NYSBA Tax Section Recommends Repeal of Preferential Dividend Rule

Dear Sirs:

I am writing on behalf of the Tax Section of the New York State Bar Association to recommend the repeal of Section 562(c) of the Internal Revenue Code as it applies to regulated investment companies (RICs) and real estate investment trusts (REITs). This 72-year-old provision, known as the "preferential dividend rule," no longer furthers its original purpose and complying with it significantly burdens RICs and REITs. For these reasons, we recommend that it be repealed as it applies to RICs and REITs. We also recommend administrative guidance that would permit RICs and REITs to correct inadvertent violations of the rule.

By way of background, RICs and REITs must satisfy a distribution requirement and otherwise rely on the "dividends paid deduction" to avoid a corporate-level tax. However, this deduction is not available for a "preferential" dividend, which generally includes any non-pro rata dividend within a class of shares or any dividend that violates a preference to which a class is entitled.

Paying a preferential dividend can therefore have dire consequences. A preferential dividend can prevent a RIC or REIT from satisfying its distribution requirement and cause it to become subject to corporate taxation on all of its income. Even a small, inadvertent error can potentially cause a dividend to be preferential and, if not remedied, result in corporate taxation. Moreover, there are currently no remedies for violations -- even inadvertent violations -- of the prohibition on preferential dividends that do not involve significant penalties and double taxation.

The enclosed report explains why the preferential dividend rule no longer effectively fulfills its purpose, why other rules more effectively fulfill that purpose and how complying with the preferential dividend rule significantly burdens RICs and REITs. The report also outlines the rule's history and current application.

In view of the rule's ineffectiveness and burdens, the report recommends that it be repealed. Until it is repealed or reformed, the report recommends that administrative guidance be issued permitting RICs and REITs to correct inadvertent violations of the rule.

We appreciate your consideration of our recommendations. We would be pleased to discuss these matters with you further or provide any other assistance that you would find helpful.

Regulated investment companies ("RICs") and real estate investment trusts ("REITs")² are generally taxed as quasi-pass-through vehicles. The principal mechanism for providing this pass-through treatment is the dividends paid deduction.³ So long as a RIC or REIT distributes a sufficient amount of its income (other than net capital gain),⁴ the amount of its ordinary income and capital gain otherwise subject to tax is reduced by a deduction for its dividends paid.⁵ Its income so distributed is not subject to a corporate level of tax and is subject to tax only at the shareholder level.

Section 562(c) limits the type of dividends that are eligible for the dividends paid deduction. Under Section 562(c), distributions that are "preferential" are not treated as distributed for purposes of the distribution requirement and are not deductible for purposes of determining tax liability (the "Preferential Dividend Rule"). In particular, Section 562(c) provides:

This Report considers the public policy goals intended and achieved by applying the Preferential Dividend Rule to RICs and REITs⁶ and compares those public benefits to the burdens that result from the imposition of the rule on these entities. The rule's primary intended benefit is the prevention of tax avoidance, though protecting RIC investors from unfair expense allocations is another purpose that has been attributed to it. Burdens resulting from the rule include taxpayer compliance costs, government enforcement and administrative costs, the unfairness that results from its violations (and from the remedy of these violations) and the economic inefficiencies that result from foregone business opportunities because of the rule. These costs are magnified by the rule's ambiguity (particularly the unfairness and inefficiencies that result from different funds taking different positions with respect to its meaning).

The Report concludes that applying the rule to RICs and REITs only modestly furthers (if at all) any tax or other public policy goal, is unnecessary (because other better means exist to further these goals) and significantly burdens RICs and REITs and their investors. In view of these conclusions, we recommend the following:

• Repeal the Preferential Dividend Rule as it applies to RICs and REITs.⁷

• Until the Preferential Dividend Rule is repealed, issue administrative guidance clarifying that (1) if two classes of shares have different distribution, redemption or liquidation rights, each such class of shares will be respected as a separate class of shares for purposes of the Preferential Dividend Rule and (2) actual or constructive distributions made in accordance with those rights will not be treated as a violation of the Preferential Dividend Rule. We also recommend that this guidance include a simple and easy means of curing inadvertent violations of the Preferential Dividend Rule.

• If it is determined that the tax law should be used to regulate the allocation of certain fees among investors in a RIC or REIT, replace the Preferential Dividend Rule as it applies to RICs and REITs with a new condition for RIC and REIT qualification that requires the burden of those fees to be allocated among investors in a RIC or REIT in accordance with that public policy objective. If this approach is adopted, we also recommend that the new qualification requirement include a simple and easy means for curing inadvertent noncompliance.

II. CONTINUING PURPOSE OF THE RULE

In the 70-plus year history of the Preferential Dividend Rule, only two purposes have been attributed to it: (1) the prevention of tax avoidance and (2) shareholder fairness. However, in our view, the rule does not effectively fulfill either of these purposes, which are now furthered more effectively by other means.

Today, however, the tax law is better equipped than in 1936 to prevent income shifting among shareholders. The dividend waiver rules disregard waivers that would otherwise shift income based on shareholder-level relationships,⁸ and Sections 305(b) and (c) are designed to prevent a corporation from making taxable distributions to one group of shareholders while allowing another group to defer the recognition of their share of the corporation's earnings and profits by means of an actual stock dividend or some other change in legal rights treated as a constructive stock dividend.⁹

By contrast, the Preferential Dividend Rule is largely ineffective against planned tax avoidance achieved by shifting income among shareholders. This is because the rule has no effect on dividends paid on different classes of stock so long as those dividends are paid in accordance with the preferences inherent in the terms of those classes. In fact, the rule actually requires dividends be so paid. By its terms, the Preferential Dividend Rule really only applies in the case of sloppy tax planning or inadvertent mistakes resulting in either non-pro rata distributions within a single class or distributions failing to conform to class preferences.

This is why the Preferential Dividend Rule provided the government with no protection against so-called "fast pay stock" transactions, which were one of the only historical examples of RICs and REITs being used to shift income to tax indifferent parties.¹⁰ The Internal Revenue Service (the "IRS") decided not to use Section 305 to attack these transactions¹¹ and instead issued regulations under Section 7701(1), which authorizes regulations to recharacterize multi-party financing arrangements in order to prevent tax avoidance.¹²

More importantly, the provisions of the rule itself do not address fairness issues. The rule does not (1) limit the type or amount of preferences to distributions, so long as those preferences are inherent within the terms of a class of shares entitled to such preferences, (2) limit class or share preferences that are unrelated to actual or constructive distributions (like redemption or liquidation preferences)¹³ or (3) impose the types of restrictions that are thought actually to promote investor protection, like independent oversight or mandatory disclosure regarding preferences that a fund has adopted in accordance with the rule.¹⁴

When the first version of the Preferential Dividend Rule was enacted in 1936, there may have been more reason to use the tax law to promote shareholder fairness (since the securities laws were in their infancy),¹⁵ but now the securities laws are well developed and the Securities and Exchange Commission (the "SEC") has expertise and resources in addressing these issues that the IRS lacks. Today, the SEC is much better equipped than is the IRS to regulate fairness issues. The SEC's Division of Investment Management has specific responsibility for administering the Investment Company Act of 1940 (as amended, the "1940 Act"),¹⁶ the principal statute governing the organization and operations of investment companies (and thus virtually RICs), as well as the Investment Advisers Act of 1940, which is a parallel law to the 1940 Act that regulates the activities of investment managers. The IRS has no professionals that specialize in understanding these issues.

Through its Division of Investment Management, the SEC has taken very seriously its charge under the 1940 Act to regulate investment companies in the public interest and for the protection of investors. In addition to issuing rules and providing interpretive guidance,¹⁷ the SEC has conducted extensive studies of the industry,¹⁸ including reports that became the basis of legislative reform strengthening investor protection.¹⁹

Throughout these regulatory and legislative developments, the appropriateness of fees borne by shareholders has been of particular concern.²⁰ For example, as a result of recent rule changes, registered fund boards are now required to disclose to investors the material reasons for the approval of advisory contracts in annual reports by explaining in detail why the board selected the investment adviser and approved the advisory fee (and any other amounts payable by the fund under the contract).²¹

In general, the SEC has determined that (1) investor fairness is best achieved through disclosure to investors, oversight by independent directors and regulators and matching investor sophistication with level of protection and (2) it is undesirable to impose ineffective rules because they result in unnecessary burdens that harm investors by stifling competition.²² By contrast, the Preferential Dividend Rule is not well designed to provide protection or ensure fairness with respect to distributions or allocations of fees, at least not by the standards of the SEC. The Preferential Dividend Rule provides no disclosure requirements (relating to the nature of preferences among the classes or otherwise), ensures no oversight by independent parties as to the appropriateness or fairness of fees and does not attempt to match levels of shareholder protection with the shareholders' need for such protection. Instead, the Preferential Dividend Rule applies in a binary fashion based solely on whether or not distributions are pro rata within a class and whether or not distributions among classes are in accordance with their terms.

In addition, the Preferential Dividend Rule does not shield investors suffering unfairness from the consequences that follow from violations of the rule. Instead of requiring those investors to be made whole from any resulting harm, the rule operates to cause all shareholders to bear the tax cost of its violation, including those shareholders suffering the unfairness that leads to the violation. Moreover, to the extent that complying with the Preferential Dividend Rule results in additional costs to RICs and their shareholders (as discussed below) without furthering a public policy goal, its application to RICs is in fact inconsistent with the SEC's objective of eliminating unnecessary rules, which tend to reduce competition and thus harm investors.

In the specific context of fee allocations, the Preferential Dividend Rule actually undermines, in some respects, safeguards provided under the 1940 Act. For example, one of the means by which the 1940 Act provides investor protection is by requiring the independent trustees to approve certain fund fees and rebates of fees to shareholders. Since rebates would not violate the Preferential Dividend Rule so long as the manager is not acting as agent for the fund in providing the rebate, the Preferential Dividend Rule discourages the disclosure of these rebates to the funds' trustees. In addition, the securities law rules generally favor correcting errors by making affected shareholders whole. The Preferential Dividend Rule disfavors this solution because such corrective action is itself susceptible to characterization as a preferential dividend (since it benefits only harmed shareholders). Moreover, because paying a deficiency dividend (or other "corrective" dividend) can be enormously expensive, the Preferential Dividend Rule puts pressure on RICs to conclude that these errors are not in fact preferential dividends.²³

It might be tempting to think concerns about investor fairness are more appropriate in the REIT context, since REITs are subject to less regulation than RICs. While that observation is perhaps true, it does not follow, however, that the Preferential Dividend Rule is intended to provide any protection to REIT investors or that it is well designed to do so. For all the reasons that the Preferential Dividend Rule is an ineffective means of promoting investor protection for RIC shareholders (including under the standards used by the SEC), it is also not well designed to provide such protection to REIT shareholders. If it is determined that REIT investors need greater protection than is provided currently under the applicable laws, those laws should be changed to provide it.

If the Preferential Dividend Rule is in fact intended to protect investors from unfair fee allocations or otherwise, we recommend that it be replaced as it applies to RICs and REITs with a qualification requirement that achieves that result and is consistent with applicable securities laws. We would be pleased to provide a more detailed recommendation for such a change if that would be useful.

III. ORIGINAL PURPOSE OF THE RULE

The Preferential Dividend Rule was initially intended to prevent the avoidance of tax on corporate income. We do not believe that it has a purpose related to the protection of shareholders from unfair fee allocations (or more generally).

Congress seems to have believed that in the 1930s the Preferential Dividend Rule was needed to prevent corporations otherwise subject to these taxes from avoiding them by making disproportionate distributions to low-bracket taxpayers. In the context of the accumulated earnings tax. at least one court argued that there was an implicit Congressional disapproval of non-pro rata dividends "as a means of evading tax" even before the first version of the Preferential Dividend Rule was enacted in 1936.³⁰ At that time, the tax law recognized that non-pro rata distributions could be taxed as dividends to only the shareholders receiving them (instead of treated as paid to all shareholders pro rata and then reallocated based on the shareholders' relationships).³¹ Citing this case law, a Sixth Circuit Court flatly stated, in the context of a case involving the accumulated earnings tax, "the purpose of . . . the so-called preferential distribution provision [in the 1936 Tax Act] . . . was to prevent special dividends to stockholders in the lower income tax brackets."³² The IRS has also acknowledged that the 1936 version of the Preferential Dividend Rule was intended principally to prevent corporations from paying non-pro rata dividends as a way of shifting their income to shareholders in lower tax brackets and thereby undermining the anti-tax-avoidance purposes of the personal holding company tax and accumulated earnings tax.³³

While the legislative history of the 1936 Tax Act is silent on the purpose of its version of the Preferential Dividend Rule, in connection with the 1938 restatement of the rule, Congress provided the following explanation, confirming the rule's anti-tax-avoidance purpose:

In 1936, it was much easier than today to accomplish this kind of income shifting -- for example, by dividend waivers or nontaxable stock dividends -- without there being an economic detriment to the shareholder receiving less than a pro rata share of dividends. This is because, at that time, other tax rules had not yet been developed to prevent income shifting among shareholders. It was not until the 1950s that the IRS issued rulings indicating that dividend waivers would not be respected for tax purposes if there are family or direct business relationships between the waiving and non-waiving shareholders.³⁵ In addition, the taxation of stock dividends was extremely unclear until 1954, when Congress defined which stock dividends would be taxable.³⁶ The 1936 Tax Act simply provided for their taxation to the extent constitutionally permitted.³⁷ Rather than attempting to prevent income shifting among shareholders, the 1936 Tax Act's taxation of stock dividends (to the extent constitutionally allowed) was intended to prevent taxpayers from claiming full basis in their untaxed stock dividends when those shares were sold.³⁸ Even after the taxation of stock dividends was clarified in 1954,³⁹ it was not until the Tax Reform Act of 1969⁴⁰ that Congress seriously tried to restrict income shifting among shareholders by treating as taxable stock dividends certain types of increases in relative interests in a corporation.⁴¹

In summary, based on this history, we believe that in 1936 Congress was very concerned about the ability of shareholders to deflect and defer the recognition of corporate income. Congress seems to have believed that this type of tax avoidance was best prevented by subjecting corporate income to additional taxes unless distributed pro rata within a class (and in accordance with any preferences as among classes). For these reasons, we do not believe that the 1938 legislative history should be interpreted to mean that the Preferential Dividend Rule was intended to protect shareholders from the pre-tax economic detriment of receiving less than their pro rata share of a dividend (or less than the amount specified by a class preference).

While Congress believed that requiring pro rata distributions within a class helped prevent tax avoidance, the 1938 committee report also indicates that "minor differences in valuations of property distributed" should not cause the distribution to be treated as "preferential" so long as shareholders are treated "with substantial impartiality and in a manner consistent with their rights under their stock-holding interests."⁴² We believe that this history means that, for a distribution to be deductible, generally shareholders must be treated the same with respect to that distribution and that minor differences in distributions among shares are permitted so long as those difference result from treatment that is the same among all shareholders. Again, we do not believe that the reference to "substantial impartiality" should be interpreted to mean that the Preferential Dividend Rule is intended to promote or reward shareholder fairness for the sake of fairness.

For example, on the basis of this legislative history, New York Stocks, Inc.⁴³ holds that a dividends paid credit for the net income attributable to shares redeemed by an open-end investment fund⁴⁴ is not a "preferential distribution" (and is therefore deductible) even though the amount treated as distributed is not pro rata among all shares within a class.⁴⁵ In its decision, the court emphasized that each shareholder had an "equal opportunity to redeem" shares, by reason of the fund's organization as an open-end investment company, and concluded that the distribution is not preferential because the distribution "is open to all shareholders 'with substantial impartiality.'"⁴⁶ Allowing the credit in this circumstance ensures that redeeming shareholders are treated, for purposes of the dividends paid credit, as receiving their share of corporate income, that non-redeeming shareholders are not distributed more than their share of corporate income, that a shareholder's decision to redeem does not shift income recognition to others and that, because all holders have the same rights to redeem, they all have equal ability to affect the timing and amount of income treated as distributed.⁴⁷ The opinion thus implies that, because differences in the amount and timing of the shareholders' recognition of that income is determined by their exercise of rights equally available to all shareholders, there is less likelihood of income deflection by some shareholders.

Prior cases had concluded that corporations without shares redeemable on demand are not permitted to deduct earnings attributable to redeemed shares because such distributions of earnings are treated as preferential.⁴⁸ The opinion in New York Stocks, Inc. distinguishes those cases by noting that they "follow a consistent pattern of special advantages offered certain shareholders and denied others." Such "special deals" were presumably thought to enhance the risk of tax avoidance by potentially allowing some shareholders to deflect more easily the recognition of their share of corporate income. For example, allowing a dividends paid credit in respect of open market redemptions would reduce the amount distributable to remaining shareholders, and such reduction would exceed the remaining shareholder's allocable share of earnings if the redemption price discounts undistributed corporate earnings.

MICs were introduced because Congress believed that shareholders of passive investment funds, whose "modest means" generally otherwise excluded them from investment products providing professional money management and risk diversification,⁵¹ should bear only one layer of tax on their investment income, as was the case for large investors, whose wealth generally enabled them to obtain professional management and diversification through individually managed accounts.⁵² It is less clear to what extent Congress intended the tax rules applicable to MICs to further other public policy goals, like protecting investors from unfair or risky business practices of investment fund managers and sponsors. It would seem reasonable to believe that, at least to some extent, concerns about investor protection might have informed some aspects of how MICs were taxed,⁵³ since in 1936 investment funds were largely unregulated by the federal securities laws⁵⁴ and since at that time many of the abuses associated with these unregulated investment funds were recognized.⁵⁵ For example, a "closed-end" fund, which was the type of investment fund most widely abused during the 1920s and 1930s,⁵⁶ was not allowed to be a MIC under the 1936 Tax Act.⁵⁷ However, there is no legislative history that specifically ties the tax rules applicable to MICs (including the 1936 version of the Preferential Dividend Rule) to particular concerns about investor protection, other than general statements about how investment funds allow diversification of risk for small investors (which is perhaps one of the reasons why MICs, like RICs today, must meet an asset diversification test).⁵⁸

Four years after introducing MICs, Congress passed the 1940 Act, a watershed for the regulation of investment funds resulting from an exhaustive study by the SEC.⁵⁹ Intended specifically to protect investors in investment funds, the 1940 Act⁶⁰ supplanted whatever ambiguous function the tax law was intended to serve in promoting investor protection by providing favorable tax treatment to MICs.⁶¹

For example, among the many provisions of the 1940 Act designed to provide this protection is Section 18, which limits the ability of an investment company to issue "senior securities." As developed administratively, this provision in effect requires, in the absence of specific relief, shareholders of an open-end fund generally to be treated identically, and it severely limits the ability of a closed-end fund to issue shares with any preferences.⁶² In this respect, Section 18 of the 1940 Act is much broader than the Preferential Dividend Rule, since it applies to all material shareholder rights, not only those relating to distributions, and thus in effect does not allow a fund to have different classes of shares.⁶³ By contrast, the Preferential Dividend Rule requires distributions among classes to differ in accordance with class preferences, whether or not those preferences are "fair."

In 1942, Congress replaced MICs with RICs, which generally included only investment companies that were registered under the 1940 Act and satisfied certain qualification requirements. These requirements were similar to, but substantially pared down from, those applicable to MICs, and their basic framework survives largely intact today.⁶⁴ These changes seemed to have been an acknowledgement that, with the passage of the 1940 Act, there was less need to rely on tax incentives as a means of regulating investment funds,⁶⁵ though the legislative history to the 1942 Tax Act does not indicate how or whether the remaining qualification requirements were intended to promote investor fairness or protection. In testimony before Congress, the Treasury has taken position that investor protection for RIC shareholders is better implemented through the securities laws rather than the tax laws.⁶⁶

Under the 1942 Tax Act, RICs were taxed like MICs in that they were allowed to determine the amount of income subject to tax after applying the dividends paid credit and that credit continued to be limited by the version of the Preferential Dividend Rule then in effect.⁶⁷ Again, however, the legislative history of the 1942 Tax Act does not indicate how the Preferential Dividend Rule related, if at all, to any tax or other public policy goal specific to RICs.

In 1960, REITs were introduced by the new Section 857.⁶⁸ Like RICs, a REIT is allowed the dividends paid deduction (subject to the Preferential Dividend Rule)⁶⁹ provided that it meets a distribution requirement that also is measured by reference to the dividends paid deduction. ⁷⁰ This aspect of REIT taxation, like RIC taxation, has remained unchanged. There is no indication in any of the history to these provisions indicating a purpose of the Preferential Dividend Rule that is specific to REITs, suggesting that the Preferential Dividend Rule serves the same purpose for REITs as for RICs. There is also no indication that the requirements for REIT qualification were intended to promote investor protection or fairness. The REIT provisions were added to "provide[] substantially the same tax treatment for real estate investment trusts as present law provides for regulated investment companies"⁷¹ in order to "accord to individuals of small means an opportunity to pool their investments in one of these companies, yet receive the same treatment as those of greater wealth can obtain by direct investments."⁷²

The Preferential Dividend Rule was most recently amended in 1986. This amendment permits a non-pro rata increase in distributions to shareholders of a RIC making an initial investment of at least $10 million so long as such increase is solely by reason of administrative cost savings.⁷³ The legislative history of this amendment explains that "differences in the rate of dividends paid to shareholders are not treated as preferential dividends (within the meaning of Section 562(c)), where the differences reflect savings in administrative costs (but not differences in management fees), provided that such dividends are paid by a RIC to shareholders who have made initial investments of at least $10 million."⁷⁴

By its terms, this amendment relaxes the Preferential Dividend Rule by allowing distributions among shares within a single class to differ based on certain differences in expenses. Some in the IRS appear to believe, however, that the reference to "management fees" in this legislative history should be interpreted to mean that Congress intends that any differences in distributions based on differences in management fees should be treated as preferential (and thus not deductible), even if such differences are required pursuant to differences in terms of different classes of shares. For example, Announcement 96-95 states, "The legislative history to the 1986 amendment to section 562(c) explains that any difference in the investment advisory fee charged to shares of a RIC results in a preference."⁷⁵ The IRS has not otherwise asserted this position in published guidence, though informally IRS officials have been very direct that they believe that the Preferential Dividend Rule prohibits non-pro rata allocations of management-related expenses.⁷⁶ The rationale for this position appears to be based on the belief that Congress intended the Preferential Dividend Rule to prevent fee allocations resulting in some shareholders unfairly subsidizing others.⁷⁷ It is unclear whether the IRS believes that, based on this legislative history, REITs are precluded from allocating management fees disproportionately among classes of shares; there is no indication in any guidance that it does.

We do not believe that the 1986 amendment to the Preferential Dividend Rule should be interpreted as expanding the rule and we do not believe that the legislative history to this amendment implies that Congress intended such an expansion. Rules issued under the 1940 Act explicitly prohibit registered investment companies (and thus virtually all categories of RICs) from allocating management fees differently among different classes of shares (or differently among shares of a single class).⁷⁸ Given that explicit prohibition under the securities laws, we believe that, if Congress had intended to expand the Preferential Dividend Rule to restrict such allocations, it would have done so directly and explicitly in the statutory language of the 1986 amendment and not indirectly through ambiguous language included in the legislative history to that amendment. We also believe, for the reasons indicated above, that the Preferential Dividend Rule is not intended to promote shareholder fairness and that, even if it were, it should not be relied on to do so because it is ineffective. The SEC and securities laws are better suited than the IRS and tax law to regulate the allocation of fees among investors in an investment fund. As discussed below, interpreting the Preferential Dividend Rule in a manner that prohibits certain types of fee arrangements materially burdens RICs, REITs and their shareholders.

IV. THE PREFERENTIAL DIVIDEND RULE UNDER CURRENT LAW

The substantive law under Section 562(c) has changed very little since its predecessor was first enacted. However, there have been a number of developments in its administration, in particular, as it applies to RICs with multiple classes of common shares. The IRS's current administration of the rule seems to be based on the legal conclusion that the rule is intended to regulate the allocation of certain fees, including among classes of RIC shares. This approach has led to greater ambiguity in the rule's application, contributing to its resulting burdens.

Because amounts treated as constructively received as dividends under Section 316 are dividends for purposes of Section 561, they are included in the dividends paid deduction and subject to the Preferential Dividend Rule.⁸⁶ However, preferences among shareholders that do not constitute actual or constructive dividends (or some other form of payment described in Section 561) are not subject to Section 562(c).

For example, Section 562(c) places no limitation on the ability of a corporation to incur or deduct expenses that are attributable to only some shareholders (or benefit only some shareholders) provided the corporation's payment of them does not result in a constructive distribution. In particular, expenses that are deductible as "ordinary and necessary" for the corporation's business will not be treated as a constructive dividend simply because the expenses incidentally benefit some shareholders more than others.⁸⁷ Consistent with this principle, when a RIC or REIT recognizes an "excess inclusion" from an investment in a residual interest in a real estate mortgage investment conduit, any resulting tax owed by the RIC or REIT⁸⁸ can be deducted by it (to reduce the amount otherwise available for distribution to shareholders) and the burden of that tax can be allocated specifically to those disqualified organization shareholders, without giving rise to a constructive dividend or a preferential dividend.⁸⁹ Similarly, the IRS has ruled privately that, where the SEC required a RIC advisor to pay to the RIC shareholders fees illegally received by it, those payments would not constitute income to the RIC or preferential dividends to the shareholders under Section 562(c) and would not affect the company's RIC status.⁹⁰

Equalization Payments are necessarily non-pro rata (unless made in liquidation of the corporation), since they are made to only redeeming shareholders (and would not be treated as exchanges under Section 302 if they were pro rata). An example in the regulations under Section 562(c) denies a deduction for a non-pro-rata redemption but does not specify whether the tender was made equally available or whether it was taxed as an exchange to the shareholders.⁹⁵ Equalization Payments have also been held to be non-deductible under Section 562(c) where the redemption right is not made available to all shareholders⁹⁶ and in the case of non-RIC corporations even if the redemption is equally available to all shareholders of the class being redeemed.⁹⁷ As discussed above, however, in the case of open-end RICs, where the ability to redeem on demand is made available to all shareholders, courts have concluded that Section 562(c) does not apply to deny the deductibility of these amounts.⁹⁸

Similarly, the IRS has ruled privately in the context of a fund-of-funds arrangement that redemptions treated as Section 301 distributions from an open-end RIC are not preferential under Section 562(c).⁹⁹ The ruling is apparently based on the rationale that these Section 301 distributions are not preferential because the right to redeem is equally 100 available to all shareholders at all times and redemptions are not tax motivated.¹⁰⁰

The only published ruling addressing non-pro rata dividends within a single class of stock relates to stock dividends issued at a discount. In Rev. Rul. 83-117,¹⁰¹ a REIT gave all shareholders an election to receive a dividend in the form of cash or stock and shareholders electing stock received an amount of stock that was discounted to reflect underwriting and other cost savings available from issuing the shares directly. The ruling holds mat a discount of up to 5% from the fair market value of the stock in these circumstances does not cause the dividend to be treated as preferential under Section 562(c) and that a greater percentage discount results in a preferential dividend. The ruling provides no analysis of the law supporting this conclusion.

If any portion of a distribution is preferential under Section 562(c), no portion of the distribution is deductible.¹⁰² For example, if a dividend on preferred and common classes is paid but the preferred receives less than the full amount to which it is entitled, the amount paid to the preferred and to which the preferred is entitled is treated as preferential (as well as the amount paid to the common).¹⁰³ There is little law bearing on the question as to how the scope of a distribution should be determined. The case law addressing this issue treats all distributions made pursuant to a single dividend declaration as a single distribution.¹⁰⁴

The IRS has issued a revenue procedure stating under what conditions it will not treat dividends paid by these RICs as preferential dividends under Section 562(c).¹⁰⁷ This revenue procedure is based on Rule 18f-3,¹⁰⁸ issued by the SEC in 1996,¹⁰⁹ which provides exemptive relief for certain of these types of arrangements otherwise prohibited under Section 18 of the 1940 Act.¹¹⁰

Rule 18f-3 conditions its relief on the satisfaction of three requirements.¹¹³ First, the classes must differ either in the manner in which they distribute their securities, in the services they provide to their shareholders, or both,¹¹⁴ and expenses relating to the distribution of a class's shares or to services provided to the shareholders of that class must be allocated to that class. Second, the investment advisory fee charge to each class generally must be the same percentage amount. (However, in the case of a multiple class fund with an advisory contract that provides for compensation to the adviser on the basis of performance, the percentage amount may vary for each class to the extent mat any difference is the result of the application of the same performance fee provisions to the different investment performance of each class.) Third, other expenses must be either allocated to different classes in different amounts to the extent that they are incurred by one class in a different amount or they must reflect differences in the amount or kind of services that different classes receive.¹¹⁵

Rule 18f-3(b) expressly allows a fund's underwriter, adviser, or other provider of services to waive or reimburse the expenses of a specific class or classes. In adopting this rule, the SEC noted that reimbursements or waivers should not become de facto modifications of the fees provided for in advisory or other contracts so as to provide a means for cross-subsidization between or among classes.¹¹⁶

To rely on the revenue procedure,¹¹⁹ each "qualified group of shares" of a RIC (1) must have a different arrangement of shareholder servicing or distribution or both and be allocated and pay the fees and expenses of that arrangement,¹²⁰ (2) may be allocated and may pay other fees and expenses (other than management-related fees) if actually incurred in different amounts or if the qualified groups receive services of a different kind or degree¹²¹ and (3) must be allocated all expenses relating to the management of the assets pro rata in accordance with relative net asset value (except to the extent differences result from the application of a performance fee provision to the different performances of different qualified groups).¹²²

One of the curious aspects of Rev. Proc. 99-40 is that it appears to require dividends to be paid in a manner that is inconsistent, at least in one respect, with how income is required to be allocated among classes under Rule 18f-3. Rule 18f-3 requires income to be allocated among classes based on the classes' relative net asset values. However, Rev. Proc. 99-40 requires that distributions vary among groups of shares only as a result of permitted differences in fee allocations. This means that a portion of income allocated to a class under Rule 18f-3 because of its relatively lower expenses could be paid as a dividend to a class with the relatively higher expenses. This slight inconsistency between the Preferential Dividend Rule and Rule 18f-3 illustrates the challenge faced by the IRS in trying to use the Preferential Dividend Rule to enforce its views regarding shareholder fairness and how even seemingly simple rules can lead to ambiguity and mistakes in their day-to-day application.

It can be argued, however, that the differences in distribution rights resulting from differences in class-specific expenses are simply "differences" and do not constitute a "preference" as that term is used in Section 562(c). If that argument is correct, those differences in distributions would violate the Preferential Dividend Rule even if they are attributable to rights inherent in the RIC's different classes of shares. This concern arises because the aggregate differences in distributions among classes distinguished on the basis of different class-specific expenses do not result from a "priority" in distributions of one class over another. Such differences are attributable only to differences in the class-specific expenses allocated to each class. It is true that some classes have a right to bear lower expenses than other classes. It is also true that, in case of class-specific expenses, the amount of dividends paid on each class is reduced by the class-specific expenses allocated to that class. However, it does not follow from these facts that the class with the lowest class-specific expenses is entitled to the highest level of dividends. This is because changes in the relative size of each class affects the relative amounts of a dividend paid on each class (and the per share dividend of each class). This relationship between changes in relative class size and changes in relative dividend amounts among classes exists because, for purposes of determining the amount of dividends paid on each class, gross income and non-class-specific expenses are allocated based on relative shares outstanding at the time that the dividend is paid, and class-specific expenses are allocated only to the class bearing that expense.

For example, assume a fund has two classes of shares (each with 10 shares outstanding), one class that bears class-specific expenses equal to 0.20% of its net assets and one that bears class-specific expenses equal to 0.30% of its net assets, and that the fund's other expenses equal 1% of net assets. Assume that at the beginning of the year, each class is worth $100, that the fund earns $20 of taxable income for the year and that at the end of the year there is no change in the net unrealized gain or loss in the fund for the year. If during the year there is no purchase or redemption of shares and no dividend, then at the end of the year, the fund must distribute its approximately $17.50 of net income ($20 of gross income minus $2.50 of expenses) to avoid tax. The class with the lower expenses would pay a dividend of approximately $8.80 ($10 of income minus approximately $1.20 of expenses) or $0.88 per share and the other class will pay a dividend of approximately $8.70 ($10 of income minus approximately $1.30 of expenses) or $0.87 per share. Based on these facts, one might argue that the class with the lower expenses has a class "preference" to distributions of $0.10 (or $0.01 per share) over the class with the higher expenses, since shares of that class are entitled to a dividend that is in the aggregate $0.10 greater than the other class. However, assume that immediately before the dividend, five shares of the low-expense class are redeemed. As a result of the redemption, the dividend on the low-expense class would be $5.80 (($20-2)*5/15) minus $0.2 of class-specific expenses) or $1.16 per share, and the dividend on the other class would be $11.70 (($20-2)* 10/15) minus $0.30 of class-specific expenses) or $1.17 per share. In other words, the class with the higher expenses received a relatively higher dividend (in the aggregate and on a per share basis) by reason of its increase in relative size over the dividend period.

In summary, as the above example illustrates, basing class distinctions on differences in class-specific expenses does not provide one class with a priority to distributions over another class, though such distinctions affect the relative amounts distributable to each class. It is therefore arguable that these differences in distributions are not in accordance with class "preferences" within the meaning of Section 562(c). It is because of this ambiguity that, prior to the issuance of Rev. Proc. 99-40, RICs routinely requested letter rulings from the IRS that their multiple pricing arrangements would not violate the Preferential Dividend Rule.

The IRS has never explicitly relied on the analysis above to argue that these multiple pricing arrangements violate the Preferential Dividend Rule. However, the IRS does take the position that these arrangements do not result in separate classes of shares, as reflected by the use of the term "group" of shares instead of "class" of shares in Rev. Proc. 99-40. Beginning in 1990, the IRS issued hundreds of private rulings that defined a "class" for purposes of Section 562(c) as "a group of shareholders whose rights are so closely aligned and so different from other shareholders' rights as to warrant a conclusion that members of the group should all be treated the same and should be protected against the infringement of shareholders outside the group with respect to distributions."¹²³ Under this definition of "class," RICs with a multiple pricing arrangement would be treated as having a single class of shares paying non-pro rata dividends and thus must either obtain a private letter ruling or rely on Rev. Proc. 99-40 to avoid the risk that the IRS will treat those non-pro rata distributions as preferential under Section 562(c). In the private rulings addressing this issue, the IRS concluded that, based on the 1938 legislative history permitting minor differences in distributions within a class, differences in distributions among groups of shares within a "class" should not be treated as violating the Preferential Dividend Rule if they are attributable to de minimis differences in certain expense allocations, or differences attributable to "indirect shareholder expenses" (such as fees paid under a Rule 12b-1 plan).

The IRS seems to have adopted this position regarding the definition of "class" in an effort to enforce its view that one of the purposes of the Preferential Dividend Rule is to prevent a RIC from allocating the burden of its investment advisory fees and other expenses related to the management of the RIC's assets, including custodial fees and RIC tax-return preparation fees, differently among different categories. As discussed above, the IRS seems to have adopted this position based on the 1986 amendment to the Preferential Dividend Rule.

These concerns arise in part because of the analysis in General Counsel Memorandum 39457.¹²⁴ This memorandum considers the tax treatment of fees paid by the shareholders of a RIC to manage its assets. Under the facts presented, shareholders bear different levels of management fees (as a percentage of assets under management) based on differences in amounts invested in the RIC. The memorandum concludes that (1) these fees are necessarily expenses of the RIC and deductible only by it, since Congress clearly does not intend for RICs to be able to pass through deductions, and (2) the RIC's dividends violate the Preferential Dividend Rule since those dividends, after being reduced by management fees paid by the shareholders, are not pro rata to shares. The memorandum does not consider whether the violation of the Preferential Dividend Rule could be avoided by providing for different classes of shares that specify, by their terms, that different classes will bear different levels of fees.

Similar recharacterization concerns arise when investors in a RIC or REIT hold their shares through a partnership that allocates either partnership expenses or dividends from the RIC or REIT differently among different investors.¹²⁵ The desire to so allocate expenses or dividends might relate to differences in the level of capital or services being contributed by the different investors and typically would be both fully agreed upon by the investors and legally required by the terms of their investment. The recast concern is greater when the economic justification for the disproportionate allocation relates to management-related services performed for the RIC or REIT. While the IRS has not provided any guidance on whether it might attempt to recharacterize these partnership arrangements as violating the Preferential Dividend Rule, it has issued guidance limiting a RIC's ability to look through its interests in a partnership for purposes of applying the asset diversification test to only those partnerships that allocate expenses pro rata to investment.¹²⁶ Informally, IRS officials have acknowledged that this position is intended to reinforce the IRS's position that the Preferential Dividend Rule generally requires fees to be borne by common shareholders pro rata to shares, except as permitted by Rev. Proc. 99-40.¹²⁷

The risk of such recast seems particularly acute because it is less clear in the RIC and REIT context than in other contexts whether an expense should be treated as incurred either principally for the benefit of the fund and thus deductible by it or principally for the benefit of its shareholder and thus not deductible by the fund and treated as a dividend to the shareholder.¹²⁸ Even the well-settled rules addressing this question illustrate the uncertainty that exists in determining whether an expense is more appropriately treated as incurred at the corporate or shareholder level. For example, sales charges for the purchase of shares are not deductible by a RIC when paid by its shareholders but are effectively deductible by an open-end RIC when financed by the RIC's distributor through charges paid by the RIC to the distributor pursuant to a distribution plan adopted by Rule 12b-1 under the 1940 Act.¹²⁹ Similarly, RICs uniformly deduct their custody fees, and shareholders uniformly bear these fees in proportion to the net assets value of their shares.¹³⁰ Rev. Rul. 68-377 holds, however, that the custody fees charged for holding assets of a unit investment trust classified as a RIC and paid from income distributable by the trust are expenses of the shareholders and deductible by them and not the RIC.¹³¹

Short of obtaining a closing agreement with the IRS, a RIC or REIT is able to avoid income tax on the income and gain attributable to a preferential distribution only by paying another non-preferential dividend. There are several issues, however, about the availability and limits of this remedy. First, unless the deficiency dividend procedure is used (discussed below), this corrective dividend must be paid within the time required for distributing income for the year,¹³² which means that the RIC or REIT must discover the violation of the Preferential Dividend Rule in a timely manner. Second, there is a technical question about whether there can be sufficient earnings and profits, which measure the amount distributable as a dividend, to support the curing distributions. This is because earnings and profits are generally reduced by all distributions.¹³³ Therefore, for a RIC or REIT to be able to cure a preferential dividend by payment of another non-preferential dividend, the preferential dividend must be treated as not reducing earnings and profits so that the subsequent distribution will be treated as a dividend.

Section 852(c) and Section 857(d) generally provide that the current earnings and profits of a RIC or REIT respectively are not reduced by deductions not allowed in the computation of the company's "taxable income," without regard to whether the distribution requirement is satisfied. If applicable to preferential dividends, this section would preserve current earnings and profits otherwise reduced by preferential dividends and thus would allow another non-preferential dividend to be paid to cure the earlier violation of rule.

However, Section 852(c) and Section 857(d) may not preserve earnings and profits otherwise charged in respect of a preferential dividends. There are two arguments supporting such an interpretation. First, the dividends paid deduction is not a deduction that is relevant in the computation of the "taxable income" of a RIC or REIT.¹³⁴ Therefore, these provisions arguably do not apply to preserve earnings and profits otherwise reduced by preferential dividends. Second, preferential dividends are not dividends for purposes of the dividends-paid deduction and thus are not "disallowed" as deductions for purposes of the dividends paid deduction. In the REIT context, the IRS has issued a private ruling indicating that Section 857(d) preserves current earnings in respect of a preferential dividend, but it is unclear to what extent these questions were considered.¹³⁵

The deficiency dividend procedure also provides a potential for curing a violation of the Preferential Dividend Rule. This procedure allows a RIC or REIT determined not to have distributed all its income for a year to make a distribution of the undistributed amount in the form of a "deficiency dividend" that will be treated as paid in the earlier year even if there is otherwise insufficient earnings and profits at the time of the distribution to support a dividend.¹³⁶ This solution, however, requires the RIC or REIT to pay potentially substantial interest and, in the case of RICs, penalty charges.¹³⁷ In addition, this solution also presents technical issues.

Section 316(b)(3) specifically provides that a Section 301 distribution qualifying as a deficiency dividend is treated as a dividend regardless of the amount of earnings and profits available to support dividend treatment of the distribution. However, for a distribution to be a "deficiency dividend" under Section 860(f), it must be a distribution that, in addition to meeting other requirements, "would have been includable in the computation of the deduction for dividends paid under [S]ection 561 for the taxable [year] with respect to which [it relates] if [it had been] distributed during such taxable year." If a RIC lacks sufficient current earnings and profits to support full dividend treatment for an amount of nonpreferential Section 301 distributions equal to its otherwise taxable income in any year in which the RIC paid a preferential dividend, then under a technical reading of Section 860, the RIC or REIT would not be able to pay a deficiency dividend in the later year.¹³⁸

Contrary to the these possible interpretations, we believe that the Section 852(c), Section 857(d) and Section 860 should be interpreted to allow preferential dividends to be corrected by the payment of subsequent non-preferential dividends. There is no public policy reason to deny a RIC or REIT the ability to cure a violation of the Preferential Dividend Rule and thereby avoid an otherwise unavoidable corporate tax. We also believe, however, that this cure is not adequate, particularly when the violation is inadvertent.

We believe that existing cures are inadequate because their cost can be enormous and disproportionate to the magnitude of the violation. There are three reasons for this. First, shareholders are subject to tax on both the preferential dividend and the dividend paid to cure it, since both are by definition taxable dividends. This is true for those shareholders whose rights were violated as well as those shareholders benefiting from the impermissible preference. Second, if the deficiency dividend procedure is used to cure a violation of the Preferential Dividend Rule, significant penalties and interest can apply to the offending R1C or REIT as part of that procedure. Third, these additional tax costs are not correlated to the nature of the violation. The tax cost is based generally on the size of the dividend that violates the Preferential Dividend Rule and the time lag until it is cured. These costs are unrelated to the amount of the preference provided or egregiousness of violation. Providing a tiny impermissible preference, even by mistake, on a large dividend results in the same tax costs as providing a large preference (or willful violation of a shareholder right) on a dividend of the same size, and this cost is spread among shareholders on an indiscrimmant basis.

We understand anecdotally that some RICs and REITs have entered into closing agreements with the IRS regarding possible violations of the Preferential Dividend Rule that have avoided the payment of a deficiency dividend (or other non-preferential dividend taxable to shareholders) and thus allowed them to cure the potential violation at a lower cost. We commend the IRS for providing such relief, particularly where the possible violations were inadvertent. However, obtaining such an agreement at present does not seem to be a practical means for curing a possible violation of the rule. We understand anecdotally that very few of these agreements have been granted, that it is very difficult to obtain such an agreement and that the standards for granting this relief are shifting and unclear. As discussed below, this is why we recommend that, until the Preferential Dividend Rule is reformed legislatively, guidence be issued that provides a simple and easy means of curing violations.

V. BURDENS RESULTING FROM THE CURRENT APPLICATION OF THE RULE TO RICS AND REITS

RICs and REITs incur material costs ensuring that they comply with the Preferential Dividend Rule. These include building and operating systems to minimize the risk of inadvertent violations of the rule, auditing those operations to ensure that such violations in fact did not occur inadvertently and that any such violations are identified quickly to minimize the cost of correcting them, correcting inadvertent violations and seeking advice from experts and guidance from the government on these compliance issues.

As with any unclear tax rule, these compliance costs are magnified by the ambiguity of the Preferential Dividend Rule's content and scope. These additional costs include heavier reliance on outside advisors, greater reliance on private guidance from the IRS on how the rule will be applied in specific circumstances (and the cost of the government in providing that guidance), lost business opportunities when RICs and REITs decide not to engage in transactions allowed by the rule because such transactions are arguably not allowed, and the unfair competitive advantage that is obtained by RICs and REITs that are willing to exploit the rules ambiguities by taking aggressive tax positions.

The costs arising because of some RICs and REITs being too cautious and others being too aggressive are particularly heavy in the context of multi-tier investment structures and alternative fee arrangements. These structures and arrangements are desirable because they facilitate pooling (including, for example, through funds of funds, master-feeder structures, wrap fee arrangements and tiered partnership-fund structures) and the resulting efficiencies to investors. Recognizing these benefits, the SEC has blessed many such structures and arrangement. However, the IRS's administrative practice suggests, though without providing clear guidance, that the IRS believes that some of these structures and arrangements potentially violate the Preferential Dividend Rule. This ambiguity, particularly about what structures and arrangements the IRS views as impermissible, leads some RICs and REITs to forego business opportunities that are fully compliant and other RICs and REITs to obtain a competitive advantage based on their willingness to take aggressive tax positions. Inevitably, investors are less well off because of the market inefficiencies that result from different RICs and REITs interpreting these rules differently.

The extremely high cost (including possible inability) to cure violations of the Preferential Dividend Rule further magnifies these compliance costs. The costs of curing a violation includes (1) potentially significant interest and penalty charges applicable under the deficiency dividend procedure, (2) the double tax on the income and gain treated as twice distributed and (3) the unfairness resulting from the inability to shield innocent investors from these costs. These high costs of curing justify even greater efforts to avoid violations, further increasing compliance costs. Moreover, since showing good faith compliance or absence of negligence cannot reduce these costs, RICs and REITs generally hold themselves to an even higher standard of diligence. Such efforts include (1) investing even more in systems to prevent inadvertent violations, (2) incurring even greater expense to have the rule clarified (whether by tax advisors or the government) and (3) turning down even more business opportunities that arguably violate the rule.

VI. RECOMMENDATION AND RELIEF FOR INADVERTENT VIOLATIONS

We recommend that the Preferential Dividend Rule be repealed or reformed legislatively as it applies to RICs and REITs because, as explained in this Report, it no longer serves its original anti-tax-avoidance purpose, has been superseded by laws that better fulfill the purposes (including shareholder fairness) attributed to it and imposes significant burdens on RICs, REITs and their shareholders. Repealing the rule would also modestly further the public good of tax simplification. Until the rule is repealed, we recommend that guidance be issued that clarifies that the rule will be applied only in accordance with its plain language and that provides a simple means for RICs and REITs to treat distributions otherwise violating the Preferential Dividend Rule as deductible under Section 561, provided that the violation was inadvertent (including by reason of an incorrect but good-faith interpretation or application of the Preferential Dividend Rule).

Most inadvertent mistakes result from human or computer error rather than good faith mistakes in the interpretation or application of the rule. Examples of these errors include (1) applying the wrong amount of expenses to a class of shares, whether by reason of a miscalculation of the value of the shares or how the expense is measured or misidentifying the class to which the expense should be charged, (2) miscalculating the number of shares outstanding in a class for purposes of allocating a dividend among classes, (3) miscalculating the amount of income allocable to a class, whether by reason of a error in determining the number of shares outstanding or the net asset value of those shares, (4) miscalculating the amount of a priority distribution payable on a class of preferred shares, (5) mistakes by a paying agent in the timing or amount of a dividend to some shareholders (except where such mistake does not result in a violation because the unpaid dividend is treated as available to the shareholder) or (5) paying dividends in the wrong order or otherwise at the wrong time.

Violations can also arise from mistaken but good faith interpretations of the rule. Because the rule is ambiguous and because the IRS's administration of the rule tends to increase rather than decrease that ambiguity, it is possible that some RICs or REITs might apply the rule in a manner that they believe complies with the rule but arguably does not. For example, some multiple-class RICs calculate their dividends by allocating gross income in the same manner as prescribed by Rule 18f-3, which is based on relative net asset value. However, as discussed above, this method is arguably not permitted under Rev. Proc. 99-40, and some of those RICs have been unable to persuade their auditors that, based on outstanding legal authorities, there is a "more likely than not" basis for their position.¹³⁹ These RICs in effect are precluded from using their preferred method for calculating dividends for future periods and have a cloud over their financial statements for prior periods. We believe a RIC should be able to remove this cloud through the relief procedure outlined below.

We suggest the following conditions for relief:

We do not recommend that relief be conditioned on the payment of a subsequent "non-preferential" distribution (unless such distribution is required to correct a material nontax law breach of obligations to shareholders). Under the terms of the shareholders' rights, such a distribution may not be required and in some cases paying it may violate shareholder rights. There is no reason to believe that a subsequent distribution would mitigate income shifting among shareholders, since the shareholder base may have changed significantly since the payment of the preferential distribution (in which case the subsequent distribution would be paid to shareholders who would not have received it if it have been paid at the same time as the preferential distribution). If there remains income or gain to be distributed after the payment of the preferential distribution (because, for example, the preferential distribution resulted from some shareholders receiving less than their allocable share of income while other shareholders received no more than their share), then such undistributed income or gain should be distributed according to the fund's dividend policies (taking into account the need to correct material breaches of its obligations to shareholders), which may require that the undistributed income or gain be distributed to the under-distributed class or may require that it be distributed in some other manner.

As a condition of relief, any obligation to shareholders (as determined under applicable non-tax law) resulting from the violation will have been corrected. After taking into account such correction, the shareholders will have been treated with "substantial impartiality" and any difference in distributions should be treated, therefore, as "minor" under the logic of New York Stocks, Inc.

This relief is also consistent with the rule's anti-tax avoidance purpose. If the violation was unintended, it could not have been tax motivated. If it resulted from a good faith but mistaken legal interpretation, then relief is conditioned on a certification that a principal purpose of the violation was not tax avoidance. The IRS can of course challenge the validity of the certification in an examination.

In most circumstances, distinctions in "types" of violations break down when analyzed. Almost all preferential dividends involve one or more shareholders receiving more than they should have received relative to the amounts distributed to other shareholders, as determined at the time that the distribution is made. This is true whether, as determined at the time of a distribution, the distribution is non-pro rata with respect to a single class or results in a preferred class failing to receive its preference relative to a subordinate class. For this reason, we recommend a single approach that provides relief for all inadvertent violations as outline above. We doubt that there are good reasons for basing the scope or type of relief on purported distinctions in the type of violation, and we suspect that any such distinction would likely be subject to manipulation.

² The terms "regulated investment company" and "real estate investment trust" are defined in Sections 851(a) and 856, respectively, of the Internal Revenue Code of 1986, as amended (the "Code"). All references to "Section" are to sections of the Code unless otherwise stated.

³ This deduction is also used in Section 545 to compute the tax on undistributed personal holding income imposed under Section 541, in the case of personal holding companies, and it is used in Section 535 to compute the tax on accumulated taxable income under Section 531, in the case of other corporations subject to that tax.

⁴ Sections 852(a) and 857(a).

⁵ Sections 852(b) and 857(b). RICs and REITs are also subject to an excise tax under Sections 4982 and 4981, respectively, on certain amounts not treated as distributed within a calendar year. To be treated as distributed for this purpose, a distribution generally must be eligible for the dividends paid deduction under Section 561 (or subject to regular income tax at the RIC or REIT level).

⁶ The Report does not address whether the Preferential Dividend Rule should be modified or repealed as it applies to personal holding companies that are not RICs. Many of the observations about the rule's effectiveness in preventing tax avoidance apply in that context. If it would be useful, we would be pleased to address this question in another report. For the same reasons that we believe that the Preferential Dividend Rule should be repealed as it applies to the calculation of income and excise tax of a RIC and REIT, we believe that it should not apply to the calculation of the personal holding company tax applicable to RICs that are personal holding companies.

⁷ Others have recommended the repeal of the Preferential Dividend Rule as it applies to RICs. See, e.g., Letter from Dale S. Collinson and Louis W. Ricker to Donald C. Lubick, Acting Assistant Secretary, Treasury (February 3, 1998), 98 TNT 48-37 ("[This memorandum] relates to the preferential dividend rule ('Preferential Dividend Rule'), a somewhat arcane and, as applied to publicly offered regulated investment companies ('RICs'), unnecessary tax requirement."); Roger J. Baneman, Preferential Dividends in the Regulated Investment Company Context, 111 Tax Notes 49, 61 (2006) ("The preferential dividend provisions of section 562(c) simply make no sense when applied in the RIC context.").

⁸ See note 35.

⁹ These provisions are specifically designed to prevent a corporation from paying dividends to one class of shares while deferring other shareholders' receipt of that income. See, e.g., H.R. Rep. No. 413 (Pt. 1), 91st Cong., 1st Sess. 88 (1969), 1969-3 C.B. 200, 391; S. Rep. No. 552, 91st Cong., 1st Sess. 150 (1969), 1969-3 C.B. 423, 519 (provisions passed because the "final regulations issued on January 10, 1969, do not cover all of the arrangements by which cash dividends can be paid to some shareholders and other shareholders can be given corresponding increases in proportionate interest."). In particular, under Section 305(b)(2), a distribution or deemed distribution of stock is taxable where it results in the receipt of property by some shareholders and an increase in the proportionate interest of other shareholders in the assets or earnings of the corporation. Section 305(c) provides for a long list of circumstances where an increase in the relative interest of a corporation will be treated as a deemed stock dividend and thus trigger taxation of it under Section 305(b)(2). Therefore, under current law, if some shareholders receive less than their pro rata share of a dividend but are protected against any lost value in their shares by an increase in their relative interest in the corporation's remaining value, the amount of that relative increase will be treated, in many circumstances, as a taxable stock dividend to such shareholders. Treas. Reg. § 1.305-3(d).

¹⁰ In these transactions, a RIC or REIT issued to a tax indifferent party (like a tax-exempt organization) a class of preferred shares that were entitled to all distributions for a fixed period and thus amortized much of the principal invested over that period. These amortization payments were intended to be treated as dividends to the full extent possible, and thus allowed the RIC or REIT a dividends paid deduction against income, to the extent of those payments, that would otherwise have to have been distributed to the other shareholders. The arrangement functioned like a financing by the holders of fast-pay stock for the benefit of the other shareholder, who maintained their economic position in the underlying assets and effectively deducted both the principal and interest costs associated with the financing by avoiding the recognition of their share of any of the income of the RIC or REIT, since all of that income was treated as distributed to the preferred holders. Reversing the benefit of these deductions of principal could be deferred until the structure was unwound.

¹¹ Under current Section 305 regulations, it is unclear that in these transactions the shareholders not receiving cash dividends would be treated as receiving constructive stock dividends by reason of their relative increase in the RIC or REIT as a result of the cash dividends paid on the fast-pay stock. While these shareholders' relative interest in the RIC or REIT increased in connection with the dividends paid on the fast-pay stock, such increase did not result from a change in their legal rights, like a change in redemption price of their shares or conversion ratio, as arguably required by Section 305(c) for such relative increase to be treated as a constructive stock distribution. However, strong arguments could be made that Section 305 provides the IRS with sufficient authority to issue rules that would treat the common holders in these transactions as receiving such dividends. Under Treas. Reg. § 1.305-3, a redemption of shares from some shareholders, if not isolated and treated as dividends under Section 301, results in a constructive taxable stock dividend to the non-redeeming shareholders under Section 305(c) by reason of the relative increase in their interest in the corporation. See Treas. Reg. § 1.305-3(e) examples 8 and 9. By analogy, this would seem to provide support in the fast-pay stock transactions for treating the holders of regular stock as receiving taxable constructive dividends under Sections 305(b)(2) and 305(c), or at least a clear path for regulations to be issued that would provide that result.

¹² Treas. Reg. § 1.7701(1)-3 is specifically targeted at "fast-pay arrangements" (defined generally to include self-amortizing stock investments) by generally treating them as financing directly between the fast-pay shareholders and the shareholders otherwise receiving the benefit of deferral.

¹³ The Preferential Dividend Rule applies to charges to earnings and profits in respect of a redemption taxed as a sale or exchange, but it does not affect the ability of a RIC or REIT to deduct pro rata dividends on shares that provide for redemption or liquidation preferences.

¹⁴ Even at the time when the rule was enacted in 1936, mandatory disclosure was recognized as a way of providing some measure of protection to the investing public, as evidenced by the Securities Act of 1933 and Securities Exchange Act of 1934. See note 59.

¹⁵ See note 54 and accompanying text.

¹⁶ 15 U.S.C. §§ 80a-1 - 80a-64.

¹⁷ Pursuant to the broad rulemaking authority granted to the SEC under the 1940 Act, over 150 rules that prescribe more specific requirements for investment companies have been promulgated by the SEC. In addition, the SEC has issued numerous interpretive releases that express its views and the SEC staffs views on issues under the 1940 Act. Furthermore, the SEC staff has issued thousands of informal no-action letters and exemptive orders to investment companies relating to various provisions of the 1940 Act in specific contexts.

¹⁸ The first comprehensive study of the industry after the passage of the 1940 Act was undertaken in 1958 by the Securities Research Unit of the Wharton School of Finance and Commerce at the University of Pennsylvania and at the direction of the SEC. The report focused on the relationships among investment companies, investment advisers, principal underwriters and portfolio brokers. The report found that these relationships tended to create conflicts of interest between fund management and shareholders and an absence of arm's length bargaining between management and investment advisers. Based on this report, the SEC submitted its views to Congress in 1966 in its Report of the Securities and Exchange Commission on the Public Policy Implications of Investment Company Growth, H.R. Rep. No. 2337, 89th Cong., 2d Sess. (1966) (the "PPI Report").

¹⁹ The PPI Report formed the basis for significant revisions to the 1940 Act that were passed by Congress in 1970.

²⁰ For example, the PPI Report concluded that mutual fund shareholders needed protection from excessive management fees and sales loads and that neither competition nor existing safeguards in the 1940 Act were sufficient to provide the necessary protection. In 1970, Congress responded by amending the 1940 Act to provide better protection against high sales charges and excessive fees, including an express requirement that advisory compensation be "reasonable." See Gerard H. Manges, The Investment Company Amendments Act of 1970, 26 Bus. Law. 1311 (1971); Gerard H. Manges, The Investment Company Amendments Act of 1970 -- An Analysis and Appraisal After Two Years, 14 B.C. L. Rev. 387(1973).

²¹ See Form N-1A. Item 22(b)(6).

²² See notes 59-63 and accompanying text.

²³ See notes 59-63 and accompanying text.

²⁴ This version of the rule provided. "No dividends paid credit shall be allowed with respect to any distribution unless the distribution is pro rata, equal in amount, and with no preference to any share of stock as compared with other shares of the same class." Pub. L. No. 74-740, § 27(g), 49 Stat. 1648, 1665 (1936) (the "1936 Tax Act")

²⁵ Subject to its version of the Preferential Dividend Rule, Section 27 of the 1936 Tax Act generally defined the dividends paid credit to include, in addition to regular cash dividends (including certain carry over amounts), dividends paid in kind or in obligations of the corporation, taxable stock dividends and charges against earnings and profits in respect of a liquidating distribution.

²⁶ Pub. L. No. 75-554, § 27(h), 52 Stat. 447, 470 (1938) (the "1938 Tax Act"):

Preferential Dividends -- The amount of any distribution (although each portion thereof is received by a shareholder as a taxable dividend), not made in connection with a consent distribution (as defined in section 28(a)(4)), shall not be considered as dividends paid for the purpose of computing the basic surtax credit, unless such distribution is pro rata, with no preference to any share of stock as compared with other shares of the same class, and with no preference to one class of stock as compared with another class except to the extent that the former is entitled (without reference to waivers of their rights by shareholders) to such preference. For a distribution made in connection with a consent distribution, see section 28.

The provision was not changed again until the 1954 Code, which replaced the dividends-paid credit with the dividends-paid deduction and introduced the current version of the preferential dividend rule, except with respect to the relief for shareholders investing at least $10 million. Sections 561 and 562(c) of the Code. The sentence of Section 562(c) relating to this relief was added to the rule in 1986, Tax Reform Act of 1986 ("TRA '86"), Pub. L. No. 99-514, § 657(a), 100 Stat. 2085, 2299.

²⁷ The personal holding company rules were first enacted in 1934 and were intended to address the "scheme of the 'incorporated pocketbook'" for individuals, thus preventing individual shareholders deferral of their share of corporate income if ownership of the corporation was concentrated in the hands of a few individuals. See H.R Rep. No. 704, 73rd Cong., 2d Sess. (1934), 1939-1 (Pt. 2) C.B. 554 at 562; S. Rep. No. 558, 73rd Cong., 2d Sess. (1934).

²⁸ The accumulated earnings tax was first enacted as a shareholder-level tax in 1913. Pub. L. No. 63-16, § 11(A)(2), 38 Stat. 114, 166 (1913). It became an entity tax in 1921 and applied to corporations "formed or fraudulently availed of for the avoidance of shareholder taxation on accumulated corporate income. Pub. L. No. 63-98, § 220, 42 Stat. 227, 247-48 (1921).

²⁹ Section 14 of the 1936 Tax Act. This surtax was intended to reinforce the accumulated earnings tax and personal holding company tax. H.R. Rep. No. 2475, 74^th Cong., 2d Sess. at 670 (1936). ("Under [existing law] a corporation may retain all its net income and the Government fails to receive the tax it should secure from the surtax on the stockholders of such corporations. It is true, that if the Government can prove a corporation is formed or availed of for the purpose of preventing the imposition of surtaxes upon its shareholders, a special surtax can be collected from the corporation under section 102 of the existing law [the accumulated earnings tax]. The difficulty of proving such purpose, however, has rendered section 102 more or less ineffective. Moreover, under existing law, a corporation secures an unfair advantage over a partnership whose members are obliged to include in their returns their pro rata share of the partnership profits whether distributed or not.")

³⁰ GPD, Inc. v. Comm'r, 508 F.2d 1076 (6th Cir. 1974) ("We construe the shareholder election provisions of the 1934 Tax Act and previous acts to be an indication of congressional disapproval of non-pro rata distribution as a means of evading the tax and to establish a statutory scheme exempting the corporation from the tax only when the entire amount of its adjusted net income is included pro rata in the shareholder's gross income. When such income has been included, a corporation has not been availed of for the proscribed purpose.")

³¹ Lincoln Nat'l Bank v. Burnet, 63 F.2d 131 (D.C. Cir. 1933) (held that a non-pro rata distribution to a shareholder was taxable as a dividend to that shareholder, who had reported the distribution as a gift in recognition of services performed by such shareholder to other shareholders, who approved the non-pro rata distribution.) Even in 1936 courts had long since recognized that a shareholder holding stock on which a dividend is declared and paid cannot avoid recognizing such dividend as income by assigning it. Warner, 5 B.T.A. 963 (1926). However, there is no indication in the Lincoln National Bank opinion that either party argued that all shareholders should be treated as receiving a pro rata dividend followed by a transfer of dividend proceeds among shareholders, presumably because the directors of the corporation had "ordered." without objection from the other shareholders, that the dividend be paid in the non-pro rata manner.

³² Black Motor Co. v. Comm'r, 125 F.2d 977 (6^th Cir. 1942). This opinion provided no other citation for this.

³³ Priv. Ltr. Rul. 8903073 (Oct. 26, 1988). The Congressional hearings cited by the ruling in support of this argument discuss the need to prevent shareholder deferral of corporate income but do not discuss whether efforts to tax such income might be defeated by non-pro rata distributions of such income to low bracket taxpayers. Revenue Act of 1936: Hearings on H.R. 12395, 74^th Cong., 2d Sess. 62 (1936).

³⁴ H.R. Rep. No. 1860, 75th Cong., 3d Sess. 23 (1938), 1939-1 (Pt. 2) C.B. 728, 744. The report also indicates that this purpose is unchanged from the 1936 Tax Act.

³⁵ See Rev. Rul. 53, 1953-1 C.B. 178 (holding that waivers for bona fide business reasons will not result in income to the waiving shareholder provided that there is no family or direct business relationship between the waiving shareholder and minority shareholders receiving the dividend); Rev. Rul. 56-431. 1956-2 C.B. 171 (disregarding a dividend waiver by a majority shareholder for the benefit of minority shareholders who are members of his family); and Rev. Proc. 67-14, 1967-1 C.B. 591 (identifying circumstances in which the IRS will issue a private ruling respecting dividend waivers).

³⁶ In 1954, Congress defined taxable stock dividends as only those that either provided shareholders with a stock or cash election or were issued to discharge preference. Pub. L. No. 83-591, 68A Stat. (1954) (as amended, the "1954 Code"). Prior to this time, there had been extensive litigation and conflicting case law on the scope of Congress' power to tax dividends. For cases applying different standards, see, e.g., Sprouse v. Comm'r, 122 F.2d 973 (9th Cir. 1941), aff'd 318 U.S. 604 (1943) and Dreyfuss v. Manning, 44 F. Supp. 383 (D. N.J. 1942) and contrast with those with Strassbugger v. Comm'r, 124 F.2d 315 (2d Cir. 1941), rev'd 318 U.S. 604 (1943) and Frank J. & Hubert Kelly Trust v. Comm'r, 38 B.T.A. 1014 (1938), vacated 106 F.2d 1002 (8th Cir. 1939).

³⁷ Section 115(f) of the 1936 Tax Act.

³⁸ Until 1936, Congress abandoned any attempt to tax stock dividends after Eisner v. Macomber, 252 U.S. 189 (1920), which held the taxation of certain stock dividends to be unconstitutional. From 1921 to 1936, all stock dividends were excludable from income by statute. Compare Revenue Act of 1921, Pub. L. No. 67-98, § 201(d), 42 Stat. 227, 228 (1921), with Section 115(f)(1) of the 1936 Tax Act. The 1936 Tax Act's taxation of stock dividends was a response to Koshland v. Helvering, 298 U.S. 441 (1936), which had been decided earlier in 1936 and held that a taxpayer had full basis in stock received as a nontaxable stock dividend because such dividend was income that Congress could have taxed under its constitutional powers.

³⁹ See note 36.

⁴⁰ Pub. L. No. 91-172, 83 Stat. 487 (1969) (the "1969 Act")

⁴¹ Section 305(c) of the 1954 Code as amended by the 1969 Act.

⁴² The report states:

The committee believes that no distribution which treats shareholders with substantial impartiality and in a manner consistent with their rights under their stock-holding interests, should be regarded as preferential by reason of minor differences in valuations of property distributed. To illustrate: Suppose the case in which a stock dividend distribution is coupled with an option equally available to all the shareholders to take cash or stock, and in which $100 in cash is paid to one shareholder, while there is issued to another shareholder one share of stock which happens to have, on the date on which the distribution is authorized, or on the dale of its receipt by the shareholder, an exchange value of $99.75. An impartial distribution of this character should not be considered to be preferential.

H.R. Rep. No. 1860, 75th Cong., 3d Sess. 23 (1938), 1939-1 (Pt. 2) C.B. 728, 744.

⁴³ New York Stocks, Inc. v. Comm'r, 164 F.2d 75 (2nd Cir. 1947). See also National Securities Series -- Indus. Stocks Series, 13 TC 884 (1949).

⁴⁴ Open end funds permit shareholders to redeem their shares on demand for an amount equal to the net asset value of the shares. Prior to 1942, mutual investment companies (the predecessor of RICs) were required to be organized as open-end funds. Section 48(e)(1)(E) of the 1936 Tax Act and Section 361(a)(5) of the 1938 Tax Act.

⁴⁵ Amounts charged against earnings in respect of a redemption were eligible for the dividends paid credit provided that such amounts were not preferential. See note 8.

⁴⁶ New York Stocks, Inc. at 78. New York Stocks, Inc., like the other cases applying the Preferential Dividend Rule, addresses permissible differences in distributions among shares within a single class of shares. Its principles, however, should also apply to differences among classes of shares in determining whether such differences violate the Preferential Dividend Rule.

⁴⁷ Id. at 79 (if the credit were not available in this circumstance, "the more the significant and necessary provision for redemption is exercised, the less the company receives a tax benefit, so upon a 100 per cent redemption it would have to pay a tax on all its earnings. Since Congress could so have intended, these strange consequences are not decisive of a contrary intent. More important is the fact that every shareholder receives his fair proportion of the earnings during the period he elects to remain an owner, those redeeming being paid when they redeem, those retaining ownership being paid at the end of the [company's] dividend period. Since the decision to retain or redeem is left to the unfettered judgment of the shareholder, there is no preference and no injustice to any shareholder.").

⁴⁸ May Hosiery Mills, Inc. v. Comm'r, 123 F.2d 858 (4th Cir. 1941) (redemption of shares on the open market is preferential, since there was "no equal opportunity" to participate in the repurchase); George E. Warren Co. v. Comm'r, 53 F. Supp. 578 (D Mass. 1944) (redemptions pursuant to a tender by an industrial company which are not equally available to all shareholders are treated as preferential). An example in the current regulations under Section 562(c) denies a deduction for a non-pro-rata redemption but does not specify whether the tender was made equally available or whether it was taxed as an exchange to the shareholders. Treas. Reg. § 1.562-2(b), Ex. (2).

⁴⁹ Section 48(e) of the 1936 Tax Act.

⁵⁰ Section 13(a)(3) of the 1936 Tax Act.

⁵¹ Statement of Paul C. Cabot and Merrill Griswold, 1936 Tax Act: Hearings, 74th Cong, at 799 ("Most shareholders [of investment trusts] are persons of moderate means . . . who do not have equal facilities with the wealthy to obtain expert supervision and diversity in their investments. It is in order to obtain these benefits that they have availed themselves of these funds. . . .").

⁵² Investment funds generally became subject to entity taxation under the Morrissey Trust decision in 1935. Morrissey Trust v. Comm'r, 296 U.S. 344 (1935) (an unincorporated business trust managed on behalf of its owners is an "association taxable as a corporation" rather than taxable as a trust, resulting in its distributed income being subject to tax at both the corporate and shareholder levels). Statement of John S. Myers, 1936 Tax Act: Hearings on HR 12395 before the Senate Comm. on Finance, 74th Cong. 777, 780 (1936) ("[T]hese trusts should be treated like any other trusts for tax purposes. . . . This type of organization has been formed and has been in existence since as early as 1923 or before. Both the sponsor and trustees, as well as the Commissioner of Internal Revenue, treated such organizations as true trusts for tax purposes.")

⁵³ For example, testimony on the provisions of the 1936 Tax Act included statement that "bad" investment trusts "ought not to be given any particular inducement to go on." Revenue Act of 1936: Hearings on HR 12395 before the Senate Comm. on Finance, 74th Cong. 10-11 (1936) (statements of Sen. Walsh and Mr. Kent, Acting Chief Counsel, Bureau of Internal Revenue); and Revenue Act of 1936: Hearings on HR 12395 before the Senate Comm. on Finance, 74th Cong. 35-37 (1936) (comments of Sen. LaFollette).

⁵⁴ Prior to the enactment of the 1940 Act, the only federal securities laws generally applicable to investment funds were the Securities Act of 1933 (which regulated public offerings) and the Securities Exchange Act of 1934 (which regulated companies with traded shares).

⁵⁵ In response to reported abuses, in 1935 Congress directed the SEC to study investment trusts and investment companies, and to report its results and recommendations to Congress. Public Utility Holding Company Act of 1935, Pub. L. No. 74-333, § 30, 49 Stat 803.

⁵⁶ See note 62.

⁵⁷ Under the MIC qualification requirements, shareholders had to be entitled to redeem their shares on reasonable notice. Section 48(e)(1)(E) of the 1936 Tax Act and Section 361(a)(5) of die 1938 Tax Act.

⁵⁸ For example, in his address to Congress on June 19, 1935, President Roosevelt stated (regarding a tax on dividends received by corporations), "Bona fide investment trusts that submit to public regulation and perform the function of permitting small investors to obtain the benefit of diversification of risk may well be exempted from this tax." H.R. Rep. No. 1681, 74th Cong., 1^st Sess. at 644 (1935).

⁵⁹ Becoming popular during the 1920s, investment companies were largely unregulated during these early years of development but proliferated rapidly. After the market crash of 1929, the values of investment companies declined precipitously. In addition, many investment companies suffered frauds from affiliates. In the 1930s, publicly offered investment companies became subject to the Securities Act of 1933 and, in the case of those traded on an exchange, the Securities Exchange Act of 1934. The Securities Act of 1933 utilized a regime of public disclosure to combat abuse. However, by the middle to late 1930s, Congress recognized that the disclosure approach of the 1933 Act was insufficient to eliminate the potential for abuse when large pools of public assets were entrusted to others for investment. In 1935, Congress directed the SEC to study investment trusts and investment companies, and to report its results and recommendations to Congress. This congressional mandate resulted in an exhaustive study of the investment company industry, titled "Report of the Securities and Exchange Commission on Investment Trusts and Investment Companies" (1939) (the "Investment Trust Study"), the recommendations of which formed the basis of the 1940 Act. legislation that fundamentally changed how investment companies are permitted to operate. The Investment Trust Study concluded that these abuses resulted primarily from the liquid, mobile, and readily negotiable nature of cash and securities that made it easy for managers of pooled assets to embezzle, steal, or use them for improper purposes or to pursue their own interests rather than those of shareholders, especially where the companies were controlled by banking, brokerage, or dealer interests. The study noted that unrestricted dealings between investment companies and affiliated persons, while permitting greater flexibility and allowing investment companies to share in the affiliates' profits, largely were used to allow insiders to treat investment companies as sources of private capital and a dumping ground for unprofitable securities.

⁶⁰ Based on recommendations made by the Investment Trust Study, Congress passed the 1940 Act and the Investment Advisers Act of 1940, 15 U.S.C. §§ 80b-1 - 80b-21. The 1940 Act imposes a significant federal regulatory structure on the organization and operation of investment companies, and explicitly and broadly empowers the SEC to regulate the activities of investment companies in the public interest and for the protection of investors.

⁶¹ Section 1(a) of the 1940 Act declares that, based on the findings of the Investment Trust Study, investment companies are affected with a national public interest, and Section 1(b)(1)-(8) of the 1940 Act lists the problems adversely affecting the national public interest and the interests of shareholders and declares that the policy and purpose of the 1940 Act are to mitigate the problems outlined therein.

⁶² This provision is designed generally to prohibit investment companies from creating overly complicated capital structures and to ensure the equitable treatment of shareholders. Section 18 of the 1940 Act defines a "senior security" as any bond, debenture, note or similar obligation constituting a security and evidencing indebtedness. This definition also includes any stock of a class having priority over any other class as to distribution of assets or payment of dividends. In a Congressional report on the necessity of regulating the investment company industry, the term "priority" is used to refer to priority of right or claim, in order to distinguish preferred stock, which has a preferred but limited claim to a company's assets, from common stock, which has a residual but unlimited claim to a company's assets. Report on Investment Trusts and Investment Companies, Pt. 3, H.R. Rep. No. 70, 76th Cong., 1st Sess. 1576 (1939). However, the SEC interprets the prohibitions of Section 18 much more broadly than indicated in the Congressional report, and today takes the position that in addition to the types of instruments included in the definition of "senior security" in Section 18(g), numerous other instruments and trading practices of investment companies could give rise to a "senior security" under Section 18. In particular, the SEC has focused on transactions that involve great material differences between the rights of investment company shareholders, and also on transactions that involve potential leveraging of an investment company's assets.

⁶³ In interpreting the provisions of Section 18, the SEC's staff takes the position that any material difference in the rights of an investment company's shareholders comes within the broad sweep of the term "senior security." See, e.g., Lance M. Brofman, Securities Act Release No. 6664, Exchange Act Release No. 3673; Investment Advisers Act Release No. 1044, Investment Company Act Release No. 15340, 36 SEC Docket 1249 (Oct. 2, 1986). As a result, the SEC has issued exemptive relief from Section 18 or has promulgated various rules under the 1940 Act, in order to permit new arrangements that were not foreseen when the 1940 Act was enacted. Such new arrangements involve the creation of multi-class or multiple series investment companies in which the only difference between share classes involves distribution-related and other shareholder expenses, or in which a single corporate entity operates with numerous different series and each series is treated as a separate investment fund by shareholders. See, e.g., Fidelity Special Situations Fund, Investment Company Act Release No. 15258 (Aug. 15, 1986); Tax-Free Cash Reserve, Inc., Investment Company Act Release. No. 14656 (Aug. 2, 1985).

⁶⁴ In addition to being registered under the 1940 Act (or having another specified type of narrowly defined status under the securities laws), RICs were required, as they are today, to elect RIC status and meet both a gross income test and diversification test. Revenue Act of 1942, Pub. L. No. 77-753, § 170, 56 Stat. 798, 880 (the "1942 Tax Act"). At that time, they were also required to limit their recognition of short term gains, but that requirement was repealed in 1997. Taxpayer Relief Act of 1997, Pub. L. No. 105-34, § 1271, 111 Stat. 788, 1036.

⁶⁵ H.R. Rep. No. 2333, 77^th Cong., 1^st Sess. at 28 (1942) ("The new provisions enlarge the category of companies entitled to special tax treatment and liberalize the standards required to be met. These results have the approval of the Securities and Exchange Commission, in light of their experience with the supervision of investment trusts and the objectives sought to be reached in the Investment Company Act of 1940, as amended.").

⁶⁶ Hearings on Bills Affecting the Tax Treatment of Pass-Through Entities Before the Subcomm. on Select Revenue Measures of the House Comm. on Ways and Means, 99th Cong. 119, 135 (1986) (statement of Dennis E. Ross, Tax Legislative Counsel, Dep't of the Treasury).

⁶⁷ In addition, the application of the dividends paid credit was extended to include the measure of distributions for a RICs satisfaction of the distribution requirement. Section 170 of the 1942 Tax Act. Previously, all dividends counted toward the satisfaction of this requirement but not toward the dividends paid credit. New York Stocks, Inc. v. Comm'r, 164 F.2d 75 (2d Cir. 1947).

⁶⁸ Pub. L. No. 86-779, § 10, 74 Stat. 998, 1006-08 (1960).

⁶⁹ Section 857(b)(2) and (3).

⁷⁰ Section 857(a)(1).

⁷¹ 1960-2 C.B. 820. The Committee on Ways and Means explained:

The omission of the corporate income tax in the case of distributed earnings, which present law provides for regulated investment companies, secures for investors in these companies essentially the same tax treatment as they would have received if they had invested directly in the operating companies. H.R. 12559 extends this same type of tax treatment to real estate investment trusts specializing in investments in real estate equities and mortgages as distinct from the stock and security holdings of regulated investment companies. Thus this secures for the trust beneficiaries the same type of tax treatment they would receive if they held the real estate equities and mortgages directly and, therefore, equates their treatment with that accorded investors in regulated investment companies.

Id. at 821.

⁷² Id.

⁷³ Section 657(a) of TRA '86. The IRS has interpreted a shareholder's "initial investment" for this purpose as the amount put in the account when opened. Priv. Ltr. Rul. 8802011 (July 27, 1987). It is unclear whether subsequent investments may be counted toward the $10 million threshold. It would seem reasonable to conclude that subsequent investments are so included and that the term "initial" is intended to exclude only capital appreciation.

⁷⁴ H.R. Conf. Rep. No. 841, 99th Cong., 2d Sess. 11-246 (1986), 1986-3 (vol. 4) C.B. 1, 246.

⁷⁵ 1996-40 I.R.B. 10. This announcement requests comments on the tax treatment of expense waivers and reimbursements by RICs.

⁷⁶ See, e.g., Zero Tolerance Policy for RIC Distribution Differentials Outlined. Highlights & Documents, June 6, 1995, at 3564.

⁷⁷ The cross-subsidization concern was frequently addressed in private rulings addressing the preferential dividend issue in the context of multiple-class RICs. See, e.g., Priv. Ltr. Ruls. 8741062 (July 15, 1987), 8806006 (Aug. 27, 1987), 8817029 (Jan. 28, 1988), 9523019 (Mar. 10, 1995). and 9549027 (Dec. 8, 1995). For an explanation of this concern, these rulings refer to GCM 39457 (Dec. 18, 1985). See note 124 and accompanying text. Of course since some fees benefit only some shareholders, allocating them pro rata necessarily results in cross-subsidization among shareholders, a result clearly permitted under the Preferential Dividend Rule and never questioned by the IRS in any published guidance or rulings.

⁷⁸ See notes 20-21.

⁷⁹ Treas. Reg. § 1.562-2(a).

⁸⁰ Sections 561(b) and 562(a).

⁸¹ Sections 561(b) and 562(b)(1)(A). A RIC can be a personal holding company, though a REIT generally cannot under Section 856(h).

⁸² Sections 561(b) and 562(b)(1)(B).

⁸³ Sections 561(b) and 562(b)(2).

⁸⁴ Section 561(a)(2).

⁸⁵ Section 562 (with respect to the first four categories listed) and Section 565 (with respect to consent dividends).

⁸⁶ It is clear that the preferential dividend rule applies to constructive Section 301 distributions. Lucas, 71 TC 838, 855 (1979). See also Schwartz, 60 T.C. 728, 748 (1973).

⁸⁷ See, e.g., United States v. Gotcher, 401 F.2d 118 (5^th Cir. 1968) and Rubber Associates, Inc. v. Comm'r, 335 F.2d 75 (6^th Cir. 1964) ("ordinary and necessary" corporate expenses that advance the corporation's business interests are deductible notwithstanding that shareholders may receive an incidental benefit).

⁸⁸ RICs and REITs are taxed on the share of such "excess inclusions" to the extent such share is allocated to a shareholder that is a "disqualified organization." Sections 860E(e)(1), 860E(e)(2), 860E(e)(5), 860E(e)(6); Treas. Reg. §§ 1.860E-1(b) (reserved): 1.860E-2(b)(1).

⁸⁹ Treas. Reg. § 1.860E-2(b)(3) and (4). In Rev. Rul. 2006-58, 2006-46 I.R.B. 1, the IRS ruled that, because a charitable remainder trust is a "disqualified organization," the distribution of an excess inclusion to such trust will trigger a tax at the distributing RIC level, the cost of which can be allocated specifically thereto without giving rise to a preferential dividend under Section 562(c), by virtue of Treas. Reg. § 1.860E-2(b)(4).

⁹⁰ Priv. Ltr. Rul. 200651022 (Sept. 13, 2006). A disgorgement of fees from a service provider to a RIC has been held to be qualifying income to the RIC (consistent with Rev. Rul. 92-56, 1992-2 C.B. 153) and does not result in a preferential dividend even if the disgorgement is allocated differently among different classes, so long as the amounts (including interest and penalties) are allocated consistently with how they were incurred, in accordance with Rev. Proc. 99-40, 1999-2 C.B. 565. Priv. Ltr. Rul. 200812016 (March 21, 2008).

⁹¹ Treas. Reg. § 1.562-2(a). The limitation imposed by section 562(c) is unqualified, except in the case of an actual distribution made in connection with a consent distribution, see Section 565, provided that the entire distribution composed of such actual distribution and consent distribution is not preferential. The regulation continues, "A corporation will not be entitled to a deduction for dividends paid with respect to any distribution upon a class of stock if there is distributed to any shareholder of such class (in proportion to the number of shares held by him) more or less than his pro rata part of the distribution as compared with the distribution made to any other shareholder of the same class. Nor will a corporation be entitled to a deduction for dividends paid in the case of any distribution upon a class of stock if there is distributed upon such class of stock more or less than the amount to which it is entitled as compared with any other class of stock." After these regulations were issued, Section 562(c) was amended to permit a non-pro rata increase in distributions of a RIC to shareholders making an initial investment of at least $10 million so long as such increase is solely by reason of administrative cost savings. TRA '86, § 657(a), 100 Stat. 2085, 2299. The IRS has interpreted a shareholder's "initial investment" for this purpose as the amount put in the account when opened. Priv. Ltr. Rul. 8802011 (July 27, 1987). It is unclear whether subsequent investments may be counted toward the $10 million threshold. Presumably the term "initial" is intended to exclude only capital appreciation.

⁹² Treas. Reg. § 1.562-2(a). However, an agreement among shareholders to participate in dividends other than in conformity with their rights as shareholders should not affect the deductibility of the dividends provided that the shareholders are treated as receiving the dividend on a pro rata basis. See. e.g., Coca-Cola Bottling Co. of Greenville, 143 F.2d 381 (5th Cir. 1944) (shareholder agreement to assign dividends among shareholders other than on pro rata basis does not affect conclusion that dividends paid by corporation are pro rata).

⁹³ Treas. Reg. § 1.562-2(a). The regulation continues. "A preference exists if any rights to preference inherent in any class of stock are violated." Our proposal would allow RICs and REITs to deduct dividends that violate rights of shareholders, since as discussed above, we do not believe that the Preferential Dividend Rule is an effective means of protecting shareholders from these violations.

⁹⁴ Treas. Reg. § 1.562-2(b), Ex (1). See also, e.g.. Safety Convoy Co. v. Thomas, 139 F.2d 219 (5th Cir. 1943) (a pro rata dividend paid in non-pro rata installments over a three-year period was preferential; court found that "a dividend must be paid in equal amount upon each share of stock in any given class, and must be paid at the same time, if the dividends-paid credit is to be made applicable"). However, a dividend will be treated as paid for this purpose when made available. See, e.g.. United Artists Theatre Circuit, Inc., 1 T.C. 424 (1943) (dividend irrevocably set aside and made available to all preferred shareholders upon conversion of their shares pursuant to a recapitalization in which they were required to participate was not preferential even though not all shareholders received the dividend before the end of the year. Emphasizing the company's uniform treatment of the preferred shareholders, the court observed, "[i]t could perhaps be said that the dividend was so far made available to them that they merely turned their backs upon income which they were free to enjoy, though, as stated, we need not go so far in this case").

⁹⁵ Treas. Reg. § 1.562-2(b), Ex. (2).

⁹⁶ May Hosiery Mills, Inc. v. Comm'r, 123 F.2d 858 (4th Cir. 1941) (redemption of shares on the open market is preferential, since there was "no equal opportunity" to participate in the repurchase); George E. Warren Co. v. Comm'r, 53 F. Supp. 578 (D. Mass. 1944) (redemptions pursuant to a tender by an industrial company which are not equally available to all shareholders are treated as preferential).

⁹⁷ H.H. King Flour Mills Co. v. Comm'r, 325 F. Supp. 1085 (D. Minn. 1971) (redemptions pursuant to tender offer equally available are preferential; the court indicated that New York Stocks v. Commissioner, 164 F. 2d 75 (2d Cir. 1947), and National Securities Series -- Indus. Stocks Series, 13 T.C. 884 (1949), both of which involved redemptions that were not deemed preferential, apply only to open-end companies).

⁹⁸ New York Stocks, Inc. v. Comm'r, 164 F.2d 75 (2nd Cir. 1947); National Securities Series -- Indus. Stocks Series, 13 TC 884 (1949).

⁹⁹ Priv. Ltr. Rul. 9449016 (Sept. 13, 1994). In the fund-of-funds structure in this ruling, all of the shares of a RIC (the lower-tier fund) were proposed to be held by other RICs (upper-tier funds). Because redemptions by the upper-tier funds from the lower-tier fund generally would have little effect on the upper-tier fund's percentage interest in the lower-tier fund, many of those redemptions would likely be treated as Section 301 distributions. If our proposal is adopted and Section 562(c) repealed as it applies to RICs and REITs, we note that allowing closed end funds to deduct Section 301 redemptions will result in an opportunity for income shifting that does not presently exist. In our view, however, Section 305 should still provide adequate protection from this possibility, and this result does not materially change the degree of income shifting that already exists under subchapter M.

¹⁰⁰ The ruling quotes United Artists Theatre Circuit v. Comm'r, 1 TC 424, 430 (1943), acq. 1943 C.B. 23: "[W]here a distribution is made available in conformity with the rights of each stockholder, where no act of injustice to any shareholder is contemplated or perpetuated, where there is no suggestion of a tax avoidance scheme, and where each stockholder is treated with absolute impartiality, the distribution is not preferential. . . ." The ruling also cites New York Stocks, Inc. v. Comm'r, 164 F.2d 75 (2d Cir. 1947) and National Sec. Series -- Indus. Stocks Series, 13 TC 884 (1949). These cases do not address whether redemptions treated as Section 301 distributions are preferential. They hold that charges against earnings and profits in respect of redemptions taxed as sales or exchanges are not preferential.

¹⁰¹ Rev. Rul. 83-117, 1983-2 C.B. 98.

¹⁰² Treas. Reg. § 1.562-2(a).

¹⁰³ Treas. Reg. § 1.562-2(b), Ex. (3).

¹⁰⁴ For example, in Black Motor Co. v. Comm'r, 125 F.2d 977 (6th Cir. 1942), a company declared a pro rata 10 percent dividend payable in four installments. It paid out only the first installment in full, on a pro rata basis, and its partial payment on the second through fourth installments was on a non-pro rata basis. Rather than finding that the first installment was a nonpreferential distribution separate from the last three, the court held thai none of the payments were entitled to the dividends-paid credit (the predecessor of the dividends-paid deduction). We are not aware of any published authority addressing the question of when distributions paid pursuant to separate declarations are treated as separate or when constructive and actual dividends are treated as separate for purposes of the Preferential Dividend Rule.

¹⁰⁵ This charge is commonly called a ""front-end load" and is paid at the time of sale of the shares.

¹⁰⁶ Before 1980, open-end RICs were prohibited by Section 18 of the 1940 Act from using their own assets to pay for the distribution of their shares, including in particular commissions on the sale of their shares. In 1980, the Securities and Exchange Commission (SEC) adopted Rule 12b-1 (under the 1940 Act) to permit these RICs to finance these sales commissions from charges to their own assets. 17 C.F.R. § 270.12b-1 (1980). The financing works as follows: The RICs distributor pays the brokerage commissions associated with the sale of the RICs shares and then collects a fee from the RIC over time ("12b-1 Fees"). The amount of the 12b-1 Fees charged to the mutual fund is typically calculated in a manner that is intended to cover all or a portion of the cost of the brokerage commissions, with interest.

¹⁰⁷ Rev. Proc. 99-40, 1999-46 I.R.B. 1, amplifying and superseding Rev. Proc. 96-47, 1996-2 C.B. 338.

¹⁰⁸ 17 C.F.R. 270.18f-3.

¹⁰⁹ Rev. Proc. 99-40 specifically provides that the conditions for its application "are to be interpreted in a manner consistent with the SEC's interpretation of analogous requirements in the rules under the 1940 Act. See Rule 18f-3(a). . . ."

¹¹⁰ This relief was needed because Section 18(f)(1) of the 1940 Act generally makes it "unlawful for any registered open-end company to issue any class of senior security. . . ." Section 18(g) defines "senior security" to include any stock of a class having priority over any other class as to distribution of assets or payment of dividends. Section 18(i) requires that every share of stock issued by a registered investment company be voting stock, with the same voting rights as every other outstanding voting stock. 15 U.S.C. §§ 80a-18(f)(1) - 80a-18(i). Rule 18f-3 largely codifies the exemptive order approach of addressing the potential for conflicts among classes by limiting the permissible differences among classes in expenses and voting rights. This rule was since amended. Investment Company Act Releases No. 22835 (Sept. 26, 1997); No. 24816 (Jan. 2, 2001); and No. 26520 (July 27, 2004).

¹¹¹ Securities Act Release No. 7143, Investment Company Act Release No. 20915 (Feb. 23, 1995).

¹¹² Id.

¹¹³ Id

¹¹⁴ A class that pays a front-end load, for example, differs from a class paying a Rule 12b-1 Fee in a spread-load or level-load arrangement in the amount, the form (by shareholders individually versus the class as a whole), and timing (at purchase versus over time) of distribution charges.

¹¹⁵ These expenses encompass both differences in actual out-of-pocket expenses among classes (for instance, blue sky fees that are incurred for some classes but not others), and differences in charges when classes receive services that are different in kind or amount. For example, some classes may use transfer agency services differently than others. Thus, the rule contemplates that allocations may be based upon the level or kind of services used. Amendments to Rule 18f-3 expanded the permissible allocation methods of income, gain, and expenses among different classes of shares. Investment Company Act Release No. 22835 (Sept. 26, 1997).

¹¹⁶ Securities Act Release No. 7143, Investment Company Act Release No. 20915 (February 23, 1995).

¹¹⁷ Rev. Proc. 99-40, 1999-46 I.R.B. 1. Under any of our recommendations, this revenue procedure would no longer be needed.

¹¹⁸ Id. Sec. 5.

¹¹⁹ In addition to these three conditions, there are two other conditions for relying on the revenue procedure: (1) the rights and obligations of the shareholders of each class must be fixed in the corporation's "organizing documents" which must provide that (a) each class is entitled to distributions calculated in the same manner and at the same time as all other classes (except in the case of differences described above) and (b) expenses are to be allocated to each class at the same time as to all other classes, and (2) the shares of each class must separately meet the requirements of Section 67(c)(2)(B), which defines "publicly offered" RICs. To be publicly offered for this purpose, the shares of each class must be continuously offered pursuant to a public offering, and such shares must be regularly traded on an established securities market or held by no fewer than 500 persons at all times during the taxable year.

¹²⁰ Any waiver or reimbursement of these fees and expenses must be applied to the group on behalf of which it was incurred. Rev. Proc. 99-40 § 4.01.

¹²¹ These expenses may be waived or reimbursed in accordance with the manner allocated. Rev. Proc. 99-40 § 4.02. This requirement has been interpreted to permit interest and penalties paid in connection with a fee reimbursement pursuant to an SEC order to be allocated among different classes of shares in the same manner as reimbursed fees. Priv. Ltr. Rul. 200812016 (March 21, 2008)

¹²² Any waiver or reimbursement of management expenses must be allocated by net asset value except to properly account for differences attributable to applicable performance fees. Rev. Proc. 99-40 § 4.03.

¹²³ See, e.g., Priv. Ltr. Ruls. 9126033 (Apr. 1, 1991), 9125015 (Mar. 21, 1991). Prior to this, in private rulings, the IRS originally took the position that a RIC had two or more classes of stock for purposes of Section 562(c) when it had two or more classes of shares for state law purposes that, as a result of having different distribution plans, had different voting, dividend, liquidation, and redemption rights. See, e.g., Priv. Ltr. Ruls. 8850055 (Sept. 21, 1988), 8822082 (Mar. 10, 1988.) Outside the investment company context, distinctions in distribution or redemption rights among shareholders give rise to separate classes. See, e.g., Treas. Reg. § 1.1361-1(1)(1), relating to S corporations.

¹²⁴ This memorandum is dated Dec. 18, 1985, and seemed to have provided the analytical basis of Priv. Ltr. Ruls. 8601016 (Sept. 30, 1985) and 8552063 (Sept. 30, 1985).

¹²⁵ For example, if the anti-abuse rules of Treas. Reg. § 1.701-2 were applied to disregard the partnership, different investors would be treated as receiving different distributions in violation of the Preferential Dividend Rule unless either (1) such differences are determined to be in accordance with preferences inherent within the classes of shares paying the distribution, or (2) they are treated as paid pro rata and then assigned by and among the shareholders pursuant to a shareholder-level agreement. See, e.g., Coca-Cola Bottling Co. of Greenville v. United States, 143 F.2d 381 (5th Cir. 1944) (shareholder agreement to assign dividends among shareholders other than on pro rata basis does not affect conclusion that dividends paid by corporation are pro rata).

¹²⁶ See, e.g., Rev. Proc. 2001-57, 2001-50 I.R.B. 577.

¹²⁷ Sheryl Stratton, IRS Wants Practitioner Input on RIC Distributions Rules, 67 Tax Notes 1418, 1418 (Jun. 12, 1995) ("The IRS will not tolerate management fee allocation differentials in master feeder structures . . . reiterated IRS officials . . . . 'Management fees always have to be proportionate,' insisted Alan B. Munro, Jr. . . .").

¹²⁸ In the context of corporate and shareholder expenses, identifying the party potentially eligible to deduct an expense is based on which party primarily benefits from its payment, not who paid it. See, e.g., Deputy v. DuPont, 308 U.S. 488, 494 (1940) ("[I]t is the origin of the liability out of which the expense accrues which is material."); South Am. Gold & Platinum Co. v. Comm'r, 8 T.C. 1297, 1301 (1947), aff'd without opin. 168 F.2d 71 (2d Cir. 1948) (legal expenses paid by a parent holding company in settling claims for and against its subsidiaries were not deductible by parent); Webb v. United States, 560 F. Supp. 150, 157 (S.D. Miss. 1982) (shareholder's satisfaction of a failure-to-pay penalty imposed on his corporation was an expense of the corporation, not the shareholder).

¹²⁹ Rev. Rul. 73-463, 1973-2 C.B. 34, as amplified by Rev. Rul. 94-70, 1994-2 C.B. 17 (permitting these fees to be deducted based on the rationale that the cost associated with the need to replenish capital is an ordinary and necessary expense of an open-end RIC). In addition, the IRS has ruled privately that frequent flier miles provided by a RIC's distributor to the RIC's shareholders are not preferential dividends, implying that this arrangement has no affect on the RICs deduction of fees paid to the distributor pursuant to its distribution plan under Rule 12b-1. Priv. Ltr Ruls. 200121004 (May 29, 2001), 199920031 (May 24, 1999) and 9746048 (Nov. 14, 1997).

¹³⁰ Rev. Proc. 99-40 lists these fees as among those of a RIC that are required to be allocated among shareholders pro rata, implying that they are required to be deducted by a RIC and are not deductible by its shareholders even if paid by them. Similarly, Treas. Reg. § 1.67-2T(j)(1) identifies custodian fees (among certain other types) as "affected expenses" and thus not deductible by a private RIC under Section 67(c), implying that the fees are an expense of the RIC and otherwise deductible.

¹³¹ 1968-2 C.B. 301. In the ruling, the investors were contractually obligated to pay these fees, and the trust's only asset were shares of another management-type RIC (as opposed to a RIC organized as a unit investment trust), which were held by the custodian. In addition to holding the managed RIC shares for the trust, the custodian maintained the trust's shareholder records and accepted investor subscriptions to the trust.

¹³² For excise tax purposes, dividends generally must be paid within the calendar year to reduce the amount subject to the excise tax for the year. Sections 4981(c) and 4982(c). For income tax purposes, dividends deductible against income for a year generally can be paid up to 12 months after the end of the tax year or the first regular dividend of the next year. Sections 858 and 855.

¹³³ Section 312.

¹³⁴ In the case of a RIC, for example, the dividends paid deduction is an adjustment to "taxable income" for purposes of computing "investment company taxable income" under Section 852(b)(2), the amount of the RIC's income (along with undistributed net capital gains) that is subject to entity tax. See also, Section 852(b)(3), with respect to the tax on undistributed net capital gain of RICs, and Section 857(b)(2) and (3) for analogous provisions for REITs.

¹³⁵ Priv. Ltr. Rul. 200729021 (July 20, 2007). This result may apply only to REITs because Section 857(d)(3) seems to provide more support for this result and the RIC provisions do not contain an analogous rule.

¹³⁶ Section 860 . There are other conditions that must be satisfied for a RIC to claim a deficiency dividend.

¹³⁷ The interest charges are calculated based on the time delay between the under distribution and deficiency dividend but are based on the full amount of the deficiency dividend instead of the tax due on it. Section 860(c)(1). RICs are liable for an additional penalty equal to the lesser of that interest charge or half of the amount of the deficiency dividend. Section 6697.

¹³⁸ This is because Section 316(b)(3) provides for dividend treatment of only distributions qualifying as deficiency dividends, and it provides for such treatment only in the year of actual distribution. Section 316(b)(3) does not provide that certain distributions would have been treated as dividends if they had been made in a prior year notwithstanding that there were insufficient earnings and profits in that prior year to support its dividend treatment. It is probably fair to understand Section 316(b)(3), added by the Tax Reform Act of 1976, Pub. L. No. 94-455, § 1601(d), 90 Stat. 1520, 1746 (1976), as intended to affect only shareholder-level (not the RIC-level) tax treatment of a deficiency dividend. Section 860 prescribes the RIC-level treatment of a deficiency dividend. In particular, Section 860(a) provides that a deduction "shall be allowed" for a deficiency dividend. Section 316(b)(3) simply requires that the actual recipients of the deficiency dividend treat it as a dividend in their hands without regard to the earnings and profits of the distributing RIC at the time of the distribution.

¹³⁹ In reporting their financial results (and, in the case of RICs, the net asset value of their shares), RICs and REITs using generally accepted accounting principles must fully accrue the tax liability in respect of any tax position for which there is not a "more likely than not" legal basis. Interpretation No. 48, Accounting for Uncertainty in Income Taxes: An Interpretation of FASB Statement No. 109.