Firm Suggests Alternative Approach Under Proposed Governmental Plan Regs

This letter provides comments on the Advance Notice of Proposed Rulemaking ("ANPRM") under section 414(d) of the Internal Revenue Code that was published on November 8, 2011. 76 Fed. Reg. 69172. This letter solely addresses the criteria for determining whether an entity is an instrumentality of one or more states or political subdivisions thereof. This letter is submitted on behalf of clients of our firm that have established and maintained retirement plans under section 401(a), as well as eligible deferred compensation plans under section 457(b), based on the understanding and belief that they are properly classified as governmental instrumentalities for all purposes under the Code and ERISA and that their plans are governmental plans under section 414(d). This letter focuses on the term "instrumentality" because we believe the scope of that term, rather than the term "agency," represents the effective boundary of section 414(d).

Since a retirement plan is structured to operate on a long-term basis and the consequences of classification or misclassification as a governmental plan are highly significant, we believe regulations under section 414(d) will serve a constructive purpose only to the extent such regulations provide greater certainty than currently exists without drawing lines that would deny instrumentality status for narrowly conceived reasons. We recognize the considerable thought and effort reflected in the ANPRM, and that the ANPRM was merely outlining a possible structure for regulations that would govern determinations of agency or instrumentality status. With all due respect, however, we do not believe the multi-factor approach to this determination that is proposed in the ANPRM, including the distinction between main and minor factors, would promote greater certainty in this area. Except to the extent of possible color-matching of an organization's facts with the examples set forth in the ANPRM, the ANPRM provides no meaningful guidance as to how an organization's status as a governmental agency or instrumentality would be determined if some but not all of the major factors or other factors set forth in the ANPRM were satisfied. Furthermore, assuming that regulations under section 414(d) were made applicable, or replicated, under section 3(32) of ERISA, regulations following the indefinite facts-and-circumstances test of the ANPRM would do little to guide judicial review of ERISA claims or state law claims based on ERISA exemption.

Moreover, although not expressly acknowledged in the ANPRM, certain of its main factors and/or its articulation of or commentary on such factors create the impression that the ANPRM is seeking to very narrowly confine the circumstances in which section 414(d) applies to plans of entities other than political subdivisions and the agencies thereof. If that is the intent, we question its legal basis. The disjunctive use of the term "instrumentality" in section 414(d) of the Code and section 3(32) of ERISA, and the longstanding use of that term under other sections of the Code, reflects congressional recognition that there are entities that are not political subdivisions or government agencies, or integral parts thereof, that nonetheless exist as arms or extensions of state and local governments and should be treated in the same manner as such government units for purposes of the relevant statutes, We can see no evidence that Congress intended the term "instrumentality" to have different meaning or scope under section 414(d) than its accepted meaning under the Code prior to the enactment of section 414(d), and we cannot see any policy reason why Congress might have intended such a distinction. The implication that an organization might be classified as a governmental instrumentality for some purposes under the Code but not others would only increase uncertainty and confusion, all the more so if the distinction were based on the kind of indefinite and discretionary weighing of factors required under the test set forth in the ANPRM. Simply put, we do not think this kind of hair-splitting is justified or administrable.

In our view, the ultimate inquiry under section 414(d) should be the same as that under other provisions of the Code that provide rules for instrumentalities of state and local government -- whether, by reason of government control and purpose, an entity is more properly characterized as a creature of one or more states or political subdivisions than as a private organization. Specifically, we recommend that an organization's classification as an agency or instrumentality of state or local governments be determined under the same or essentially the same criteria that apply for purposes of determining whether it is a wholly-owned instrumentality of a state or political subdivision under section 3121(b)(7) of the Code and thereby eligible for exemption from the OASDI portion of the FICA tax. One practical consequence of applying the same criteria is that, if an organization were determined by the IRS (or a court) to be an instrumentality under section 3121(b)(7), it would be deemed to be an agency or instrumentality under section 414(d). There are a number of reasons why we believe the standards should be consistent.

In the first place, we can see no reason why the incidental differences in wording between sections 414(d) and 3121(b)(7) would support differing interpretations. Although section 3121(b)(7) is limited to "wholly-owned" instrumentalities, the ANPRM rightly indicates that section 414(d) is inapplicable to an organization that has any private ownership. (Indeed, it is unclear to us why the ANPRM lists this as an "other" factor rather than a dispositive factor if not satisfied.) Further, while section 3121(b)(7) refers to instrumentalities "of any one or more of the foregoing" and section 414(d) more imprecisely refers to agencies or instrumentalities "of any of the foregoing," it seems clear and is acknowledged by the ANPRM that section 414(d) may apply to organizations that are instrumentalities of multiple governmental units. (The ANPRM's further commentary on this issue is discussed below.) Finally, although section 414(d) refers to an "agency or instrumentality," whereas section 3121(b)(7) refers only to instrumentalities, it is impossible to envision the circumstance in which an organization that failed to qualify as an instrumentality under section 3121(b)(7) might nonetheless be classified as an agency of state or local government under section 414(d).

As a matter of policy, since the scope of section 3121(b)(7) directly and materially affects the retirement benefits provided to the employees to which it applies, we can see no reason why section 414(d) should be given narrower scope or why Congress would have intended that it have narrower scope than section 3121(b)(7). As in the case of section 414(d), the employment tax exemption provided by section 3121(b)(7) is primarily based on principles of federalism.¹ When section 414(d) was enacted in 1974, the employees of section 3121(b)(7) instrumentalities did not accrue Social Security benefits and their wages were not subject to the OASDI portion of FICA tax unless the employer had entered into a coverage agreement under section 218 of the Social Security Act, without regard to whether any form of retirement plan was provided for such employees. With section 3121(b)(7) instrumentalities having been given the same complete responsibility for the retirement income security of their employees as states and their political subdivisions, it is hard to conceive how Congress could have contemplated that section 414(d) would not apply to those instrumentalities of state and local government.

After the effective date of the amendments to section 3121(b)(7) made by the Omnibus Budget Reconciliation Act of 1990, the employees of such instrumentalities are exempt from Social Security coverage (and their wages are exempt from OASDI tax) only if they are also participants in a retirement system that meets the minimum requirements set forth in section 31.3121(b)(7)-2 of the regulations. Needless to say, however, there is a substantial difference, both qualitatively and quantitatively, between a defined contribution plan funded with a combination of employer and employee contributions totaling 7.5% of compensation up to the Social Security wage base, which qualifies as a retirement system under section 31.3121(b)(7)-2 of the regulations, and Social Security coverage at a combined OASDI cost of 12.4% of wages up to the Social Security wage base. Thus, even under current law, it is fair to say that the applicability or inapplicability of section 3121(b)(7) more directly affects the retirement income of affected employees than does the applicability or nonapplicability of section 414(d).

Furthermore, for purposes of determining instrumentality status under section 3(32) of ERISA, the corresponding provision to section 414(d), the courts have generally looked to the Service's interpretation of section 3121(b)(7). Revenue Ruling 57-128, 1957-1 C.B. 311, sets forth the following six-factor test for determining whether an entity is an agency or instrumentality under section 3121(b)(7):

Although these court decisions obviously do not preclude the adoption of different regulations under section 414(d) (or section 3(32) of ERISA), they provide confirmation that the courts have not viewed the six-factor test as inconsistent with the purpose and terms of the ERISA definition of governmental plans. And, while the six-factor test requires a limited form of facts-and-circumstances analysis, in none of the cases have the courts struggled to determine how the six-factor test should apply to the organization under consideration. Apart from being time-tested and accepted by the courts in determining whether a plan is a governmental plan, we believe the six-factor test more directly addresses the core characteristics of an instrumentality of state or local government than the test set forth in the ANPRM and does so without attempting to specify exactly how those characteristics must be evidenced.

Relevance of treatment of employees as government employees: Whether an organization's employees are treated as government employees is not included in the six-factor test of Revenue Ruling 57-128, whereas it is listed as a main factor in the ANPRM. Further, this is listed as a factor in Revenue Ruling 89-49, 1989-1 C.B. 117, which is the only prior published guidance addressing section 414(d) itself. We recommend that this factor be eliminated or, at a minimum, that it be made clear that the absence of this factor does not militate against an organization's instrumentality status.

We accept that, if an entity's employees are treated as government employees for other purposes (e.g., civil service protections)², there is a high likelihood that the entity would be properly classified as a governmental agency or perhaps as an instrumentality. However, we believe such classification would be proper only for reasons directly related to the entity itself. The statute addresses the status of the employer, not its employees. That point was directly addressed in Berini v. Federal Reserve Bank of St. Louis, supra, in which the court followed the Second Circuit's decision in Rose and applied the six-factor test of Revenue Ruling 57-128 in dismissing the plaintiff's contention, based on Revenue Ruling 89-49, that the plan in question should not be classified as a governmental plan because employees of the Federal Reserve Bank were not treated as employees of the federal government. The Court stated: "Congress did not define 'governmental plan' by reference to the status of employees; it simply states that a plan established by 'any agency or instrumentality' of 'the Government of the United States' [the relevant government under the facts of the case] was a governmental plan and thus exempt." 420 F. Supp. 2d at 1029.

In applying the Rose test, the court in Berini declined to follow the decision of the D.C. Circuit in Alley v. Resolution Trust Corp., 984 F.2d 1201 (D.C. Cir. 1993), in which the court found "no indication that Congress meant the governmental plan exemption to reach an entity that relates to its employees as would a private business -- an entity whose employees are not subject to laws governing public employees generally." 984 F.2d at 1206. Notably, the court in Alley had itself acknowledged that, because of federalism concerns, "a Rose-style test focusing broadly on the extent of governmental contacts may be more appropriate where state-affiliated entities are concerned." 984 F.2d at 1205 n. 11. In Koval v. Washington County Redevelopment Authority, 574 F.3d 238 (3d Cir. 2009), the Third Circuit squarely ruled, based on federalism concerns, that the Alley approach is not suited to cases involving state entities. In contrast with the Second Circuit's decision in Rose, the court in Koval did not address the distinction between political subdivisions and agencies and instrumentalities. In concluding that a county redevelopment authority was a political subdivision for ERISA purposes, the Court looked to the Supreme Court's decision in National Labor Relations Board v. Natural Gas Utility Dist. of Hawkins County Tennessee, 402 U.S. 600 (1971), under which an entity is treated as a political subdivision if it is either (1) "created directly by the state, so as to constitute departments or administrative areas of the government," or (2) "administered by individuals who are responsible to public officials or to the general electorate." 402 U.S. at 604-605.³

This letter is not submitted to recommend criteria for identifying instrumentalities of the federal government. Because of the federalism concerns addressed by the Third Circuit in the Koval decision (and acknowledged by the court in Alley), we accept that it might be more appropriate to give weight to the classification of employees with respect to entities affiliated with the federal government. However, even with respect to such entities, we believe the analysis in Berini has much stronger legal basis than that of the court in Alley. The facts of Berini are sufficient to make the point: it strains the imagination to suggest that a Federal Reserve Bank is not an agency or instrumentality of the federal government because its employees are not treated as employees of the federal government.

In any case, we think it would be inappropriate for regulations under section 414(d) to create any implication or suggestion that the determination of an organization's status as an instrumentality of state or local government is negatively affected by the fact that its employees are not otherwise treated as government employees. It is squarely within the province of the states and their political subdivisions to decide which of their employees and the employees of their instrumentalities should be accorded civil service or other employment protections, and the fact that a given state or political subdivision has determined not to extend such protections to a class of employees should not be taken as an indication that the employing entity is not otherwise a public instrumentality.

Relevance of sovereign powers: For much the same reasons, we do not believe the determination of an organization's status as an instrumentality of state or local government should be negatively affected by the fact that it does not exercise sovereign powers, which is listed as a main factor in the ANPRM. The possession of sovereign powers has been the litmus test for determining whether an entity is properly classified as a political subdivision for federal tax purposes. See Revenue Ruling 78-138, 1978-1 C.B. 314 ("The term political subdivision has been defined consistently for all federal tax purposes as denoting either a division of a state or local government that is a municipal corporation or a division of such state or local government that has been delegated the right to exercise sovereign power."); compare Estate of Shamberg v. Commissioner, 144 F.2d 998 (2d Cir. 1944), cert. denied, 323 U.S. 792 (1945) (Port of New York Authority to be a political subdivision because it had been delegated substantial sovereign powers), with Philadelphia Nat'l Bank v. United States, 666 F.2d 834 (3d Cir. 1981) (holding Temple University not to be a political subdivision of the state of Pennsylvania because it was authorized to exercise only one small aspect of the police power).

Based on the Third Circuit's decision in the Koval case and the Seventh Circuit's decision in the Shannon case, discussed above, an entity might be classified as a political subdivision under ERISA even if it possesses no sovereign powers. But without regard to how the term political subdivision is defined for these purposes, since section 414(d) applies with equal force to instrumentalities of state or local government that are not political subdivisions, we believe it would be wrong for the regulations to create any implication that an organization lacks an attribute of instrumentality status because it does not possess any of the sovereign powers that political subdivisions ordinarily possess. If there is a perceived need to do so, the regulations under section 414(d) might be cast in a form that would identify an organization's possession of sovereign powers as a positive indicator that it is a governmental employer (whether a political subdivision or otherwise) within the scope of section 414(d). However, under the form of the multi-factor test set forth in the ANPRM, the absence of sovereign powers might be viewed as a strike against recognition of instrumentality status, and we can see no justification for that position.

Indicia of control by one or more states or political subdivisions: In our view, the most important indication that an organization exists and serves as an instrumentality of state or local government is its ultimate control by public authorities, rather than private interests. However, given the wide variety of governance and other relationships between state and local governments and entities that serve as arms of such governmental units, we think it would be preferable for the indicia of control to be described in less detailed and specific terms than in the ANPRM. For example, whereas Revenue Ruling 57-128 simply asks "whether control and supervision of the organization is vested in public authority or authorities," the ANPRM lists public control of an entity's governing board as one factor, public nomination and election of the entity's governing board as a separate factor, and public control of the entity's operations as a third factor (listing each of the first two as separate "main" factors and the third as an "other" factor).

Similarly, whereas Revenue Ruling 57-158 broadly refers to the necessity of statutory or other authority for the creation and/or use of the instrumentality, the ANPRM provides that a specific enabling statute is an indication of an organization's governmental status only if it "prescribes the purposes, powers, and manners in which the entity is to be established and operated" and then only if the entity is not incorporated under a State's general corporation laws. We find it difficult to understand why it would not be supportive of instrumentality status if a state statute mandated the creation of a state or municipal-controlled corporation for prescribed governmental purposes, but the statute did not also prescribe the specific "manner" in which the organization was to carry out its purposes or if the statute directed that the corporation be organized under the state's general statute governing nonprofit corporations. We believe this is precisely the kind of hair-splitting that regulations under section 414(d) should avoid.

In the same vein, the ANPRM is highly specific and exclusive in terms of the financial relationships that may be taken into account as supportive of instrumentality status, whereas Revenue Ruling 57-158 simply looks to the degree of an organization's financial autonomy and the source of its operating expenses. For example, if a political subdivision has the authority to direct an organization to pay surplus cash over to the political subdivision, that would properly be taken into account as a clear indication of instrumentality status under Revenue Ruling 57-158, but does not fit under any of the specific terms of the ANPRM. On the other hand, the ANPRM lists as a main factor that a state or political subdivision has fiscal responsibility for the liabilities of the entity, including for the funding of its employee benefit plans; we expect that this factor would not be satisfied by the vast majority of entities that have been recognized as instrumentalities of state or local government under section 3121(b)(7). While we accept that the presence of this atypical factor would be a positive indicator of instrumentality status, it is difficult to see why the absence should be counted as a strike against instrumentality status. Similarly, the ANPRM would count an entity's funding through public sources as a factor supporting instrumentality status only if the entity was not paid under a government contract or funded through governments grants, but this limitation would only apply if the entity was "not otherwise an agency or instrumentality," Surely this level of specificity and circularity cannot be regarded as an improvement on the more general inquiry relating to financial autonomy and the source of operating expenses posed by Revenue Ruling 57-128.

Finally, though one of the examples in the ANPRM would treat an entity controlled by five states as an instrumentality of the states, the preamble to the ANPRM suggests that control by states or political subdivisions might be a "mere legal possibility" if "control is shared among so many governing entities that none of them can be said to be responsible in the event of a failure to exercise control." We fail to understand the policy or rationale underlying this statement, much less how it would be given operative effect under regulations. If multiple states or political subdivisions have determined that they can more efficiently and effectively carry out certain governmental functions through a commonly controlled entity, it would pose obvious federalism concerns for federal regulators to judge whether governmental control of that entity was too dispersed to be consistent with its classification as an instrumentality of the states or political subdivisions. Moreover, if an organization was controlled by, say, fifty rather than five states, we cannot see why that would lessen the reality of governmental control over the organization. Is the ANPRM suggesting that, unless officials from each state supervise or actively participate in the organization's decisionmaking, state control would not be meaningful? Why not? Lastly, we cannot imagine how would it be determined, other than on an impressionistic basis, whether an organization was controlled by too many states or political subdivisions. We do not mean to be overly critical, but if there is an objective to provide principled and coherent guidelines that can be applied with reasonable consistency and a minimum of subjectivity, we believe this statement in the preamble exemplifies the shortcomings of the ANPRM.

We believe the ANPRM is testament to why an effort to provide detailed and fine-tuned guidelines for determining whether an entity is classified as an instrumentality of state or local government would be ill-advised. In our view, the multi-factor and multi-tiered test set forth in the ANPRM raises as many questions as it seeks to answer, it fails to distinguish positive from negative factors, it defines certain factors in specific terms that appear to foreclose consideration of other relevant facts, it describes other factors in ways that invite arbitrary determinations, and it would fail to give decisionmakers (whether employers, regulators, or the courts) a coherent sense of the ultimate inquiry. In our view, the broader strokes of the six-factor test set forth in Revenue Ruling 57-128 better serve to frame the core question whether an entity is truly an arm of government, without drawing arbitrary or indefensible lines or attempting to address every factual variation. This is not to say that the six-factor test is perfect or that it eliminates all uncertainty, nothing will. However, that test has long served as the basis for determining instrumentality status under a section of the Code that directly relates to the retirement benefits of public employees, it has been accepted by the courts for purposes of making governmental plan determinations under ERISA, and we can see no justification for inventing a different wheel for purposes of regulations under section 414(d).

² Apart from such employment protections, it is unclear what the ANPRM means when it refers to whether an organization's employees are treated as government employees. For example, if an organization's employees participate in a public employees retirement system, would that to be taken as an indication of such treatment, or would it merely pose the possibility that such employees must cease to participate in the public employees retirement system?

³ The Seventh Circuit had earlier followed essentially the same analysis in Shannon v. United Services Automobile Ass'n 965 F.2d 542 (7^th Cir. 1992), giving no weight to to whether the organization's employees were otherwise classified as government employees.