Menu
Tax Notes logo

Firm Finds Endowment Tax Regs Reach Too Far

OCT. 1, 2019

Firm Finds Endowment Tax Regs Reach Too Far

DATED OCT. 1, 2019
DOCUMENT ATTRIBUTES

October 1, 2019

CC:PA:LPD:PR (REG-106877-18)
Room 5203
Internal Revenue Service
P.O. Box 7604
Ben Franklin Station, Washington, DC 20044

RE: Proposed Guidance on the Determination of the Section 4968 Excise Tax Applicable to Certain Private Colleges and Universities

Ladies and Gentlemen:

We represent a variety of nonprofit organizations, including churches, universities, community foundations, hospital systems, donor advised fund sponsoring organizations, and supporting organizations to the foregoing entities. On behalf of some of our clients, we are providing these comments to the proposed regulations (the “Proposed Regulations”) under section 4968 of the Internal Revenue Code (the “Code”) released via a notice of proposed rulemaking (the “NPRM”) July 3, 2019.1

We have strong concerns with the Proposed Regulations' definitions of related and controlled organizations, which if not modified could cause many entities that control, or are under common control with, one or more educational institutions, to be treated as though they are controlled by the educational institutions. This would result in their assets being taken into account as funds of an applicable educational institution even when they are dedicated to unrelated charitable endeavors of the controlling or commonly controlled organizations, in direct contravention to the clear intent of Congress in drafting section 4968 of the Code. We also believe that the regulations under section 4968 should clarify that general unrestricted funds of a parent entity primarily engaged in charitable or religious activities other than supporting one or more educational institutions should not be considered intended or available for such institutions until such time as the parent entity affirmatively authorizes such funds to be used for them.

As currently drafted these regulations could affect many organizations without Harvard-size endowments; as explained below, one in three private nonprofit schools participating in Title IV funding programs would be subject to the tax with attributable assets of as little as $500 million, and more than half of all such schools would be subject to the excise tax if their attributable assets reached $1 billion. Overbroad control tests and constructive ownership rules could lead many entities to reach these thresholds inadvertently and unwittingly because of assets elsewhere in their affiliated groups. And even organizations not ultimately having enough assets to meet the threshold for applicability of the tax under section 4968(b)(1)(D) could be required to engage in burdensome collection of information across their affiliated groups and beyond to confirm that fact.

I. The Proposed Definition of Control is Too Broad and Inconsistent with the Statutory Scheme

Section 4968 requires educational institutions to take into account certain assets of related institutions — including those that control, are controlled by, or are under common control with the educational institutions. The House version of the bill originally treated the assets of all related organizations — parents, subsidiaries, or commonly controlled organizations — as belonging to the educational institution.2 However, the version ultimately enacted included a Senate amendment (the “Senate Amendment”) that limited this inclusion of related organization assets and income:

[U]nless such organization is controlled by such institution or is described in section 509(a)(3) with respect to such institution for the taxable year, assets and net investment income which are not intended or available for the use or benefit of the educational institution shall not be taken into account.3

This provision creates an asymmetry between controlled organizations and other related organizations. It recognizes that while an educational institution can reasonably be treated as having access to all of the assets of any entity that it directly or indirectly controls, it does not typically have unrestricted access to assets of other organizations that control it and their subsidiaries. For instance, in the case of a large religious denomination or religious order controlling one or more educational institutions, the educational institutions do not have access to the general assets of the religious denomination, which would typically be used primarily to conduct religious activities of the denomination or order. Without the Senate Amendment, educational institutions would potentially be taxed with respect to assets over which they had no real control, direct or indirect; by the same token, larger systems of entities could be penalized if, anywhere in their system, they had an applicable educational institution.

Any adequate regulatory definition of control for purposes of section 4968 must preserve the Senate Amendment's core distinction between subsidiaries on the one hand and parent or sister organizations on the other. However, as indicated below, the Proposed Regulations fail to preserve that distinction, and are thus inconsistent with the statute as finally enacted by Congress.

A. The Definition in Treasury Regulation § 1.512(b)-1(l)(4)

The Proposed Regulations indicate an intent to adopt the definition of control in Section 512(b)(13)(D) of the Code “and the regulations thereunder.” They then adopt a standard taken entirely from Treasury Regulation § 1.512(b)-1(l)(4) under which an organization is controlled by another whenever more than half its directors or trustees are either (i) removable and replaceable by another organization, or (ii) “representatives” of the other (directors, trustees, agents, or employees).4 We have no quibble with the part of this definition based on removal and replacement powers. However, the “representatives” prong (the “Representatives Test”) often incorrectly determines which entity controls the other.

1. The Representatives Test is Not Current Law

As a threshold matter, this regulatory provision is not actually an interpretation of current section 512(b)(13)(D). Rather, it was likely superseded by current section 512(b)(13)(D), which was enacted by Congress in 1997 to revamp the rules for which organizations would be controlled within the meaning of section 512(b)(13).5 Prior to amendment, the definition of control in section 512(b)(13) was based on an 80% control test without any constructive ownership rules for indirect ownership. The regulations provided (and still provide) additional definitions of control for corporations and nonstock corporations.6 The new provision took a modified version of the regulatory control test for corporations and included it in the statute, along with a new statutory control test for partnerships and for “any other case.” Unlike the corporate control test, the Representatives Test was not included in any form, even though the relevant House Reports did note that the Representatives Test had been part of the prior law. The House Reports indicated that the new statutory language was not only adding constructive ownership rules, but also modifying the existing test for control:

The bill modifies the test for determining control for purposes of section 512(b)(13). Under the bill, "control" means (in the case of a stock corporation) ownership by vote or value of more than 50 percent of the stock. In the case of a partnership or other entity, control means ownership of more than 50 percent of the profits, capital or beneficial interests.7

Both this legislative history and the language of the statute itself show that Congress intended to provide a comprehensive new definition of control, consisting in more than 50% profits or capital interest in a partnership, more than 50% of the stock (by vote or value) in a corporation, or more than 50% of the beneficial interest in any other organization. The Committee Report indicates that the corporate stock test was meant for “stock corporations,” so it appears that most nonstock corporations would be under the catchall 50%-of-beneficial-interest test. But even if they should be under an appropriately adjusted version of the corporate prong of section 512(b)(13)(D), that would not be consistent with the Representatives Test, which does not follow any of the prongs of section 512(b)(13)(D) as passed by Congress in 1997. Rather, the Representatives Test was the only part of the old regulation that was left entirely out of the new statutory control definition. Treasury and the IRS seem to be treating section 512(b)(13)(D) as though it simply left the regulations in place and changed their percentage thresholds from 80% to 50%, but that is not the case.

It thus appears that the Representatives Test has been superseded by the new statutory definition. Indeed, after section 512(b)(13)(D) the IRS itself seems to have recognized that it had superseded the Representatives Test. In the only subsequent private letter ruling we have found in which the IRS has considered whether a nonstock organization is controlled within the meaning of section 512(b)(13) after 1997, the IRS found control solely by finding that the tax-exempt parent owned all the beneficial interests of the tax-exempt subsidiaries within the meaning of section 512(b)(13)(D); it did not even cite Treasury Regulation § 1.512(b)-1(l)(4) or its Representatives Test.8 In fact, for audit years after 1997 (a period of more than 20 years), we have found no rulings applying Treasury Regulation § 1.512(b)-1(l)(4) to determine whether an organization is controlled within the meaning of section 512(b)(13).

2. The Representatives Test Leads to Obviously Erroneous Results

It is clear that the presence of common individuals in the leadership of two organizations does not demonstrate that one organization controls the other. At best, such overlap by itself can show only that the two organizations are commonly controlled by the same persons (the overlapping persons). Depending on the circumstances, the overlapping persons may be controlling organization A in their capacity as agents of organization B or vice versa, or may be controlling each organization completely separately. Such overlap therefore cannot be used as the criterion for distinguishing between organizations controlled by educational institutions and those that are merely commonly controlled with them — but that distinction is critical to the section 4968 statutory scheme.

It is worth noting that the Representatives Test is not used for stock corporations, partnerships, or trusts. If it were, obviously nonsensical conclusions would follow. For instance, any trust with a single trustee (for instance a family trust or testamentary trust) would find itself classified as controlled by any educational institution of which that trustee became an employee or board member. This would be inappropriate because, although the individual would be serving both as trustee of the trust and as a representative of the educational institution, the educational institution would have no right to direct that individual trustee in the trustee's administration of the trust. Neither the educational institution nor its directors, in their capacity as such, would have any control over the trust. The same problems would arise for a private company having a single owner/director. If applied, the Representative Test would wrongly treat the educational institution as having control in such situations because it disregards the different capacities in which individuals act, essentially assuming that representatives of one organization are always acting on its behalf whenever they exert influence over another entity.

The Proposed Regulations rightly reject this standard of control for entities other than nonstock corporations. But the same considerations that make the Representatives Test inappropriate for other types of entities also make it inappropriate for nonstock exempt entities. If the trust in the preceding example were a private foundation structured as a trust or as a single-director nonstock corporation, it might be more difficult to identify its beneficial owners with particularity, but it would be equally clear that the directors of the educational institution would have no right to control the private foundation in their capacity as directors of the educational institution. It would make no sense to treat the private foundation's assets as automatically part of the educational institution's assets for purposes of section 4968.

Other control tests in the exempt organizations context do not have this defect. For instance, the test used to determine when a private foundation controls another entity sufficiently to cause it to engage in an indirect self-dealing transaction explicitly limits control to situations where the foundation managers can exercise such control acting solely in their capacity as foundation managers.9 A control test that followed this model would be more likely to comport with the purposes of section 4968.

When most or all of an organization's board are also representatives of a school, there are a variety of factors that can clearly indicate that the organization, and not the school, controls: (1) The organization appoints the board of the school and not vice versa; (2) the school's trustees are all accountable to the organization, but common board members do not report to the school on their management of the organization; (3) the school furthers the purposes of the organization, but most of the organization's activities are not designed to advance the purposes of the school; (4) the organization's leaders, in their capacity as such, could cause the assets of the school to be made available for the organization's use, but the school's leaders, in their capacity as such, have no authority to appropriate additional organization assets for their use. Perhaps not accidentally, these kinds of factors are close to the dominant factors in determining control for other entities besides nonprofits: focusing on appointment powers is akin to focusing on having voting stock in a corporation; focusing on ability to control the disposition of assets is akin to focusing on ownership by value or beneficial interest.

The Representatives Test could be especially devastating to hierarchical religious organizations using corporations sole or similar structures. The test does not apply by its terms to corporations sole, which have neither directors nor trustees. However, if one were to make a natural extension of the rule by treating the incumbent officer of a corporation sole as a sole director or trustee, it would follow that the ecclesiastical officeholder could not be on the board of any educational institution in the world without causing the corporation sole to be deemed controlled by that educational institution. Similarly, suppose a religious organization had a small board of three directors who could appoint all of the trustees of an educational institution, and chose to appoint themselves and six others with educational experience to that board, appointing themselves so that they could make sure the religious organization would maintain a strong voice in the governance of the educational institution. It seems clear that the religious organization would control the educational institution. But the Representatives Test would conclude that the educational institution also controlled the religious organization. And if the religious leaders also served on the boards of other affiliates with only one or two other individuals, the Representatives Test could well cause the educational institution to be deemed to control these other institutions as well.

These are not merely theoretical hypotheticals. Particularly in hierarchical churches, it is very common for entities to be controlled by a single ecclesiastical officer, and it may be a requirement of a church's polity that this officer be included on the board of various affiliated nonprofits, including religious orders, social ministry entities, healthcare institutions, and the like. This is not so that each of these other nonprofits can control the core church entity, but rather to make sure proper ecclesiastical oversight is maintained. Treasury should not adopt a test for control that discourages these ecclesiastical leaders from being on the boards of religiously affiliated universities; to do so would constitute unwarranted government interference with sensitive internal affairs of religious organizations and would raise grave Constitutional concerns.10 Furthermore, even if the Representatives Test is not applied to corporations sole themselves, in many cases nonprofit entities affiliated with those corporations sole will have the required overlap because of the presence of the same ecclesiastical officials on many of their boards.

3. Alternatives to the Representatives Test

Rather than employ the Representatives Test, Treasury should provide guidance establishing tests for nonstock control that are closer to the control tests for other organizations, and which take into account the fact that the same individual may act in more than one capacity.

  • Treasury could simply delete the Representatives Test and focus on the right to remove and replace more than 50% of the directors of an organization.

  • Treasury could define control using a variant of the control test for indirect self-dealing by private foundations, for instance: “For purposes of section 4968, an organization is controlled by an educational institution if the educational institution or one or more of its managers or employees (acting only in such capacity) may, only by aggregating their votes or positions of authority, remove and replace more than 50% of the organization's highest governing body, control the disposition of the assets of the organization, or regularly control the annual expenditures and distributions by such organization.”

  • Treasury could potentially also find control where more than 50% of the beneficial interest in an organization is vested in the educational institution, for instance where the educational institution is entitled to receive more than half of the assets upon dissolution or regularly receives more than 50% of the annual distributions from the organization. However, where an educational institution does not have sufficient control to obtain access to all of the assets and income of an organization, it is unclear why it should be forced to take all of those assets and income into account for purposes of section 4968.

These alternatives have the virtue of being more closely parallel to definitions of control for other entities, which focus on voting power in the election of directors or rights to a majority of the assets or income of an entity. Those alternatives that focus on voting control or control over assets also comport more closely with the purposes of the Senate Amendment, by finding control and automatically including the assets of another organization in an educational institution's assets only if the educational institution has real power over the use of those assets.

B. The Attribution Rules of Section 318 Are Inconsistent With the Senate Amendment

Section 318 provides constructive ownership rules used to identify owners of stock for various purposes under the tax code. Section 318 provides for both “upward” and “downward” attribution. Under upward attribution, partners in a partnership are each deemed to own their proportional shares of the partnership's holdings; beneficiaries of a trust are deemed to own a proportionate share of the trust's assets based on their actuarial interests; and an owner of more than 50% of a stock corporation by value is treated as owning a proportionate share of the stock corporation's holdings.11 Under downward attribution, partnerships and trusts are considered to own everything owned by their partners and beneficiaries, respectively; corporations are considered to own a proportionate part of any other stock owned by a parent organization that owns more than 50% of their stock by value.12

The Proposed Regulations provide that “[t]he principles of section 318 apply for purposes of determining ownership of stock in a corporation, and similar principles apply for purposes of determining ownership of interest in any other entity.”13 On its face at least, this language appears to use section 318's principles to determine indirect or constructive control of all entities, including control of nonstock corporations and tax-exempt organizations.

It is not clear exactly how to apply these principles in the nonstock context. For instance, for most corporations, both downward and upward attribution depend on more than 50% of the value of a corporation being owned by the parent entity; what facts would satisfy that requirement in the nonprofit context? We note that the Form 990 Schedule R instructions, which also indicate that the constructive ownership principles of section 318 should be applied, appear to apply section 318's principles only to determine ownership in ordinary stock corporations, partnerships, and private trusts; there are no examples of imputing control of nonstock exempt organizations using section 318, and indeed the instructions appear to have separate rules for determining indirect control for exempt organizations.14

The IRS has, however, applied section 318 principles to nonstock corporations in the section 512(b)(13) context. In Private Letter Ruling 199941048 (July 20, 1999), the Service considered a university and two of its supporting organizations, one of which had a wholly-owned for-profit subsidiary. Concluding that the University would be considered to own all the beneficial interests in both supporting organizations, the Service concluded that, by upward attribution, the University would be considered to indirectly control the for-profit subsidiary. Furthermore, by downward attribution, the University's control of the for-profit was imputed to the sister supporting organization that did not own the for-profit. By the same reasoning, each supporting organization would be deemed to control the other.

Diagram 1: PLR 199941048 Constructive Ownership

As this example demonstrates, the “downward attribution” rules of section 318(a)(3) are fundamentally incompatible with section 4968, because attributing holdings from a parent organization to its subsidiaries causes each subsidiary to constructively control the others, wrongly converting what is in reality a situation of common control to be deemed direct control of one subsidiary by the other. As the number of direct and indirect controlled subsidiaries grows, for instance in a national organization with numerous local chapters under its control, these problems become exponentially larger, potentially treating a school at the bottom of a large corporate pyramid as though it controlled all of the other organizations in the pyramid besides its direct and indirect parent entities.

The downward attribution rules of section 318 can have other even stranger consequences. For instance, in the example below, modified from an example provided by the Service in Notice 2018-26,15 an educational institution controlled by a church could be deemed to own assets of another unrelated church just because the two were invested in the same investment partnership:

Diagram 2: Based on Notice 2018-26 Section 3.01 Constructive Ownership

In this example, the downward attribution rules attribute all control held by partners of the Investment Partnership (including ownership of Educational Institution and For-Profit) to the Investment Partnership, which therefore controls both of these entities constructively.

Then, because they are commonly controlled by Investment Partnership, Educational Institution and For-Profit are deemed to control each other by a second application of downward attribution. If these rules were applied in the section 4968 context, Educational Institution would have to take into account the assets of For-Profit, even though Charity A and Charity B are not related at all, other than by common investment in a partnership.

The way to avoid these issues is simple: Regulations defining control for purposes of section 4968 should not apply the downward attribution rules of section 318(a)(3). Furthermore, if they apply the upward attribution rules of section 318, the Service should provide clearer guidelines in the nonstock corporation context as to how such attribution would apply.

C. Treasury Should Clarify the Treatment of 509(a)(3) Organizations Supporting Multiple Supported Organizations

Section 4968 treats assets and income of two types of related organization as automatically attributable to an educational institution: organizations controlled by the institution, and organizations “described in section 509(a)(3)” with respect to the educational institution.16 Section 509(a)(3) provides public charity status to certain organizations that are, among other things, (i) “organized, and at all times thereafter . . . operated exclusively for the benefit of, to perform the functions of, or to carry out the purposes of” their designated section 509(a)(1) and (2) public charity “supported organizations,” and (ii) operated, supervised, or controlled by or in connection with such designated supported organizations.17 Thus, an organization described in section 509(a)(3) with respect to an educational institution is one organized and operated exclusively for the benefit of that institution and having the requisite governance and operational connections with that organization (these are typically classified as Type I, II, and III relationships).

In the simple case where a supporting organization only supports a single educational institution, it is organized and operated exclusively for the benefit of that institution, and so it makes sense that its assets would automatically be treated as assets of the educational institution, without any further inquiry into whether the assets were intended or available for the institution. Such assets are, in effect, all earmarked for the institution's use.

However, many supporting organizations support multiple supported organizations. For instance, a Type I supporting organization that is controlled by X City Community Foundation might support a class consisting of all public charities in X City, including Y University. It would be described in section 509(a)(3) with respect to that class of charities as a whole, because it would be organized and operated exclusively for their benefit and would be operated, supervised, or controlled by or in connection with that class as a whole. However, it would not be described in section 509(a)(3) with respect to Y University, because it would not be organized or operated exclusively for the benefit of, to perform the functions of, or to carry out the purposes of Y University — any more than an organization operated for the exclusive benefit of the United States of America could be said to be operated exclusively for the benefit of Alaska, or Utah, or Maine. Support for Y University might be only one small portion of the supporting organization's purpose, not its exclusive or even primary purpose. Indeed, in many such cases Y University might never have received a distribution from the supporting organization and would have no right to such a distribution. It would make no sense for the assets of such a supporting organization to be included in full in Y University's endowment assets for purposes of section 4968.

Accordingly, Treasury should clarify that an organization described in section 509(a)(3) with respect to a class of supporting organizations is not thereby described in section 509(a)(3) with respect to each of its supported organizations individually.18 Rather, it is described in section 509(a)(3) with respect to a particular organization only if it meets the organizational, operational, and relationship tests with respect to that particular organization considered in isolation.

The supporting organization regulations explicitly provide for situations where a Type I supporting organization is controlled by one or more organizations but supports other organizations. This is permitted “only if it can be demonstrated that the purposes of the former organizations are carried out by benefitting the latter organizations.”19 In such a case, the supporting organization would often be described in section 509(a)(3) with respect to the controlling organizations, because all of the supporting organization's activities, including support of other supported organizations in the class, would at least “carry out the purposes of” the controlling organizations. Thus, in our example above, the supporting organization would be considered described in section 509(a)(3) with respect to the X City Community Foundation but not with respect to Y University.

The current instructions to Form 4720 are ambiguous, but could be read to apply the automatic inclusion rule for supporting organizations far more broadly. The instructions define the term “related organization” to include any “[s]upporting organizations described in section 509(a)(3) that support the educational institution during the tax year.”20 Further, the instructions provide that the “intended or available for” test does not apply to the assets of any “supporting organization of the educational institution.”21 By deleting the crucial qualifier requiring the supporting organization to be described in section 509(a)(3) “with respect to” the particular institution, and replacing it with a much weaker requirement that the supporting organization support the educational institution, these instructions impermissibly broaden the scope of supporting organizations subject to automatic inclusion. This expansion automatically includes in the assets of educational institutions the assets of organizations having in many cases only very attenuated connections with the educational institutions (such as supporting organizations of community foundations and similar grantmaking charities). This cannot have been Congress's intent, and the phrase “described in section 509(a)(3) with respect to such institution”22 does not require it.

Educational institutions are popular beneficiaries, and it stands to reason that many educational institutions may be within some supporting organization's set of specified supported organizations. Furthermore, in the case of large hospital systems, federated organizations with many chapters in a group ruling, and other systems of organizations containing many public charities, a natural structure is for supporting organizations to be controlled by the national headquarters organization but have as supported organizations all of the public charities in the group. The mere fact that one of the publicly supported affiliates is an educational institution should not cause a supporting organization of the group to be treated as controlled by the educational institution. Treasury may be concerned about situations in which a university might seek to evade taking income and assets of a supporting organization into account by having that supporting organization provide nominal support to other institutions. However, if the supporting organization were still related to the educational institution, it would still have to take into account that portion of the assets “intended or available for” the educational institution. Further, if it were a Type I or Type II organization controlled by the educational institution, the educational institution's control would cause all the supporting organization's assets to be included in the section 4968 calculations. Treasury could further forestall the possibility of abuse by providing that an organization will be deemed to be described in section 509(a)(3) with respect to an educational institution so long as (a) it is described in section 509(a)(3); (b) it bears a Type I, Type II, or Type III relationship to the educational institution; and (c) substantially all, or at least a majority, of its activities are for the benefit of, perform the functions of, or carry out the purposes of the educational institution or other public charities controlled by the educational institution.

Finally, we note that the Proposed Regulations provide certain relief for educational institutions having Type III supporting organizations on December 31, 2017.23 Because such supporting organizations are not controlled by their supported organizations, and because they may not provide information to the educational institutions about their income and assets, the Proposed Regulations allow the an educational institution to take its Type III supporting organization's assets and income into account only to the extent they are intended and available for the educational institution.24 The Proposed Regulations appear to limit this relief to Type III supporting organizations based on a faulty assumption that a Type I or II supporting organization would be controlled by the educational institution. However, as these comments have explained, that is not necessarily the case. Rather, the supporting organization could be controlled by one or more other charities and simply have the educational institution as a beneficiary. Because the educational institution in such a case has the same control and information limitations as it would with respect to a Type III supporting organization, any relief provided should reach any supporting organization that is not controlled by the educational institution in question. As argued above, however, we believe that many such supporting organizations should not need relief, because they are not “described in section 509(a)(3) with respect to” the educational institution in the first place.

II. Assets Intended or Available For the Use or Benefit of an Educational Institution

A. Assets that are “Fairly Attributable” to an Educational Institution

The NPRM requests comments on when assets of a related organization, not specifically earmarked or restricted for the benefit of any particular organization, should be treated as “fairly attributable” to the educational institution or to another organization.25 Since the statute itself includes in the section 4968 calculations only those assets “intended or available for” the educational institution, “fairly attributable” assets must be assets that are intended or available for the educational institution without being earmarked or restricted for that institution.

The private foundation rules state that a grant is “earmarked if the grant is given pursuant to an agreement, oral or written, that the grant will be used for specific purposes.”26 Earmarked funds are thus subject to contractual restrictions on how they may be used. Restricted funds, similarly, are subject to legally binding limits imposed on the organization that the organization cannot unilaterally set aside. Thus, for example, donations received from third parties on the condition that the assets will be dedicated to certain uses give rise to restricted funds, which the organization would be prohibited from diverting. On the other hand, internal board decisions to allocate funds to certain purposes do not typically convert funds into restricted funds, because the board still has authority to reverse its previous decision without obtaining any outside consent. Earmarked and restricted funds are subject to legally binding requirements that the assets be used solely for the earmarked recipient. Such funds are thus a clear example of assets “intended for” a particular organization. But other assets that have been affirmatively designated for use by an educational institution, but not subject to a binding restriction, could be “intended for” the institution, and thus “fairly attributable”. Similarly, the term “available” calls to mind the doctrine of constructive receipt; in that context, assets are available to an institution when the assets may be drawn upon unilaterally by the institution.27 Again, a related organization could make funds available to be drawn upon by one or more affiliates without the funds being earmarked or restricted; in such cases the funds (or in the case of multiple affiliates having access to the funds, perhaps only a reasonably allocable portion of the funds) would be available to each affiliate and thus “fairly attributable” to the affiliate.

For instance, most parent entities of multi-entity systems or networks apply some sort of regular budget process whereby they determine how much in assets each entity will receive each year from headquarters; amounts in the budget approved for distribution to a university but not yet paid over to the university at year-end would appropriately be treated as “intended for” and thus “reasonably attributable to” the university. This is true even though the funds are not true restricted funds because the appropriation decision is an internal one and could be reversed without violating any legal obligation. Similarly, if a supporting organization holds and invests certain funds that are available for a university upon request, the funds are “available” to the university and thus “fairly attributable” to it.

In contrast, an educational institution usually will have no right to or ability to appropriate the general unrestricted funds of a controlling organization — particularly when the controlling organization has a much larger and distinct purpose. At least where the controlling organization has significantly broader purposes and does not primarily operate for the benefit of an educational institution, we believe that it requires some affirmative act by the controlling organization to make assets “intended for” or “available for” a school.

The mere fact that a related organization is likely to continue to fund an educational institution in the future does not cause its assets or income to be treated as intended or available for the educational institution in advance; rather, each year's appropriation should be treated as intended and available for the educational institution at the time it is made. Treasury should clarify that if an educational institution does not control a related organization, and the related organization primarily serves purposes other than those of the educational institution, the unrestricted, undesignated assets of the organization will not be treated as intended or available for the educational institution unless they have been affirmatively designated or appropriated for the educational institution, or made available for the educational institution to draw upon at will.

B. The Previous Year's Spending Should Not Be Treated As Presumptively Available

The NPRM asks whether a related organization's prior-year spending out of unrestricted funds for an educational institution should be treated as presumptively available in the following year.28 We believe such a rule would be counterproductive.

The section 4968 tax is calculated after the conclusion of the tax year. Thus, any annual appropriation from a related organization occurring during the year will already be known and can be taken into account either because the assets have actually been transferred to the institution (thus increasing its assets at the end of the year) or because they have been affirmatively designated for the institution but not yet transferred (in which case they could be treated as “intended for” the educational institution). As a result, there is no need to use the previous year's spending to estimate what the institution's related organizations will make available in the current year. Furthermore, to add in the previous year's distributions would arguably be double-counting, since those distributions were made to the educational institution and thus have already increased its assets.

If the Service does apply such a rule, it should distinguish between funding of normal operating budget and special capital outlays; guidance should make explicit that if a charity can demonstrate that a portion of the previous year's amounts distributed were for special-purpose capital projects rather than ordinary operational expenses, that portion will not be treated as “intended or available for” the educational institution the following year.

C. Current Distributions Do Not Reflect Availability of Underlying Assets

In its special rule for pre-existing Type III supporting organizations, the NPRM suggests that an appropriate method to determine whether a related organization's assets are “intended or available for” an educational institution is to allocate the entirety of the related organization's assets to the educational institution in proportion to the related organization's distributions during the previous year.29 But the fact that a parent organization chooses to direct a particular subsidiary to make a distribution to an educational institution in one year in no way implies that the educational institution has similar amounts available to it in future years, let alone to a corresponding portion of the related organization's entire asset base. Only the amounts actually distributed or appropriated should be treated as intended or available for the educational institution.

The suggested method would in many cases lead to strange results. Many large systems of affiliated entities have various section 509(a)(3) organizations, for-profit subsidiaries, and joint ventures or investment subsidiaries or partnerships that they control and that hold investments for the general use of the larger group. A group of hospitals might have an organization that holds assets in a self-insurance fund for the group; a convention of churches might have a subsidiary designed to provide a construction fund for new chapels; many organizations with significant assets might have entities that hold various portions of their investment portfolio. As a whole, these investment entities are used to produce earnings to finance current operations and to maintain a reasonable amount of savings for potential future needs. But any one entity may not make distributions every year, particularly if it holds illiquid investments.

Typically, when an entity in the affiliated group has particular needs, the parent could choose to use the assets of any investment affiliate to meet those needs. The selection of which entity to provide funds may be based on a variety of factors such as which entity currently has liquidity, which board members are most accessible at the time, the general purposes of the particular investment entity, and so forth. For entities that are essentially holding and investing a portion of the group's reserves, such distributions might be well in excess of the investment entity's annual income, or well below it. Thus, an investment entity might have $10 million of assets, and happen to make a distribution of $500,000 to hospital A in year 1, no distributions in year 2, and distributions of $50,000 to university B in year 2. It would produce very volatile results out of keeping with economic realities to say that all $10 million were intended and available for Hospital A in the first year and then that all $10 million were intended and available for University B in year 3. We believe the correct analysis of this situation is that in all three years the investment entity's funds are intended and available for the parent entity' use in furtherance of its overarching charitable purposes, and only the portion actually approved and directed to be used by an affiliate should be treated as intended or available for that affiliate.

III. The Proposed Regulations' Defects Could Affect Many Small Institutions Without Substantial Endowments

It is easy to focus too much on the application of the section 4968 tax to traditional large and mid-size institutions, for whom the $500,000-per-student asset threshold protects most entities, and even most systems of entities, unless they have very large endowments. However, large and mid-size institutions are not the norm. We represent clients that have affiliated educational institutions far smaller than the 5,000-student threshold that typically separates “small” from “medium” colleges. And according to 2017 data from the National Center for Education Statistics, they are not exceptional: of private nonprofit institutions of higher education that have at least 500 students and participate in Title IV programs, 35% report fewer than 1,000 students enrolled (on a twelve-month full-time-equivalent basis), and another 30% report between 1,000 and 2,000 enrolled.30 This suggests that as little as $500 million in assets could cause one in three eligible institutions to become subject to the section 4968 excise tax, and almost two-thirds would be subject to the tax if they were attributed $1 billion in assets.31

That may seem like a lot of assets, but it is not if the affiliated institution belongs to a larger control group with many affiliates across the country. Many such institutions could have these kinds of affiliations. Nearly one-third of all private educational institutions participating in Title IV programs in the 500 to 2,000 student size range are religiously affiliated, and this class of institutions includes a variety of specialty postgraduate programs such as seminaries and theological institutes that might be controlled by a church, and nursing or other schools for the health professions that might be expected to be affiliated with one or more hospital systems.32 Particularly if the breadth of the attribution rules described above in Part I of these comments caused some or all of these related organizations to be wrongly treated as “controlled by” a related small educational institution, so that all of the assets of the related organization would have to be taken into account, the asset threshold of section 4968(b)(1)(D) could easily be exceeded.

The potential problem is exacerbated by an ambiguity in the parenthetical excluding from consideration those assets “used directly in carrying out the institution's exempt purpose.”33 For educational institutions themselves, this generally means that buildings and equipment used in providing education are not included, leaving the institution's endowment and similar investment assets as the primary assets potentially subject to tax. However, the statute does not say how to apply the “used directly in carrying out the institution's exempt purpose” test in the case of a related organization with exempt purposes differing from those of the educational institution. Assets used directly for the related organization's purposes — for example, a related church's worship structures or a related hospital's facilities and medical equipment — would in many cases be excluded from section 4968 calculations because they are not intended or available for the educational institution. However, even if that is not the case because the entity is controlled by the educational institution, we believe that regulations under section 4968 should clarify that these kinds of assets are still excluded from section 4968 calculations because they meet the “used directly” test, so that a related organization's assets would never be taken into account to a greater extent than they would be taxed under section 4940 if the related organization were a private foundation. If Treasury were to reject this reasonable reading of the statute, however, a related church or hospital's assets could put the educational institution over the threshold even more quickly, even if neither the educational institution nor the related organization have significant cash and investment assets, merely because the related organization has valuable core buildings and other exempt-purpose assets that it does not use in the operation of the educational institution.

To be sure, even broad application of control rules will not always result in a tax due under section 4968. Indeed, those control rules may also apply to other educational institutions within the same network, decreasing the amount of assets attributable to any one institution because of the rule that no assets can be taken into account with respect to more than one educational institution.34 And in many cases, its related organizations' assets will not be “fairly attributable” to the educational institution. However, even if these carve-outs ultimately cause the educational institution not to be subject to section 4968, overbroad control and constructive ownership rules can still exact a significant cost on the educational institution by making it necessary for it to gather information about many other entities' assets and operations, even though it does not control them.

In a multi-level hierarchical structure, particularly one that extends to affiliates nationwide or throughout the world, it can be prohibitively difficult for an educational institution to gather information about assets available throughout the control group, let alone to figure out how affiliates' assets should be allocated among multiple educational institutions dispersed throughout the group. For example, if a small educational institution in Chicago is controlled by a diocese in Chicago, which in turn is under direct or indirect common control with every other diocese in the country, and a diocese in California also controls an educational institution, it would be virtually impossible for the educational institution in Chicago to check every diocese nationwide for assets “fairly attributable” to it, to identify the educational institution in California that may share some related institutions, and to determine how to allocate the assets between them. The situation becomes even worse if the educational institution also has to identify any supporting organizations —  including supporting organizations controlled by unrelated community foundations and the like — that might include the educational organization in its class of supporting organizations.

Properly tailored rules like those proposed by these comments improve administrability by allowing educational institutions to determine whether the section 4968 tax applies to them without having to obtain, compile, and potentially disclose exhaustive information about the assets of their entire network of affiliates. Instead, they can look only to assets of those organizations actually controlled by them, and to assets of other affiliates only to the extent they have affirmatively designated or made available for the educational institutions.

We note that when the larger control group includes churches, integrated auxiliaries, and the exclusively religious activities of religious orders, requiring an educational institution to engage in extensive analysis and monitoring of the assets of the overall group could also raise First Amendment entanglement concerns, especially if the information were required to be reported to the IRS or publicly disclosed.35 It would indirectly require the religious group to share financial information at least with the educational institution, and potentially to have information about the finances of its churches and integrated auxiliaries reported to the IRS — even though Congress has forbidden the IRS from collecting such information directly by exempting churches and integrated auxiliaries from Form 990 reporting.36 There is no evidence that Congress intended such a result, and so we believe that section 4968 should be interpreted to reach only those assets within the control group that have a substantial and tight nexus to the educational institution in question — assets that the organization can be expected to know about without having to receive information about and analyze the affairs of the entire control group.

IV. Conclusion

We appreciate the opportunity to comment on the Proposed Regulations on behalf of our clients and would welcome the chance to discuss these issues further. In their current form, the Proposed Regulations could do significant harm to large affiliated groups containing one or more educational institutions, potentially subjecting many of their assets to the section 4968 regime even though those assets are clearly intended to be used primarily for purposes other than those of the affiliated educational institutions. This is inconsistent with the intent of section 4968, as evidenced by the Senate Amendment, and must be corrected.

Please do not hesitate to call me at 801-321-4836 if you have any questions about these comments to the Proposed Regulations.

Very truly yours,

Michael W. Durham
KIRTON McCONKIE
Salt Lake City, UT

FOOTNOTES

1See Department of the Treasury, Guidance on the Determination of the Section 4968 Excise Tax Applicable to Certain Private Colleges and Universities, Docket No. REG-106877-1884, Fed. Reg. 31,795 (July 3, 2019).

2Tax Cuts and Jobs Act, H.R. 1, 115th Cong., sec. 5103, § 4969(d) (as passed in the House on Nov. 16, 2017), available at https://www.congress.gov/bill/115th-congress/house-bill/1/text/eh (last visited Sept. 27, 2019).

3See 26 U.S.C. § 4968(d)(1)(B); cf. Tax Cuts and Jobs Act, H.R. 1, 115th Cong., sec. 13701(a), § 4968(d)(1)(B) (as passed by the Senate Dec. 14, 2017), available at https://www.congress.gov/bill/115th-congress/house-bill/1/text/eas (last visited Sep. 27, 2019).

4A similar definition is used for determining whether a tax-exempt organization must disclose another nonprofit corporation or similar entity “without owners or persons having beneficial interests” as a “related organization.” See Internal Revenue Service, 2018 Instructions to Schedule R, at 2.

5See Taxpayer Relief Act of 1997, sec. 1041(a), § 512(b)(13), Pub. L. No. 105-34, 111 Stat. 787, 938-39 (1997).

6Treas. Reg. § 1.512(b)-1(l)(4).

7H.R. Rep. No. 105-148, at 492 (1997); H.R. Rep. No. 105-220, at 562 (1997).

8See PLR 199941048 (July 20, 1999).

9Treas. Reg. § 53.4941(d)-1(b)(5).

10See, e.g., Kedroff v. Saint Nicholas Cathedral of Russian Orthodox Church in North America, 344 U.S. 94, 116 (1952) (declaring that under the First Amendment's church autonomy doctrine that religious organizations possess “an independence from secular control or manipulation — in short, power to decide for themselves, free from state interference, matters of church government as well as those of faith and doctrine”); Corp. of the Presiding Bishop v. Amos, 483 U.S. 327, 341 (1987) (“[R]eligious organizations have an interest in autonomy in ordering their internal affairs, so that they may be free to: select their own leaders, define their own doctrines, resolve their own disputes, and run their own institutions.”).

11I.R.C. § 318(a)(2).

12I.R.C. § 318(a)(3).

13Prop. Reg. § 53.4968-1(c)(1)(iii), 84 Fed. Reg. 31,806, 31,808 (July 3, 2019).

14See Internal Revenue Service, 2018 Instructions for Schedule R (Form 990), at 2-3.

15See Notice 2018-26, § 3.01, 2018-16 I.R.B 480, 484.

17I.R.C. § 509(a)(3)(A)-(B).

18In contrast, the Form 990 appears to treat all organizations in a class of supported organizations as related organizations, because its definition of relatedness makes a supported organization related to each of its supporting organizations, and a supporting organization related to each of its supported organizations. See Internal Revenue Service, 2018 Instructions for Schedule R (Form 990), at 1.

19Treas. Reg. § 1.509(a)-4(g)(ii).

20See Internal Revenue Service, 2018 Instructions for Form 4720, at 18.

21See id. at 19.

23Prop. Reg. § 53.4968-1(c)(3)(ii), 84 Fed. Reg. at 31,808.

24See id.; see also NPRM, 84 Fed. Reg. at 31,802-31,803.

25NPRM, 84 Fed. Reg. at 31,802.

26Treas. Reg. § 53.4945-2(a)(5)(i).

27Constructive receipt of income hinges on whether the income is “available” to the recipient:

Income although not actually reduced to a taxpayer's possession is constructively received by him in the taxable year during which it is credited to his account, set apart for him, or otherwise made available so that he may draw upon it at any time, or so that he could have drawn upon it during the taxable year if notice of intention to withdraw had been given. However, income is not constructively received if the taxpayer's control of its receipt is subject to substantial limitations or restrictions. Thus, if a corporation credits its employees with bonus stock, but the stock is not available to such employees until some future date, the mere crediting on the books of the corporation does not constitute receipt.

Treas. Reg. § 1.451-2(a) (emphasis added). In this provision, assets are “available” for a person when the person has power to obtain it upon request. Similarly, several sections governing pension plans and similar deferred compensation arrangements make clear that the assets and earnings, even if held in an account for a particular individual, will not be considered income to the individual until they are paid, distributed, or “made available” to the person. See, e.g., Code §§ 402, 403, and 457. Finally, we have found that the few courts using the unusual phrase “available for the use or benefit” or similar language often use the phrase to describe assets that have so passed into the taxpayer's dominion and control that they can be included in the taxpayer's income. See, e.g., Furstenberg v. Comm'r, 87 T.C. 755, 789 (1984) (“[B]ecause the accumulation distribution was actually received by petitioner and available for her use and benefit before her expatriation, we think it is includable in her income. . . .”); Lawyers' Title Ins. Corp. v. Early, 30 A.F.T.R. (RIA) 1757 (E.D. Va. 1942), aff'd, 132 F.2d 42 (4th Cir. 1942) (“The amount of reserves representing unearned premiums is not 'gain' and is not available for the use, benefit or disposal of the plaintiff; accordingly it is not 'income' within the meaning of the 16th amendment.”). 

28NPRM, 84 Fed. Reg at 31,802.

29Prop. Reg. § 53.4968-1(c)(3)(ii), 84 Fed. Reg. at 31,808.

30Data available for download from Integrated Postsecondary Education Data System, at https://nces.ed.gov/ipeds/use-the-data (last visited Oct. 1, 2019).

31The available data is not precisely matched to section 4968 definitions because Title IV participating institutions are not the same as Title IV eligible institutions, which are the group subject to section 4968, and because they do not take into account the fact that many institutions have at least some students who do not pay tuition, so that some institutions with more than 500 students will not have enough tuition-paying students to qualify as applicable educational institutions. Still, the figures should be considered broadly illustrative.

32Data available for download from Integrated Postsecondary Education Data System, at https://nces.ed.gov/ipeds/use-the-data (last visited Oct. 1, 2019).

35See Church of Scientology v. City of Clearwater, 2 F.3d 1514, 1535 (11th Cir. 1993) (holding that rules requiring disclosure of all expenditures of a church created impermissible government entanglement); Surinach v. Pesquera de Busqets, 604 F.2d 73, 74 (1st Cir. 1979); Archbishop of Roman Catholic Apostolic Archdiocese v. Guardiola, 628 F. Supp. 1173, 1183 (D.P.R. 1985) (“Compelling the Church to produce balance statements, statements of income and expenses, salaries paid to religious personnel, [and] depreciation of lots and buildings and other equipment, creates the likelihood of entanglement that the First Amendment proscribes.”).

END FOOTNOTES

DOCUMENT ATTRIBUTES
Copy RID