuse his power by promoting his personal interests at the expense of corporate interests.11 Moreover, *138the directors also have a fiduciary duty to promote the interests of the corporation.12 However great Byrum’s influence may have been with the corporate directors, their responsibilities were to all stockholders and were enforceable according to legal standards entirely unrelated to the needs of the trust or to Byrum’s desires with respect thereto.

The Government seeks to equate the de facto position of a controlling stockholder with the legally enforceable “right” specified by the statute. Retention of corporate control (through the right to vote the shares) is said to be “tantamount to the power to accumulate income” in the trust which resulted in estate-tax consequences in O’Malley. The Government goes on to assert that “[tjhrough exercise of that retained power, [Byrum] could increase or decrease corporate dividends . . . and thereby shift or defer the beneficial enjoyment of trust income.” 13 This approach seems to us *139not only to depart from the specific statutory language,14 but also to misconceive the realities of corporate life.

There is no reason to suppose that the three corporations controlled by Byrum were other than typical small businesses. The customary vicissitudes of such enterprises — bad years; product obsolescence; new competition; disastrous litigation; new, inhibiting Government regulations; even bankruptcy — prevent any certainty or predictability as to earnings or dividends. There is no assurance that a small corporation will have a flow of net earnings or that income earned will in fact be available for dividends. Thus, Byrum’s alleged de jacto “power to *140control the flow of dividends” to the trust was subject to business and economic variables over which he had little or no control.

Even where there are corporate earnings, the legal power to declare dividends is vested solely in the corporate board. In making decisions with respect to dividends, the board must consider a number of factors. It must balance the expectation of stockholders to reasonable dividends when earned against corporate needs for retention of earnings. The first responsibility of the board is to safeguard corporate financial viability for the long term. This means, among other things, the retention of sufficient earnings to assure adequate working capital as well as resources for retirement of debt, for replacement and modernization of plant and equipment, and for growth and expansion. The nature of a corporation’s business, as well as the policies and long-range plans of management, are also relevant to dividend payment decisions.15 Directors of a closely held, small corporation must bear in mind the relatively limited access of such an enterprise to capital markets. This may require a more conservative policy with respect to dividends than would be expected of an established corporation with securities listed on national exchanges.16

*141Nor do small corporations have the flexibility or the opportunity available to national concerns in the utilization of retained earnings. When earnings are substantial, a decision not to pay dividends may result only in the accumulation of surplus rather than growth through internal or external expansion. The accumulated earnings may result in the imposition of a penalty tax.17

These various economic considerations are ignored at the directors’ peril. Although vested with broad discretion in determining whether, when, and what amount of dividends shall be paid, that discretion is subject to legal restraints. If, in obedience to the will of the majority stockholder, corporate directors disregard the interests of shareholders by accumulating earnings to an unreasonable extent, they are vulnerable to a derivative suit.18 They are similarly vulnerable if they make an unlawful payment of dividends in the absence of net earnings or available surplus,19 or if they fail to exer*142cise the requisite degree of care in discharging their duty to act only in the best interest of the corporation and its stockholders.

Byrum was similarly inhibited by a fiduciary duty from abusing his position as majority shareholder for personal or family advantage to the detriment of the corporation or other stockholders. There were a substantial number of minority stockholders in these corporations who were unrelated to Byrum.20 Had Byrum and the directors violated their duties, the minority shareholders would have had a cause of action under Ohio law.21 The Huntington National Bank, as trustee, was one of the minority stockholders, and it had both the right and the duty to hold Byrum responsible for any wrongful or negligent action as a controlling stockholder or as a director of the corporations.22 Although Byrum had reserved the right to remove the trustee, he would have been imprudent to do this when confronted by the *143trustee’s complaint against his conduct. A successor trustee would succeed to the rights of the one removed.

We conclude that Byrum did not have an unconstrained de facto power to regulate the flow of dividends to the trust, much less the ‘‘right” to designate who was to enjoy the income from trust property. His ability to affect, but not control, trust income, was a qualitatively different power from that of the settlor in O’Malley, who had a specific and enforceable right to control the income paid to the beneficiaries.23 Even had Byrum managed to flood the trust with income, he had no way of compelling the trustee to pay it out rather than accumulate it. Nor could he prevent the trustee from making payments from other trust assets,24 although admittedly there were few of these at the time of Byrum’s death. We cannot assume, however, that no other assets would come into the trust from reinvestments or other gifts.25

*144We find no merit to the Government’s contention that Byrum’s de facto “control,” subject as it was to the economic and legal constraints set forth above, was tantamount to the right to designate the persons who shall enjoy trust income, specified by § 2036 (a)(2).26

*145II

The Government asserts an alternative ground for including the shares transferred to the trust within Byrum’s gross estate. It argues that by retaining control, Byrum guaranteed himself continued employment and remuneration, as well as the right to determine whether and when the corporations would be liquidated or merged. Byrum is thus said to have retained “the . .. enjoyment of . . . the property” making it includable within his gross estate under §2036 (a)(1). The Government concedes that the retention of the voting rights of an “unimportant minority interest” would not require inclusion of the transferred shares under § 2036 (a)(1). It argues, however, “where the cumulative effect of the retained powers and the rights flowing from the shares not placed in trust leaves the grantor in control of a close corporation, and assures that control for his lifetime, he has retained the 'enjoyment’ of the transferred stock.” 27 Brief for United States 23.

It is well settled that the terms “enjoy” and “enjoyment,” as used in various estate tax statutes, “are not terms of art, but connote substantial present economic benefit rather than technical vesting of title or estates.” Commissioner v. Estate of Holmes, 326 U. S. 480, 486 *146(1946).28 For example, in Reinecke v. Northern Trust Co., 278 U. S. 339 (1929), in which the critical inquiry was whether the decedent had created a trust “intended ... ‘to take effect in possession or enjoyment at or after his death,’ ” 29 id., at 348, the Court held that reserved powers of management of trust assets, similar to Byrum’s power over the three corporations, did not subject an inter vivos trust to the federal estate tax. In determining whether the settlor had retained the enjoyment of the transferred property, the Court said:

“Nor did the reserved powers of management of the trusts save to decedent any control over the economic benefits or the enjoyment of the property. He would equally have reserved all these powers and others had he made himself the trustee, but the transfer would not for that reason have been incomplete. The shifting of the economic interest in the trust property which was the subject of the tax was thus complete as soon as the trust was made. His power .to recall the property and of control over it for his own benefit then ceased and as the trusts were not made in contemplation *147of death, the reserved powers do not serve to distinguish them from any other gift inter vivos not subject to the tax.” 278 U. S., at 346-347.

The cases cited by the Government reveal that the terms “possession” and “enjoyment,” used in § 2036 (a) (1), were used to deal with situations in which the owner of property divested himself of title but retained an income interest or, in the case of real property, the lifetime use of the property. Mr. Justice Black’s opinion for the Court in Commissioner v. Estate of Church, 335 U. S. 632 (1949), traces the history of the concept. In none of the cases cited by the Government has a court held that a person has retained possession or enjoyment of the property if he has transferred title irrevocably, made complete delivery of the property and relinquished the right to income where the property is income producing.30

The Government cites only one case, Estate of Holland v. Commissioner, 1 T. C. 564 (1943),31 in which a decedent had retained the right to vote transferred shares of stock and in which the stock was included *148within the decedent’s gross estate. In that case, it was not the mere power to vote the stock, giving the decedent control of the corporation, which caused the Tax Court to include the shares. The court held that “ 'on an inclusive view of the whole arrangement, this withholding of the income until decedent’s death, coupled with the retention of the certificates under the pledge and the reservation of the right to vote the stock and to designate the company officers’ ” subjects the stock to inclusion within the gross estate. Id., at 565. The settlor in Holland retained a considerably greater interest than Byrum retained, including an income interest.32

As the Government concedes, the mere retention of the right-to-vote shares does not constitute the type of “enjoyment” in the property itself contemplated by § 2036 (a)(1). In addition to being against the weight of precedent, the Government’s argument that Byrum retained “enjoyment” within the meaning of § 2036 (a)(1) is conceptually unsound. This argument implies, as it must under the express language of § 2036 (a), that Byrum “retained for his life ... (1) the possession or enjoyment” of the “property” transferred to the trust or the “income” therefrom. The only property he transferred was corporate stock. He did not transfer “control” (in the sense used by the Government) as the trust never owned as much as 50% of the stock of any corporation. Byrum never divested himself of control, as he was able to vote a majority of the shares by virtue of what he owned and the right to vote those placed in *149the trust. Indeed, at the time of his death he still owned a majority of the shares in the largest of the corporations and probably would have exercised control of the other two by virtue of being a large stockholder in each.33 The statutory language plainly contemplates retention of an attribute of the property transferred — such as a right to income, use of the property itself, or a power of appointment with respect either to income or principal.34

Even if Byrum had transferred a majority of the stock, but had retained voting control, he would not have retained “substantial present economic benefit,” 326 U. S., at 486. The Government points to the retention of two “benefits.” The first of these, the power to liquidate or *150merge, is not a present benefit; rather, it is a speculative and contingent benefit which may or may not be realized. Nor is the probability of continued employment and compensation the substantial “enjoyment of . . . [the transferred] property” within the meaning of the statute. The dominant stockholder in a closely held corporation, .if he is active and productive, is likely to hold a senior position and to enjoy the advantage of a significant voice in his own compensation. These are inevitable facts of the free-enterprise system, but the influence and capability of a controlling stockholder to favor himself are not without constraints. Where there are minority stockholders, as in this case, directors may be held accountable if their employment, compensation, and retention of officers violate their duty to act reasonably in the best interest of the corporation and all of its stockholders.35 Moreover, this duty is policed, albeit indirectly, by the Internal Revenue Service, which disallows the deduction of unreasonable compensation paid to a corporate executive as a business expense.36 We conclude that Byrum’s retention of voting control was not the retention of the enjoyment of the transferred property within the meaning of the statute.

*151For the reasons set forth above, we hold that this case was correctly decided by the Court of Appeals and accordingly the judgment is

Affirmed.

with whom Mr. Justice Brennan and Mr. Justice Blackmun join, dissenting.

I think the majority is wrong in all substantial respects.

I

The tax code commands the payment of an estate tax on transfers effective in name and form during life if the now deceased settlor retained during his life either (1) “the possession or enjoyment of” the property transferred or (2) the right to designate the persons who would enjoy the transferred property or the income therefrom. 26 U. S. C. §§2036 (a)(1) and (2). Our cases explicate this congressional directive to mean that if one wishes to avoid a tax at death he must be self-abnegating enough to totally surrender his property interest during life.

“[A]n estate tax cannot be avoided by any trust transfer except by a bona fide transfer in which the settlor, absolutely, unequivocally, irrevocably, and without possible reservations, parts with all of his title and all of his possession and all of his enjoyment of the transferred property.” Commissioner v. Estate of Church, 335 U. S. 632, 645 (1949).

In this case the taxpayer's asserted alienation does not measure up to this high standard. Byrum enjoyed the continued privilege of voting the shares he “gave up” to the trust. By means of these shares he enjoyed majority control of two corporations. He used that control to retain salaried positions in both corporations. To my *152mind this is enjoyment of property put beyond taxation only on the pretext that it is not enjoyed.

Byrum's lifelong enjoyment of the voting power of the trust shares contravenes § 2036 (a) (2) as well as § 2036 (a)(1) because it afforded him control over which trust beneficiaries — the life tenants or the remainder-men — would receive the benefit of the income earned by these shares. He secured this power by making the trust to all intents and purposes exclusively dependent on shares it could not sell in corporations he controlled.1 Thus, by instructing the directors he elected in the controlled corporations that he thought dividends should or should not be declared Byrum was able to open or close the spigot through which income flowed to the trust’s life tenants. When Byrum closed the spigot by deferring dividends of the controlled corporations, thereby perpetuating his own “enjoyment” of these funds, he also in effect transferred income from the life tenants to the remaindermen whose share values were swollen by the retained income. The extent to which such income transfers can be effected is suggested by the pay-out record of the corporations here in question, as reflected in the trust's accounts. Over the first five years of its existence on shares later valued by the Internal Revenue Service at $89,000, the trust received a total of only $339 in dividends. In the sixth year, Byrum died. The corporations raised their dividend rate from 100 a share to $2 per share and paid $1,498 into the trust. See “Income Cash Ledger,” App. 25-26.

*153Byrum’s control over the flow of trust income renders his estate scheme repugnant to § 2036 (a) (2) as well as § 2036 (a)(1).

To me it is thus clear that Byrum’s shares were not truly, totally, “absolutely, unequivocally” alienated during his life. When it is apparent that, if tolerated, Byrum’s scheme will open a gaping hole in the estate tax laws, on what basis does the majority nonetheless conclude that Byrum should have his enjoyment, his control, and his estate free from taxes?

II

I can find nothing in the majority’s three arguments purporting to deal with §2036 (a)(1), that might justify the conclusion that Byrum did not “enjoy” a benefit from the shares his estate now asserts are immune from taxation.

1. The majority says that in Reinecke v. Northern Trust Co., 278 U. S. 339 (1929), “the Court held that reserved powers of management of trust assets, similar to Byrum’s power over the three corporations, did not subject an inter vivos trust to the federal estate tax.” This reading of Northern Trust is not warranted by the one paragraph in that antique opinion on the point for which it is now cited, see 278 U. S., at 346-347, nor by the circumstances of that case. No one has ever suggested that Adolphus Bartlett, the settlor in Northern Trust, used or could have used the voting power of the shares he transferred to a trust to control or, indeed, exercise any significant influence in any company. A mere glance at the nature of these securities transferred by Bartlett (e. g., 1,000 shares of the Northern Trust Co., 784 shares of the Commonwealth Edison Co., 300 shares of the Illinois Central R.Co., 200 preferred shares of the Chicago & North Western R. Co., 300 common shares of the Chicago & *154North Western R. Co.) 2 shatters any theory that might lead one to believe that the Court in Northern Trust was concerned with anything like the transactions in this case. On what basis, then, does the majority say that Northern Trust involved a decision on facts "similar to Byrum's power over the three corporations”? And on what basis does it say that the Government’s position that Byrum’s use of trust shares to retain control renders those shares taxable is “against the weight of precedent?”

2. The majority implies that trust securities are taxable only if the testator retained title or the right to income from the securities until death. But this ignores the plain language of the statute which proscribes “enjoyment” as well as “possession or . . . the right to income.”

3. The majority concludes with the assertion that By-rum secured no “substantial present economic benefits” from his retention of control.3 It is suggested that con*155trol is not important, that it either cannot be held by a private shareholder or that it is of so little use and relevance the taxpayer can hardly be said to have “enjoyed” it. This view of corporate life is refuted by the case law; 4 by the commentators; 5 and by everyday transactions on the stock exchange where offers and trades repeatedly demonstrate that the power to “control” a corporation will fetch a substantial premium.6 Moreover, the majority’s view is belied by Byrum’s own conduct. He obviously valued control because he forbade the bank that served as trustee to sell the trust shares in these corporations without his — Byrum’s — approval, whatever their return, their prospects, their value, or the trust’s needs. Trust Agreement ¶ 5.15, App. 14.

In sum, the majority’s discourse on §2036 (a)(1) is an unconvincing rationalization for allowing Byrum the tax-free “enjoyment” of the control privileges he retained through the voting power of shares he supposedly “absolutely” and “unequivocally” gave up.

*156Ill

I find no greater substance in the greater length of the majority’s discussion of §2036 (a)(2).

A

Approaching the § 2036 (a) (2) problem afresh, one would think United States v. O’Malley, 383 U. S. 627 (1966), would control this case. In O’Malley the settlor “had relinquished all of his rights” to stock, but he appointed himself one of three trustees for each of the five trusts he created, and he drafted the trust agreement so that the trustees had the freedom to allocate trust income to the life tenant or to accumulate it for the remainderman “in their sole discretion.” The District Court held that the value of securities transferred was includable in the settlor’s gross estate under § 811 (c) and (d) of the Internal Revenue Code of 1939, as amended, §811 (c)(1)(B) being the similarly worded predecessor of § 2036 (a), because the settlor had retained the power to allocate income between the beneficiaries without being constrained by a “definite ascertainable standard” according to which the trust would be administered. O’Malley v. United States, 220 F. Supp. 30, 33 (1963). The court noted “plaintiff’s contention that the required external standard is imposed generally by the law of Illinois,” but found this point to be “without merit.”

“The cases cited by plaintiff clearly set out fundamental principles of trust law: that a trust requires a named beneficiary; that the legal and equitable estates be separated; and, that the trustees owe a fiduciary duty to the beneficiaries. These statements of the law are not particular to Illinois. Nor do these requirements so circumscribe the trustee’s powers in an otherwise unrestricted trust so as to hold such a trust governed by an external standard *157and thus excludable from the application of § 811 (c) and (d).” 220 F. Supp., at 33-34.

It was another aspect of that case that brought the matter to the Court of Appeals, 340 F. 2d 930 (CA7 1964), and then here. We were asked to decide whether the lower court’s holding should be extended and the accumulated income as well as the principal of the trust included in the settlor’s taxable estate because the settlor had retained excessive power to designate the income beneficiaries of the shares transferred. We held that, though capable of exercise only in conjunction with one other trustee, the power to allocate income without greater constraint than that imposed “is a significant power ... of sufficient substance to be deemed the power to ‘designate’ within the meaning of [the predecessor of § 2036 (a) (2)].” 383 U. S., at 631.

O’Malley makes the majority’s position in this case untenable. O’Malley establishes that a settlor serving as a trustee is barred from retaining the power to allocate trust income between a life tenant and a remainder-man if he is not constrained by more than general fiduciary requirements. See also Commissioner v. Estate of Holmes, 326 U. S. 480 (1946),7 and Lober v. United States, 346 U. S. 335 (1953). Now the majority would have us accept the incompatible position that a settlor seeking tax exemption may keep the power of income allocation by rendering the trust dependent on an income flow he controls because the general fiduciary obligations of a director are sufficient to eliminate the power to designate within the meaning of § 2036 (a)(2).8

*158B

The majority would prop up its untenable position by suggesting that a controlling shareholder is constrained in his distribution or retention of dividends by fear of derivative suits, accumulated earnings taxes, and “various economic considerations . . . ignored at the director’s peril.” I do not deny the existence of such constraints, but their restraining effect on an otherwise tempting gross abuse of the corporate dividend power hardly guts the great power of a controlling director to accelerate or retard, enlarge or diminish, most dividends. The penalty taxes only take effect when accumulations exceed $100,000, 26 U. S. C. § 535 (c); Byrum was free to accumulate up to that ceiling. The threat of a derivative suit is hardly a greater deterrent to accumulation. As Cary puts it:

“The cases in which courts have refused to require declaration of dividends or larger dividends despite the existence of current earnings or a substantial surplus or both are numerous; plaintiffs have won only a small minority of the cases. The labels are ‘business judgment’; ‘business purpose’; ‘non-inter*159ference in internal affairs.’ The courts have accepted the general defense of discretion, supplemented by one or more of a number of grounds put forward as reasons for not paying dividends or larger dividends . . . W. Cary, Cases and Materials on Corporations 1587 (4th ed. 1969).

And cf. Commissioner v. Sunnen, 333 U. S. 591, 609 (1948).

The ease with which excess taxes, derivative suits, and economic vicissitudes alike may be circumvented or hurdled if a controlling shareholder chooses to so arrange his affairs is suggested by the pay-out record of Byrum’s corporations noted above.

C

The majority proposes one other method of distinguishing O’Malley. Section 2036 (a)(2), it is said, speaks of the right to designate income beneficiaries. O’Malley involved the effort of a settlor to maintain a legal right to allocate income. In the instant case only the power to allocate income is at stake. The Government’s argument is thus said to depart from “the specific statutory language” 9 and to stretch the statute beyond endurance by allocating tax according to the realities of the situation rather than by the more rigid yardstick of formal control.10

This argument conjures up an image of congressional care in the articulation of § 2036 (a) (2) that is entirely at odds with the circumstances of its passage. The 1931 legislation, which first enacted what is now § 2036 (a)(2) in language not materially amended since that date, *160passed both Houses of Congress in one day — the last day of the session. There was no printed committee report. Substantial references to the bill appear in only two brief sections of the Congressional Record.11 Under the circumstances I see no warrant for reading the words in a niggardly way.

Moreover, it appears from contemporary evidence that if the use of the word “right” was intended to have any special meaning it was to expand rather than to contract the reach of the restraint effected by the provision in which it appeared. The House Report on *161the Revenue Act of 1932 notes in relation to amendment of the predecessor of § 2036 (a)(1) that:

“The insertion of the words 'the right to the income’ in place of the words ‘the income’ is designed to reach a case where decedent had the right to the income, though he did not actually receive it. This is also a clarifying change.” H. R. Rep. No. 708, 72d Cong., 1st Sess., 47.

And see S. Rep. No. 665, 72d Cong., 1st Sess., 50, to the same effect.

I repeat the injunction of Mr. Justice Frankfurter, 25 years ago: “This is tax language and should be read in its tax sense.” United States v. Ogilvie Hardware Co., 330 U. S. 709, 721 (1947) (dissenting opinion).

Lest this by itself not be considered enough to refute the majority’s approach, I must add that it is quite repugnant to the words and sense of our opinion in O’Malley to read it as though it pivoted on an interpretation of “right” rather than power. The opinion could hardly have been more explicitly concerned with the realities of a settlor’s retained power rather than the theoretical legal form of the trust. Thus we said:

“Here Fabrice [the settlor] was empowered . . . . This is a significant power ... of sufficient substance to be deemed the power to ‘designate’ within the meaning of [the predecessor of § 2036 (a) (2)] ... .” 383 U. S., at 631 (emphasis supplied).

And we said:

“With the creation of the trusts, he relinquished all of his rights to income except the power to distribute that income to the income beneficiaries or to accumulate it and hold it for the remaindermen of the trusts.” 383 U. S., at 632 (emphasis supplied).

*162And we spoke of:

“This power he exercised by accumulating and adding income to principal and this same power he held until the moment of his death. ...” 383 U. S., at 634 (emphasis supplied).

Other passages could be quoted.

IV

Apparently sensing that considerations of logic, policy, and recent case law point to the inclusion of Byrum’s trust in his estate, the majority would blunt these considerations by invoking the principle that courts should refrain from decisions detrimental to litigants who have taken a position in legitimate reliance on possibly outdated, but once established, case law. This principle is said to bring great weight to bear in Byrum’s favor.

I need not quarrel with the principle. I think, however, that its application in this context is inappropriate.

The majority recites these facts: This Court has never held that retention of power to manage trust assets compels inclusion of a trust in a settlor’s estate. In fact, Reinecke v. Northern Trust Co., 278 U. S. 339 (1929), specifically held a trust arrangement immune from taxation though the settlor retained power to decide how the trust assets were to be invested. Though Northern Trust was decided before the passage of § 2036 (a) (2), it has been followed by “several” more recent lower court decisions. Though most of the lower court decisions provide only the most oblique reference to circumstances like those of this case, a 1962 unappealed Tax Court decision, Estate of King v. Commissioner, 37 T. C. 973, is squarely in point.

On the basis of these two authorities, a 1929 Supreme Court decision and an unreviewed 1962 Tax Court decision, the majority concludes that there exists a “gener*163ally accepted” rule that Byrum might do what he had done here. It is said that the hypothesized rule “may” have been relied upon by “hundreds” of others; if so, its modification “could” have a serious impact, especially upon settlors who “happen” to have been controlling shareholders in closely held corporations and who “happen” to have transferred shares in those corporations to trusts while forbidding the trustee to sell the shares without approval and while retaining voting rights in those shares. Therefore the rule ought not to be “modified” by this Court.

A

The argument, apparently, is that what “appear [s] to be established” should become established because it has appeared. Judges can and will properly differ on the questions of what deference to accord reliance on a well-established rule, but I doubt that we are precluded from reaching what would otherwise be a right result simply because in the minds of some litigants a contrary rule had heretofore “appear [ed] to be established.” If the majority’s approach were widely accepted, artful claims of past understanding would consistently suffice to frustrate judicial as well as administrative efforts at present rationalization of the law and every precedent— even at the tax court level — might lay claim to such authority that the Government and the tax bar could afford to leave no case unappealed.

B

Of course, the reliance argument is doubly infirm if the majority’s rule cannot be said to have “appear[ed] to be established.” Did Byrum have a sound basis for calculating that there was no substantial risk of taxation when he persisted in retaining the powers and privileges described above?

*1641. Again the majority turns to Reinecke v. Northern Trust Co., but it is no more credible to use Northern Trust as a foundation for Byrum’s § 2036 (a) (2) position than it was to use it as a basis for the Court’s § 2036 (a)(1) argument. Northern Trust was decided on January 2, 1929, two years and three months before Congress passed the first version of §2036 (a)(2). Section 402 (c) of the Act of 1921, the provision under which Northern Trust was decided, proscribed only transfers in which the settlor attempted to retain “possession or enjoyment” until his death. It is thus not surprising that Northern Trust focused on the question of the settlor’s “power to recall the property and of control over it for his own benefit,” 278 U. S., at 347 (emphasis added), and made no mention of possible tax liability because of a retained power to designate which beneficiaries would enjoy the trust income. A holding in this context cannot be precedent of “weight” for a decision as to the efficacy of a trust agreement made — as this trust was — 27 years after the predecessor of § 2036 (a)(2) was enacted.

I note also that Northern Trust rests on a conceptual framework now rejected in modern law. The case is the elder sibling of May v. Heiner, 281 U. S. 238, a three-page 1930 decision which quotes Northern Trust, at length. May in effect held that under § 402 (c) a settlor may be considered to have fully alienated property from himself even if he retains the very substantial string of the right to income from the property so long as he survives. The logic of May v. Heiner is the logic of Northern Trust. As one authority has written:

“When retention of a life estate was not taxable under the rule of May v. Heiner, it followed that mere retention of a right to designate the persons to receive the income during the life of the settlor was not taxable . . . .” 1 J. Beveridge, Law of Federal Estate Taxation §8.06, p. 324 (1956).

*165That logic no longer survives. When three Supreme Court per curiams affirmed May on March 2, 1931, and thus indicated that this view would not be confined to its facts, the Treasury Department, on the next morning, wrote Congress imploring it to promptly and finally reject the Court’s lenient view of the estate tax system. Congress responded by enacting the predecessor of § 2036 (a) (2) that very day. The President signed the law that evening. Thus the holding of May and the underlying approach of Northern Trust have no present life. I note further that though Congress has refused to permit pre-1931 trusts to be liable to a rule other than that of May, in 1949 this Court itself came to the conclusion that May was wrong, and effected “a complete rejection” of its reasoning. Commissioner v. Estate of Church, 335 U. S. 632,12 645.

I seriously doubt that one could have confidently relied on Reinecke v. Northern Trust Co. when Byrum drafted his trust agreement in 1958. This Court is certainly not bound by its logic, in 1972. I do not mean any disrespect, but as Mr. Justice Cardozo said about another case, Northern Trust is a decision “as mouldy as the grave from which counsel . . . brought it forth to face the light of a new age.” B. Cardozo, The Growth of the Law, in Selected Writings 244 (M. Hall ed. 1947).

2. The majority argues that there were several lower court cases decided after the enactment of § 2036 (a) (2) *166upon which Byrum was entitled to rely, and it is quite true that cases exist holding that a settlor’s retention of the power to invest the assets of a trust does not by itself render the trust taxable under § 2036 (a) (2). But the majority’s emphasis on these cases as a proper foundation for Byrum’s reliance is doubly wrong. First, it could not have evaded Byrum’s attention and should not escape the majority that all cited prior cases — save King (the tax court case written four years after Byrum structured his trust) — -involved retention of power to invest by the settlor’s appointment of himself as a trustee; that is, they posed instances- in which the set-tlor’s retained power was constrained by a fiduciary obligation to treat the life tenant beneficiaries and re-mainderman beneficiaries exactly as specified in the trust instrument. Thus, the "freedom” to reallocate income between life tenants and remaindermen by, e. g., investing in wasting assets with a high present return and no long-term value, was limited by a judicially enforceable strict standard capable of invocation by the trust beneficiaries by reference to the terms of the trust agreement. See Jennings v. Smith, 161 F. 2d 74 (CA2 1947), the leading case. Byrum must have realized that the situation he was structuring was quite different. By according himself power of control over the trust income by an indirect means, he kept himself quite free of a fiduciary obligation measured by an ascertainable standard in the trust agreement. Putting aside the question of whether the situation described should be distinguished from Byrum’s scheme, surely it must be acknowledged that there was an apparent risk that these situations could be distinguished by reviewing courts.

Second, the majority’s analysis of the case law skips over the uncertainty at the time Byrum was drafting his trust agreement about even the general rule that a settlor could retain control over a trust’s investments *167if he bound himself as a trustee to an ascertainable method of income distribution. While Byrum and his lawyer were pondering the terms of the trust agreement now in litigation, the Court of Appeals for the First Circuit was reconsidering whether a settlor could retain power over his trust’s investments even when he bound himself to a fiduciary’s strictest standards of disinterested obligation to his trust’s beneficiaries. Early in 1958 the United States District Court for the District of Massachusetts had ruled that a settlor could not maintain powers of management of a trust even as a trustee without assuming estate tax liability. State Street Trust Co. v. United States, 160 F. Supp. 877. The estate’s executor appealed this decision and argued it before the First Circuit panel on October 7, 1958. Byrum’s trust agreement was made amidst this litigation, on December 8. On January 23, 1959, the First Circuit affirmed the District Court. 263 F. 2d 635.13

The point is not simply that Byrum was on notice that he risked taxability by retaining the powers he retained when he created his trust — though that is true. It is also that within a month of the trust’s creation it should have been crystal clear that Byrum ran a substantial risk of taxation by continued retention of control over the trust’s stock. Any retained right can be resigned. That Byrum persisted in holding these rights can only be viewed as an indication of the value he placed upon their enjoyment, and of the tax risk he was willing to assume in order to retain control.

The perception that a settlor ran substantial risk of estate tax if he insisted on retaining power over the *168flow of trust income is hardly some subtle divination of a latter-day observer of the 1958-1959 tax landscape. Contemporary observers saw the same thing. A summary of the field in the 1959 Tax Law Review concluded: “Until the law is made more definite, a grantor who retains any management powers is proceeding at his own risk. . . . [T]here is no certainty. . . Gray & Covey, State Street — A Case Study of Sections 2036 (a)(2) and 2038, 15 Tax L. Rev. 75, 102. The relevant subcommittee of the American Bar Association Committee on Estate and Tax Planning hardly thought reliance appropriate. It wrote in 1960 that:

“The tax-wise draftsman must now undertake to review every living trust in his office intended to be excluded from the settlor’s estate in which the settlor acts as a trustee with authority to:

“1. Exercise broad and virtually unlimited investment powers . . . .” Tax-Wise Drafting of Fiduciary Powers, 4 Tax Counsellor’s Quarterly 333, 336.

More could be said, but I think it is clear that the majority should find no solace in its reliance argument.

V

The majority, I repeat, has erred in every substantial respect. It remains only to note that if it is1 wrong in any substantial respect — i. e., either in its § 2036 (a)(1) or (a)(2) arguments — Byrum’s trust is by law liable to taxation.