C. B. Zeller

If they want to be more restrictive, then passage of time is not going to accomplish that. They probably want to use a counting hours method, possibly with an equivalency if they don't/can't track actual hours.

Passage of time typically means that the participant becomes eligible after a period of time without regard to any service or breaks in service. It's similar to elapsed time, but without the service spanning rules - or rather, if the service spanning period were forever, instead of 12 months. Say an employee was hired on 11/6/2025, works for 2 months and quits on 1/6/2026. Then they show up again a year later and are re-hired on 3/1/2027. They are in the plan immediately on 3/1/2027, because more than 6 months have passed since their original date of hire. Contrast that to elapsed time, where they wouldn't get credit for their period of severance (because it was more than 12 months) and would have to work another 4 months after being re-hired in order to have earned a total of 6 months of elapsed time. It's probably a good idea to keep the 1000 hours failsafe in the document. While I can't think of a situation where it would override a 6 month passage-of-time requirement, maybe there are some class exclusions that it would be needed for. It also gives you some assurance, because it sounds like this passage-of-time provision is custom language added to the document. So, on the off chance that the IRS finds issue with it on audit, then at least you have the standard language to fall back on. As far as applying break-in-service rules, I don't think they would apply. But ultimately the interpretation of the plan document is up to the Plan Administrator, so they should abide by their best judgement, taking into account what has been communicated to participants about the rule, and probably with they lawyer's advice.

The 1-rollover-per-year rule only applies to distributions from IRAs, which are rolled over to another IRA. They can roll over as many distributions from plans as they like. They could also roll over multiple IRA distributions to plans without violating the rule.

By "AI consultant," do you mean a human who consults on AI matters, or a piece of software which is intended to function in place of a human consultant? I don't have any recommendations either way, just wanted to clarify the question.

Whether a plan is subject to mandatory auto enrollment is based on the date the CODA is established. Presumably the CODA was established when the profit sharing plan was established, so it would be after December 2022 and therefore not exempt.

The proposed regulations say that you count employees using the rules of 54.4980B-2 Q&A-5. Which says... If the half-time employees come out as 50% under this rule, then I would agree with your calculation and the plan would be exempt from automatic enrollment.

Assuming she's catch-up eligible, then yes. Yes. My understanding is that this is how the IRS applies the rule under audit. A participant's comp is not counted unless they benefit under the plan for the year, meaning that they actually receive some dollars, and deferrals don't count.

When talking about coverage testing, don't forget about the average benefits test. Depending on the size and demographics and contribution rates, this might be an easy win. Assuming that coverage is not going to pass separately, then they will have no choice but to change their plan designs. Obviously they could make both plans safe harbor, or make both plans ADP-tested. Or they could merge the plans into one, which would be either safe harbor or not. Other options might include fixing the coverage test by excluding some HCEs, or giving additional benefits to NHCEs.

The rule is that a participant has to receive at least some profit sharing in order for their comp to count towards the 25% limit. I don't know what source you are citing, but I imagine the author just meant that the definition of compensation used to apply the 25% limit includes deferrals, which is true. You can't just reclassify deferrals as catch-up because you want to, you have to exceed a limit. That's why I suggested giving some profit sharing in order to a) intentionally exceed the 415 limit, causing some deferrals to be reclassified, and b) make her compensation eligible to be included in the 25% limit. If she's under 50, and hasn't deferred the entire $23,000 yet, then she could also reduce her deferrals to make room for some profit sharing.

No. Elective deferrals are disregarded for purposes of the 25% limit - see IRC 404(n). Therefore a participant whose only contributions are elective deferrals is not considered to be benefiting for purposes of 404(a)(3). Therefore their compensation is not included in applying the limit. In other words, in order to count her comp towards the deductible limit, she would have to receive some profit sharing. Is the wife over age 50? If so, give her up to $7,500 (or more, if she's 60-63) in profit sharing. Her deferrals will be reclassified as catch-up, due to exceeding the 100% of comp 415 limit, and she will have some profit sharing so her comp will count towards the deductible limit.

1.410(b)-7(c)(3) is the reference. The relevant part states, with formatting added to make the point clear:

You would file the 2025 5500-EZ and check the retroactive adoption box. Of course an SB is still required to be provided to the plan sponsor, it just doesn't need to be attached or filed with the 5500-EZ.

Conversely, one might wonder what idiot came up with the idea of auditing something that you don't need to. Congress and the IRS have given plan administrators explicit permission to rely upon participant self-certification of hardships, unless they have actual knowledge to the contrary. So why would the plan administrator go looking for trouble when they had no duty to do so, nor liability for not doing so? There can be no good that comes from this, only problems.

To put some dates on the original question ... am I understanding it correctly, that you are talking about a plan adopted (say) 8/15/2025 with an initial effective plan year of 1/1/2024-12/31/2024? If that's the case, the first Form 5500 will be the 2025 form, due 7/31/2026 or 10/15/2026 on extension. You will check the box for a retroactively adopted plan on the 2025 form. You do not file a 2024 Form 5500 at all. However you will need to attach both the 2024 and 2025 Schedule SB to the 2025 filing.

The 204(h) notice applies to all DB plans, plus DC plans that are subject to IRC 412, such as money purchase plans (but not profit sharing plans). The notice has to be given 15 days in advance for small plans, but 45 days in advance for large plans. There are some exceptions in cases of mergers and acquisitions - the rules are in 54.4980F-1. https://www.ecfr.gov/current/title-26/chapter-I/subchapter-D/part-54/section-54.4980F-1

I think it's pretty clear from the reg section you quoted that UCE benefits are not included in the exception. The sole exception is for prohibited payments under 436(d) and 1.436-1(d) that are made to carry out the termination of the plan in accordance with applicable law. That is separate from UCE benefits under 436(b) and 1.436-1(b). Now, let me ask you - do you have an actual plan that is terminating, provides UCE benefits, has an AFTAP less than 60%, and has had an unpredictable contingent event occur after the plan's termination date? Or is this simply a question of academic curiosity? If it's a practical situation, and if the sponsor feels strongly about having the plan continue to provide UCE benefits after termination, then you might want to put it in the termination amendment and see what the IRS says when you submit the 5310.

Yes.

If the 402(g) excess was also a 401(a)(30) violation, then the plan would have to distribute the excess now in order to correct the qualification failure. If it was not a 401(a)(30) violation - for example, because the participant exceeded the 402(g) limit only when combining their deferrals in this plan with another plan of an unrelated employer - then that excess couldn't be distributed until the participant has a distributable event from this plan. In either case, the participant will experience double taxation on the excess; once because it exceeded the maximum they could deduct on their 2022 income taxes, and again when it is distributed.

This is an illegal exclusion unless the plan provides a 1000 hours fail-safe. In other words, the plan must provide that, regardless of the excluded class, a part-time employee who actually works 1000 hours enters the plan. Given that, the plan should satisfy the coverage test when using the option to disaggregate otherwise excludable employees.

Because the plan document specifies a new comp allocation, which is not a safe harbor formula, you have to satisfy the general test (aka the rate group test). However if you calculate your rate groups on allocation rates, instead of on accrual rates, then everyone should be in the same rate group and the test should pass easily. You don't need the gateway because you're not cross-testing. You also don't need the ABPT unless your rate group is less than 70% coverage (which is unlikely to happen, unless, as Lou mentioned, there is an allocation condition that isn't met by a large number of NHCEs).

The plan document will indicate whether the top heavy minimum is computed based on years of service or participation.

10% excise tax on the unpaid minimum, IRC 4971. There is a second tier tax that can increase to 100% if it remains unpaid. You must also notify the PBGC of the failure.

See this thread from yesterday:

Not to push you back towards that ledge, but I do think there is a 415(h) controlled group here. So you do have to aggregate, but only for 415 purposes.

The only "requirement" that I have seen is from the 5500 instructions. This is from the instructions to 5500-SF, there is similar (but more complex, addressing DFEs and such) wording on the full 5500.