Peter Gulia

If the Form 5500 instructions are ambiguous and there’s a choice about which value to report, one practical approach (for a plan administrator that hasn’t already decided which of the many valuation and accounting methods is most informative for the purposes of the report) might be to ask its auditor which value the auditor would prefer to support a professional opinion that the plan’s financial statements “fairly present” the plan’s net assets available to pay plan benefits. In theory, the plan administrator should select (if Form 5500 instructions don’t control the point) the method that it, as a prudent-expert fiduciary, finds is most informative to readers of the Form 5500. But often a real-world path of least resistance is for a plan administrator to say that it reasonably relied on an accounting expert’s professional opinion. The accounting guidance suggests that a CPA shouldn’t go along with a fully-benefit-responsive book-value position unless, along with other conditions, an “event” that would limit the plan’s right to transact at book value was not “probable of occurring”. For many a plan’s report, there might be about nine months between the financial-statements date and the fieldwork-close date; and some auditors might feel that the occurrence or non-occurrence of an event during that time is partial evidence that might help one judge whether the event was or wasn’t “probable” as at the financial-statements date.

There are some right ways for a retirement plan (IRA or § 401 plan) to buy original-issue shares in a new corporation, or otherwise invest in a start-up business. The hard bit is doing so without triggering a prohibited transaction, taxable unrelated business income, or other nasty consequences. Moreover, Labor department interpretations (which also are authority concerning tax-Code prohibited-transaction rules) suggest that the Government might challenge transactions that, even if entirely proper in form, seem to involve some “arrangement or understanding” to evade exclusive-benefit, fiduciary-responsibility, or prohibited-transaction rules, or to do indirectly that which could not have been done directly. So one needs the advice, and usually order-of-transactions hand-holding, of a lawyer who understands this business idea and knows every inch of the prohibited-transactions rules along with the Labor and Treasury departments’ and courts’ interpretations of them. Sometimes, a service provider suggests that it has or will get an IRS determination that a plan using this investment method is “legitimate”. (Others present an IRS determination truthfully, but without explaining the limits on what it means.) An IRS determination that a document in form states a qualified plan doesn’t consider actual transactions and never opines on prohibited-transaction questions. To defend an implementation, a smart participant wants the support of his or her expert lawyer’s candid outlook, reputation, quality opinion letter, and malpractice coverage.

Getearl, you describe what many might consider a common-sense view – that a way, sometimes a good way, to construe or interpret a statute’s word that’s not specially defined by the statute is to consider the meaning or use of the word in a related context. And I empathize with a sense that many of the practitioners who use these bulletin boards prefer “tidy” answers, hoping that the law is clearly stated. But let’s recall that joel’s originating question asked: “Is there a possible scenario where a voluntary salary reduction only 403(b) arrangement can come under ERISA?”. Internal Revenue Code § 403(b) governs only whether contributions to a contract will or won’t be excluded from gross income for Federal income tax purposes. While § 403(b)(1)(A) restrains which kind of employer can pay over a § 403(b) contribution, this text doesn’t by itself preclude the possibility that a person other than the employer might “establish” or “maintain” a “plan or program” as non-tax ERISA uses those words. And the text of ERISA expressly recognizes the possibility that a plan can be “established” or “maintained” by a person other than an employer that contributes to the plan. Whatever one thinks about the merits of the possible interpretations and arguments (and I advocate none of them), we should be mindful of the possibility that some set of facts could result in an ERISA-governed plan, or at least that a court might so find.

Before we too quickly conclude that ERISA never governs any 403(b) arrangement for public-schools employees, we should allow for at least the possibility that there could be facts that might show that an employee organization (such as a labor union) "established" OR "maintained" a plan or program that provides retirement income to employees. A variety of arguments might be made concerning whether a plan or program maintained by an employee organization could be or can't be something different than a plan (if any) that's established or maintained by a government or by an employer of the employees. See ERISA 3(2)(A), 4(b)(1), 29 U.S.C. 1002(2)(A), 1003(b)(1); compare 29 C.F.R. 2510.3-2(d)&(f).

In helping a claimant consider his or her negotiation strength or weakness, be aware of the possibility that State law might impose special procedural requirements on a claim against a government entity. In some States, a claim against the State or its agency or political subdivision (which might include a public-schools district) can be submitted only to a special-jurisdiction court. In some States, this court sits only in the State's capital city. For claims against a government entity, often a shorter statute of limitations and statute of repose applies. Also, a failure to perform special service on the Attorney General or a similar official might destroy the claim.

The instructions for Schedule C seem to say that an administrator completing Part II should consider as the service provider not the natural person but rather the corporation, partnership, or other business organization through which an accountant or actuary practices. See page 33.

An administrator’s evaluation of a participant’s claim for a hardship distribution involves two prongs: (1a) Has the participant claimed an “immediate and heavy financial need” of a kind for which the plan provides a distribution? (1b) What’s the amount of that need? (2) Can the participant relieve his or her need by means other than taking the hardship distribution? The paragraph that Belgarath helpfully cites allows reliance on the participant’s statement “[f]or the purposes of paragraph (d)(3)(iv)(B)” – that is, the no-other-means prong. Some administrators interpret the rule’s express statement about relying on the participant’s statement for one specified purpose as suggesting that the Treasury department’s interpretation of IRC § 401(k)(2)(B)(i)(IV) doesn’t allow such reliance for other purposes. [see generally Bryan A. Garner, Black’s Law Dictionary (8th ed. 2004) at 1717 (“Expressio unius est exclusion alterius. The expression of one thing is the exclusion of another.”)] Other administrators ignore this hoary maxim, and interpret a plan to permit a distribution based on a participant’s claim without further evidence if the facts and circumstances make it reasonable to do so. One of those circumstances might be the administrator’s belief that a participant ought to be deterred from lying if the plan’s claim form tells a claimant that submitting a false claim is a Federal crime that can be punished by a fine and up to five years’ imprisonment. See 18 U.S.C. §§ 664, 1027. Without considering whether such a belief is reasonable, I’m aware of IRS examinations in which an employer has successfully defended a claim-form-only procedure. Moreover, even if an administrator believes that some participants would submit a false statement, it’s also easy to imagine that they could almost as easily fabricate ostensibly independent evidence. I’ve seen forgeries of death certificates (complete with accurate colors and raised gold seal) that to any but the most eagle-eyed reviewer look the same as the real thing. How hard would it be to fake a physician’s bill or a servicing bank’s eviction notice – especially when many genuine documents of this kind are merely laser-printed on cheap paper? ERISA § 404(a)(1)’s standard of care permits a fiduciary to resolve the tensions in the sometimes competing elements within that standard. An administrator could find that it’s prudent to design a claims procedure to leave out a step that wouldn’t screen out many false claims and that would find them only at unreasonable expense. Please understand that none of the views described above reflects the advice that I’d give a client.

Readers, please don't be offended that this post isn't a simple yes-or-no answer. The finding asked involves facts-and-circumstances interpretations, and a professional would want to do that after checking into all of the surrounding facts. The finding that the plan administrator needs is a "reasonable risk of liability for breach of a fiduciary duty[.]” What the reg looks for is whether the fiduciary's fear that it could be liable is a reasonable fear. While J Simmons' thoughtful follow-up inquiry asks us to assume a finding that there was no breach, real-world facts often aren't tidy and could be open to a range of interpretations. For example, a desire to exit an investment that hasn't yet matured sometimes is not investment-based and instead relates to poor administrative services provided by the investment's issuer or by a record-keeper that's somehow tied-in with the investment issuer. It often doesn't take much to find that the earlier selection of the poor-service provider might have been made with less diligence than a prudent-expert standard of care required. Further, it's possible to have a risk of liability for a fiduciary breach without having a fiduciary breach. As any of us who have defended litigation know, once a plaintiff alleges a claim, the defendant bears a risk that a judge or jury will decide the claim differently than what the defendant believes is correct. A lawyer's advice about whether a particular set of facts and circumstances involves a "reasonable risk of liability ..." might consider that a reasonable person could fear liability, not because the fiduciary breached its duty, but rather because there are enough unfortunate facts for a claimant to state a plausible-enough allegation. One also might look at what the facts suggest about potential motivations for the fiduciary's willingness to pay restoration. If 90% of the payment would be allocated to a business owner and people she has reasons (other than fear of liability) to favor, that might seem suspicious - at least in the absence of more persuasive evidence of a fiduciary breach. Conversely, a business owner's restoration payment that would be mostly allocated to a substantial group of non-highly-compensated employees might suggest that the business owner is willing to pay in the money because she recognizes sincerely that she acted as less than a prudent expert for the plan's investment and service-provider selections.

A “restorative payment” that a fiduciary pays into a plan not to make up a plan’s losses from market fluctuations but rather because there is a reasonable risk that the fiduciary otherwise would be liable for his, her, or its breach of a fiduciary duty owed to the plan is not an annual addition counted against the IRC § 415© limit. Treasury Reg. § 1.415©-1(b)(2)(ii)©. Likewise, if a plan administrator allocates the payment as full or partial restoration of investment loss (including opportunity loss), rather than as one would allocate a contribution, the restoration wouldn’t count in most non-discrimination tests. The plan fiduciaries must allocate such a restorative payment among all individual accounts that are affected by the breach, and must allocate it under an allocation method that fairly relates to each account’s losses attributable to the breach. Of course, it remains for factual and legal-interpretation analysis whether the provision that the “GIC” contract describes as a “market value adjustment” really involves something other than “losses due merely to market fluctuations” and also that “there is reasonable risk of liability for breach of a fiduciary duty[.]” For both issues, the plan administrator that wants comfort that a payment is or is not an annual addition could get some protection under both ERISA and the Internal Revenue Code if it reasonably relies in good faith on an expert lawyer’s written opinion.

Whether it’s wise to name the employer as its employee-benefit plan’s trustee is a separate question. But it’s not at all clear that a plan’s creator can’t establish a trust that names such a corporation as the trustee. Even if a State statute prohibits a non-natural person other than a bank or trust company from acting as a fiduciary, a common exception is for such an organization to serve as the trustee regarding a plan to provide benefits to the organization’s employees. Some States do this expressly in the banking statutes, others in the general-business-organizations statute, and yet others by courts’ interpretations. While I haven’t done a 50-States survey, I’ve rendered advice on enough States’ law concerning this point to believe that there’s usually a way to make it work. Like many practitioners, I usually suggest that it’s “safer” to choose a trust company (even if it’s one that has some business relationship concerning an investment or service provider). If an employer decides to “self-trustee” a plan, naming as trustee the corporation, rather than a particular natural person, sometimes can help keep documenting which natural persons make the trustee’s decisions internal to the corporation’s resolutions and minutes, rather than a matter that might involve documents exchanged with the plan’s other service providers. A natural person who is or might become involved in making the trustee’s decisions could get his or her lawyer’s advice about how ERISA defines a fiduciary and how courts have interpreted, and might further interpret, that definition. Such a person also could get advice about how to protect himself or herself using a combination of ongoing advice (much of which can be paid for from plan assets), liability insurance, and the employer’s additional defense and indemnity.

If the person whose consent is called for is “legally incompetent”, ERISA and the Internal Revenue Code do not preclude a plan from recognizing a guardian’s act as the spouse’s consent, even if the electing participant is the spouse’s guardian. 26 C.F.R. § 1.401(a)-20/Q&A-27. Under a U.S. government-organization plan, the Treasury department’s interpretation also is relevant concerning ERISA § 205. 43 Federal Register 47713 (August 10, 1978). Even if a plan is ERISA-governed, a guardian’s authority is grounded on State law. A guardian must act in the best interests of his or her ward. A guardian serves under a court’s supervision (at least indirectly), and must account for his or her actions. Further, some guardianship decisions require a court’s approval before the guardian implements the decision. Depending on the powers granted and other State law, the guardian might need to persuade a court that choosing against a survivor annuity is the ward’s best interest.

An important principle of the necessary-and-reasonable-services exemption is that the compensation of a fiduciary must be decided by another fiduciary that is independent of the service-performing fiduciary. With an orphan plan, it’s often hard to find any fiduciary, and even harder to find one that’s independent of the one who’s willing to do the wind-up work. That’s why a wind-up fiduciary who isn’t a qualified termination administrator (see below) or an uncompensated volunteer gets the Federal court to approve the appointed fiduciary’s services and compensation. When a wind-up fiduciary has paid himself without independent approval, the Secretary of Labor has sued for restoration of the prohibited transaction. Arguing that the plan really needed the work isn’t a defense to the lack of independent approval. The QTA rule can help only if an eligible bank, trust company, or insurance company that already has assets of the abandoned plan is willing to serve as the plan’s QTA. If such a company (not a natural person) volunteers, its fees for QTA service must be both “consistent with industry rates” and no more than the company’s fee for similar services for similar customers that are not orphan plans.

If the trustee you describe is an employee of the employer that maintains the plan, it's unlikely that the plan could pay for such a trustee's services. Even if an expense otherwise would qualify for the necessary-and-reasonable-services exemption [29 C.F.R. 2550.408b-2] with approval by an independent fiduciary that had no interest in the payee, another regulation interpreting that exemption declares unreasonable (and thus a PT) "compensation to a fiduciary who is already receiving full-time pay from an employer ... (any of whose employees are participants in the plan)[.]" 29 C.F.R. 2550.408c-2(b)(2). What the employer feels like paying for from its resources is the employer's question.

Also, be careful to compute (but see an exception below) contributions only on that portion of the minister's earned income that's attributable to his or her ministry, rather than to work that's not for the church. Usually, only an employee (but not an independent contractor) of a charitable organization may make (or have made for him or her) § 403(b) contributions. Many ministers are self-employed for most tax purposes. Special rules permit § 403(b) contributions for a minister, even if he or she is self-employed. A minister’s earned income from ministry (which might include services for the church or a church-related charity) is treated as his or her § 403(b) includible compensation. Instead of an exclusion, a self-employed minister may deduct his or her § 403(b) contributions. A special $10,000 limit sometimes helps a church employee who has little or no "cash" compensation. After 2001, a § 403(b) contribution of a church employee (see below) meets the annual-additions limit if the contribution is no more than $10,000. The total contributions using this rule (including a pre-EGTRRA version of it) for all years (including pre-2002 years) can’t exceed $40,000. Example: How a church’s housing allowance affects the annual-additions limit Frank is the pastor of Fellowship Church. The Church doesn’t provide Frank any salary or wages; rather, his only compensation is his obligation to live in the Church’s parsonage. If a parsonage allowance doesn’t count in § 415 compensation, Frank’s compensation is $0.00. If the usual annual-additions limit applied, Frank’s limit would be $0.00 [100% x $0.00 = $0.00]. Internal Revenue Code § 415©(7)(A) permits a contribution of up to $10,000 into Frank’s contract. But if Frank continues to have no compensation and contributions to his § 403(b) account are $10,000 each year, he may use this rule only for four years [$10,000 x 4 = $40,000]. This rule applies to an employee of a church, a convention or association of churches, or an organization that’s controlled by or associated with either a church or a convention or association of churches. 26 U.S.C. §§ 403(b)(1)(A)(iii), 404(a)(10), 414(e)(5), 415©(7)(A)-(E); 26 C.F.R. § 1.415©-1(d)(1)-(4) (not yet officially codified). See also Proposed Treasury Reg. § 1.403(b)-9(b)&©. Following the spirit of this bulletin board, please understand that this general information isn't tax advice to anyone.

Although it might not be the kind of obvious public authority that you seek, here's another way to look. It's common for a sponsor of a group trust to include in its trust declaration that the last of the kinds of permitted users is another group trust that has an IRS determination and restricts its use to the same kinds of retirement plans and other eligible users that could directly use the group trust. In my experience, such a provision doesn't cause any difficulty with getting a "clean" routine IRS determination that the group trust is a tax-exempt 81-100 trust. So one starting point might be to look at the text of each group trust's governing documents: does the "investor" group trust authorize investment in another group trust? does the "investee" group trust permit another group trust as a user?

The default-rollover idea depends on both the terminating plan and the receiving plan having the right provisions that support the rollover. First, a terminating plan doesn't allow a participant to receive a distribution, but instead requires the participant to receive a distribution. Even for a plan that no longer has any profit-sharing contributions or other benefit accruals, a qualified retirement plan isn't terminated until it pays the money (or delivers the property or rights) of its final distribution. Many employers' plan amendments to implement a plan termination provide that the final distribution is a single-sum distribution. However, I've seen (and written) provisions for a final distribution that set a series of payments long enough to meet an investment-related purpose but short enough that the distribution still is an eligible rollover distribution. The notice of such a final distribution gives a participant a choice between receiving the distribution as a payment of money or as a payment of a direct rollover to the eligible retirement plan specified by the participant. It's important to keep in mind the likelihood that something less than all of the terminating plan's amounts would move to the retirement plan that's named as the default. Some participants choose against a rollover, and some choose a rollover to a retirement plan other than the default. But I've seen situations in which 80% to 99% of the terminating plan's amounts became rollover contributions (not a merger or transfer) to the "suggested" plan. Of course, the receiving plan must exist before it can receive rollover contributions, and it's smart to make sure that the receiving plan properly can receive, and is willing to accept, the rollover contribution. If there is any significant doubt, the employer writes or amends the documents of the receiving plan at the same time that it writes the plan amendment of the terminating plan. The employer can write the terminating plan's documents and notices so that the plan administrastor had no choice (and thereby little fiduciary responsibility) concerning which retirement plan receives a default rollover of a participant who had not communicated his or her instructions. Nonetheless, an employer likely uses this default-rollover idea because the employer believes that the "suggested" retirement plan is good for a typical non-instructing participant.

Although these are not the only choices, the practical options are: transfer the 457(b) plan's assets and liabilities to another 457(b) plan that is maintained by an eligible State or local government employer "within the same State", or amend the 457(b) plan to conform to then-current IRC 457(b) at the termination time, and provide for and pay or deliver a prompt final distribution to every participant, beneficiary, and alternate payee. See 26 C.F.R. 1.457-10(a)(2). State and local law might impose further conditions.

Concerning your #2, a plan may provide - or, to the extent not precluded by the plan or applicable law, a plan administrator's procedure may provide - a "default" on whether an eligible rollover distribution (which might include a terminating plan's final distribution) will be paid to the distributee or paid to an eligible retirement plan (if not subject to an IRC 401(a)(31)(B) mandatory $1,000-to-$5,000 rollover to an IRA). See 26 C.F.R. 1.401(a)(31)-1/Q&A-7. If such a default is payment to an eligible retirement plan, the plan administrator's IRC 402(f) notice and an explanation of the default provision or procedure must identify the receiving plan, which might be a 403(b) plan. Although this isn't a merger, it can have some practical effects that achieve an employer's goal of maintaining one plan going forward.

Whichever a plan provides - resuming a 401(k) election once the required sitting-out period has expired, or treating a claim for a hardship distribution as including an election for "cash" compensation until the participant "affirmatively" and expressly elects to restart 401(k) contribution, it would be smart (and might be required) to explain this provision in the summary plan description. Further, a plan administrator might have a better answer to "why didn't you tell me" if an explanation of this provision is on the form by which a participant claims a hardship distribution.

Beyond the church plan administrator’s consideration of (1) the plan documents, (2) relevant State law, and (3) smart fiduciary communication, the church authority that has power to amend the plan might consider whether the internal law of the church affects a pension plan amendment or notice of it. For example, a church might have rules about decreasing a minister’s compensation. And even for an amendment that’s fully permitted, there might be rules about how to adopt, implement, and communicate an amendment. Some of the internal law of a church might be legally enforceable, and much might not be; but church people might seek to obey the rules made within their own authority.

Today’s follow-up question is whether insurance company or custodian of regulated investment company shares must be covered by fidelity-bond insurance. But a more immediate question is whether the employer and its people must be so covered. ERISA § 412(a) requires not only “[e]very fiduciary” but also “every person who [or that] handles [money] or other property of” an ERISA-governed plan to be bonded. (There is an exception for an insurance or trust company, but this usually doesn’t excuse the employer and its people.) Further, ERISA § 412(b) makes it a crime to serve as a plan fiduciary or handle plan assets while not bonded as required. “Handling” plan assets can include an opportunity to pay or delay contributions to the plan; a power to decide or instruct whether a participant’s or other person’s claim for a benefit is approved or denied; a power to end the service of a trustee, insurer, or custodian; a power to direct a trustee, insurer, or custodian; or almost anything that can cause a loss from dishonesty. See 29 C.F.R. §§ 2580.412-6 and -14. Beyond what might be required, why wouldn’t an employer WANT to get the fidelity insurance? Imagine a theft. Imagine it’s a theft that the insurance contract, if the plan held it, would respond to. Now imagine the participants’ claim. You were a fiduciary. Even without expert testimony about what would be “the care, skill, prudence, and diligence” of an expert retirement plan fiduciary, it must be a per-se breach to fail to do that which the statute expressly commands. Had you caused the plan to get the fidelity insurance, the plan would have been covered. Therefore, you are personally liable to make good the plan’s loss. While there’s a logic gap in that plaintiff-style argument, I’d hate to be a defendant who needs to hope that the judge is a strict constructionist. If a liability is based on an ERISA violation, a Federal court order or a Labor department settlement agreement can require an offset of a breaching fiduciary’s plan account to help restore the plan’s uninsured losses. A plan fiduciary need not pay for the insurance personally (other than his or her share along with other participants). Just pay (or reimburse) the insurance premium from plan assets. It ought to be a “reasonable expense[] of administering the plan” to do a specific thing that the statute expressly commands.

One method that some employers might consider is delivering the required notices and the summary plan description as enclosures that accompany an employment offer letter (if the employer regularly uses such a letter). With some new employees, doing so might increase how much notice time elapses before the first pay date and work period that might be affected by an "automatic" implied election.

No advice, but a few random thoughts: Before (not instead of) pursuing further corrections, shouldn't the plan fiduciaries start by getting the plan's money back? One would like to think that the drawer's bank has no right to debit its customer's account unless the bank paid the check as the drawer specified. After restoring the plan's bank account, the mistakenly-paying bank can recover from the mistakenly-collecting bank. Then, that bank could reverse the credit to the account of its customer - the person who presented a payment that wasn't hers. All this should happen after the plan's bank account has been corrected. What the banks do about their errors should be their problems. Once the plan is whole, the plan administrator might revisit some of its earlier decisions, and might decide not to pay an alternate payee any sooner than the time provided by a court order that is a qualified order.

Beyond ERISA's requirement to keep the indicia of ownership of all plan assets in the jurisdiction of the district courts of the United States, a trust that isn't a United States person trust could tax-disqualify a plan that otherwise would qualify under IRC 401(a). Even if all other conditions for a U.S.-person trust are met, a trust won't qualify unless "one or more United States persons have the authority to control all substantial decisions of the trust." Consider the difference between directed and decision-making trustees. Also, a non-U.S. person's power to remove a trustee or add another trustee can, depending on the surrounding facts, mean that the U.S. persons lack "the authority to control all substantial decisions of the trust." While it isn't necessarily unlawful (if the ERISA requirement is met) for a retirement plan to have a foreign trust, few employers and participants desire a trust that isn't tax-qualified.

An aside: If a concern is the expense of ADP or other non-discrimination testing, perhaps an employer might inform the plan fiduciary that the employer never had obliged itself to pay for that testing, and that the employer won't pay for it. After receiving this information, the plan fiduciary might find that the plan's documents don't preclude the plan from paying expenses of administering the plan. Following this, the plan fiduciary might find that ADP and related non-discrimination testing is necessary to the correct administration of the plan according to its terms, and meets other conditions as a proper plan-administration expense. Then, the plan fiduciary would (if the plan's documents don't provide the allocation) decide the allocation of that expense among participants', beneficiaries', and alternate payees' accounts.