Peter Gulia

jpod, while we might like to think there must be a citation, I haven't yet found it. Thank you for pointing out Marty Slate's article; I'll look. ESOP Guy, thank you for confirming what we saw (and saw missing) in some IRS publications. QP_Guy, the "definitely determinable" idea is some support IF the plan provision an administrator failed to follow is one that could affect the allocation of a profit-sharing contribution. (For this conversation, I'm ignoring pension benefits.) But it doesn't explain why there is a tax-qualification failure if the provision not administered cannot affect an allocation of a contribution. Thank you, everyone, for continuing to help me.

MoJo, thank you for this further analysis.

Mojo and Belgarath, thank you for helping. An administrator’s, trustee’s, or other fiduciary’s failure to administer an ERISA-governed plan “in accordance with the documents and instruments governing the plan” (insofar as those documents are consistent with ERISA’s title I and title IV) is, under ERISA § 404(a)(1)(D), a breach of the fiduciary’s responsibility. That’s so even if what is done otherwise breaches no duty under § 404(a)(1)(A)-(C). But such a fiduciary breach doesn’t explain why a failure to administer a plan according to the written plan is a failure of the Internal Revenue Code § 401(a) conditions for treatment as a tax-qualified plan. While Basch Engineering, Buzzetta Construction, and Ludden recite that a plan must meet the tax-treatment conditions in operation, none of them based its finding on a mere failure to follow the written plan that was not also a failure of a particular tax-qualification condition. I’m thinking about situations in which the only ground for tax-disqualifying the plan is a failure to follow a provision in the plan’s governing document when the provision is not one the Internal Revenue Code requires to be stated in a plan’s document. So far, my reasoning goes like this: Under 26 C.F.R. § 1.401-1(a)(2), a qualified plan must be “a definite written program and arrangement which is communicated to the employees and which is established and maintained by an employer[.]” A failure to administer a plan according to the “definite written program” might mean the employer does not maintain that plan. It would be nice to find a law source that states the point more directly, or at least expressly adopts the reasoning. But maybe there’s nothing to be found. Again, Mojo and Belgarath, thank you for helping.

I assume the law requires operating a plan according to its written plan as a condition for tax-qualified treatment. I'm just looking to find a law source that says so. Thank you for the pointers to the IRS's explanations. But I'm looking for a regulation or something that has the effect of law. And it's not because I doubt the point of law. Rather, I hope to support it with a citation.

A further point to consider if the plan is self-funded. A stop-loss insurance contract might not provide anything for a continuation that is not compelled under the public law (rather than the written plan's provisions) that applies to the employer and its plan. Lacking stop-loss insurance could leave an employer exposed to claims on what might be (through adverse selection, and perhaps other factors) a higher-risk population.

A retirement plan, to tax-qualify under Internal Revenue Code § 401(a), must meet all conditions not only in the written plan but also in actual administration according to that written plan. The Internal Revenue Service’s Employee Plans Compliance Resolution System presumes the point; the Revenue Procedure defines an Operational Failure as one “that arises solely from the failure to follow plan provisions.” Rev. Proc. 2019-19, 2019-19 I.R.B. 1086, 1099 § 5.01(2)(b) (May 6, 2019). (Thank you, C.B. Zeller.) Unlike some other points made in the Revenue Procedure, this one cites no Treasury regulation as support for the point. Writers often say a plan must tax-qualify not only “in form” but also “in operation”. There are Treasury regulations and court decisions that support the in-form point. But I’m not (yet) seeing a regulation, court decision, or other law source that clearly states or supports the in-operation point. I’m hoping BenefitsLink mavens will teach me. Will you please help me?

If January 2020 arrives without the Treasury department having done anything (beyond publishing a proposed rule), is a plan's sponsor/administrator free to not change the plan, and not change its administration? About those changes for which Congress directed the Secretary of the Treasury to make or change a rule but he has not done so, does Congress's Act require a plan's sponsor or administrator to do anything?

Anything new on this in the past eight months?

About QDROphile’s last point: My article “May an employer restrict who can amend the plan?” in the December 2018 issue of Wolters Kluwer’s 401(k) Advisor explains that the Supreme Court interpreted ERISA to treat “The Company may amend the Plan.” as an amendment procedure that meets the command of ERISA § 402(b)(3). What do BenefitsLink mavens think about a TPA inviting a plan’s sponsor to consider whether it wants to add to a preapproved document’s general provision some further details about which people can or can’t amend the plan, and what kind of written act is valid to amend the plan? Or is that impractical for TPAs too?

Beyond Luke Bailey's suggestion, consider also the plan's provisions about what act and writing is sufficient to amend the plan. Many plan documents allow as an amendment anything that is the act of the plan's sponsor under the law that governs that organization. A governing body's resolution, if expressed or recorded in writing, might be a plan amendment. Whether something is a plan amendment and what effect an amendment has on the plan's tax treatment (or whether a document user may rely on an IRS pre-approval) are distinct questions. About making allocation provisions that are sufficiently determinable to meet a tax-qualification condition, one might specify the HCEs' maximum elective-deferral percentage or the rule or formula by which it is determined. C.B. Zeller wisely suggests some related issues and requirements.

ESOP Guy, thank you for the thoughtful information.

Even if Bird is right and furnishing the information is harmless, I lack a good way to prove that beliefs about how the IRS might use the information are mistaken. So, my question remains: If an applicant avoids the website system, and on the paper form fills-in for the “responsible party” information a corporation’s or other non-natural person’s name and Employer Identification Number, will the IRS process it and issue the retirement plan trust’s identifying number?

This discussion shows how a beginning-of-year count of participants, instead of merely repeating the preceding year’s end-of-year count, might add participants who had their entry date on the first day of the year. Does the opposite change happen? Could there be someone who was a participant on the last day of the preceding year, but is not a participant on the first day of the next year—because the plan paid (or otherwise distributed) the participant’s entire benefit on the last day of the preceding year?

I’ve heard from some clients that some businesspeople believe the IRS would, if the IRS thinks something is due from the retirement plan’s trust, use the “responsible party” information to pursue that human. Based on this, no one wants his or her name and Social Security Number on an SS-4 application. ERISAcpr, if an applicant avoids the website system, and on the paper form fills-in for the “responsible party” information a corporation’s or other non-natural person’s name and Employer Identification Number, will the IRS process it and issue the retirement plan trust’s identifying number?

If the employer seeks to exclude employees (rather than retirees) from its regular group health plan, consider the employer's powers under State law, age-discrimination law, and law designed to make Medicare secondary to employers' group health plans.

Beyond the issues you mention, consider whether you want to politely suggest that the employer seek its lawyer’s advice about whether the employer has power to establish a plan of the kind it wants. A governmental employer other than the State itself has only the powers provided by State law. Many States’ laws limit a local government’s powers in establishing employee-benefit plans. Even if the issue is beyond your scope, you might prefer to flag the issue. If the employer implements something beyond the employer’s powers and gets caught (often by an auditor), you might avoid or at least answer the rhetorical question “Why didn’t you tell me . . .?

Thank you, ESOP Guy and Belgarath. Anyone with a different outlook?

29 C.F.R. § 2520.104-46 provides its excuse from an audit of a plan’s financial statements only if, among other conditions, at least 95% of the plan’s assets are qualifying plan assets—much of which involves regulated banking, insurance, and securities businesses. Imagine a small-business retirement plan with 100% of its assets in non-qualifying assets. An officer of the plan’s sponsor serves as the plan’s trustee. If a TPA goes about its work normally, how likely or unlikely is it that a TPA would see information from which the TPA would know that the assets don’t qualify for a § 2520.104-46 waiver?

If your client wants an IRA asset not to be counted for one or more Medicaid purposes, a relevant measure might be none of the tables established for minimum-distribution rules under Internal Revenue Code of 1986 § 401(a)(9) and provisions that refer to it. A State’s Medicaid program might use different rules and measures for how much one may, must, or must not receive from a retirement account. Some of Medicaid’s law is Federal law, and some is State law. You might find the law in a combination of statutes, rules or regulations, other administrative-law interpretations, and court decisions. The law varies considerably from State to State.

Luke Bailey, thank you for the further ideas.

CBZ and CuseFan, thank you for confirming I wasn’t overlooking an exception or excuse. CBZ, thank you for the helpful citation. The situation described above shows how a plan sponsor might needlessly burden itself by using a preapproved document without getting advice about provisions a user might change or delete. The plan-document sentence quoted above might be an “administrative provision” a user might delete without losing reliance on the IRS’s opinion letter. Had it been deleted, the plan’s administrator might undo the prohibited transactions without also needing an IRS correction procedure for a qualification failure.

An individual-account § 401(k) retirement plan provides participant-directed investment. Following participants’ directions, the plan’s trustee engages in transactions with a party-in-interest. No exemption applies, and there is no doubt these are prohibited transactions under ERISA § 406 and Internal Revenue Code § 4975. Yet these transactions are not necessarily an IRC § 401(a)(2) exclusive-benefit violation. Among other facts, each investment has an above-market return, and the plan’s counterparty has strong creditworthiness and liquidity. The plan’s governing document includes this: “The Trustee shall not engage in any prohibited transaction within the meaning of the Code and ERISA.” Does something that might not have tax-disqualified a plan have that effect because the plan’s fiduciaries failed to administer the plan according to its governing document?

Revenue Ruling 2002-45 [http://www.irs.gov/pub/irs-drop/rr02-45.pdf] describes a restorative payment (the ruling’s antidote against counting an amount as a contribution) as a payment “made to restore losses to the plan resulting from actions by a fiduciary for which there is a reasonable risk of liability for breach of a fiduciary duty under title I of the Employee Retirement Income Security of 1974 (ERISA)[.]” Beyond the examples given in the ruling, the IRS in practice has treated a payment as restoration if the employer made a written finding that the selection or negotiation of the insurance or investment contract was (or might have been) a breach of the employer’s fiduciary responsibility, whether under ERISA or other law, and the finding is plausible. The Treasury department’s interpretation requires also that “participants who are similarly situated are treated similarly with respect to the [restorative] payment.” For a limitation year that began or begins on or after July 1, 2007, the ruling’s principle is included in the annual-additions-limit rule. 26 C.F.R. § 1.415(c)-1(b)(2)(ii)(C). https://www.ecfr.gov/cgi-bin/text-idx?SID=9ee62d943d4b498039cba586af9d3fc9&mc=true&node=se26.7.1_1415_2c_3_61&rgn=div8 The Treasury adopted my suggestion about looking beyond ERISA to other Federal law, and to State law. The key driver is that there is “a reasonable risk of liability”.

Imagine that a TPA has made the business decision, and a service agreement specifies the communications services the TPA provides (and, perhaps, a fee for those services). And assume the TPA prefers to avoid discretion and otherwise to provide services only as a non-fiduciary. In those circumstances, the TPA might prefer that a plan's administrator specify the methods the TPA uses to identify whether an inquirer is a participant. That returns to my question: What identifiers and methods might a TPA suggest?

For those TPAs that are or might become willing to talk with a participant: What identity controls might a TPA use to satisfy yourself that it's reasonable to believe an inquirer is the participant (or even a participant) rather than an impostor? Big recordkeepers use regimes of controlled identifiers, passwords, and other restraints. But which methods are feasible for a TPA? Are there software solutions that are useful to help a TPA check an inquirer's identity? (I'm not asking anyone to reveal a particular company's methods. Rather, I'm asking generally about what TPAs might do.)