My 2 cents

Not surprised to hear that it is a defined benefit plan. Don't otherwise-vested account balances under defined contributions always get paid in full when the participant dies? My vote is for age as of date of death unless the plan clearly provides otherwise. The beneficiary's status would be locked in as of the date the participant died, irrespective of the extent to which there would be reasonable administrative delay. If the eligibility decision is not as of the date of death, when else? Does the plan pay lifetime survivor benefits to someone who was a child under 21 when the participant died or are they only temporary until attainment of age 21? If only until age 21, then if the child had reached age 21 by the time the benefits were processed, then only back payments would be required for the period from the participant's death to the date the child attained age 21.

1. How does the plan define beneficiary or the death benefit payable when the participant is unmarried at the time of death? 2. What happens if the participant is not survived by either a spouse or children under the age of 21? 3. Whether the child is married or not could only matter if the plan contains explicit language covering that point. If drafted by an experienced attorney, if the plan refers to dependent children, it should explicitly state what exactly that would mean. 4. If it were a defined benefit plan, as often as not any survivor benefits payable to minor children would stop when the child reaches age 21. It is even possible that if the child marries before 21, then the payments would stop, but that would happen only if the plan said so.

Please make no mistake - cash balance plans are always defined benefit plans, no matter how much they are disguised to look like defined contribution plans. None of the defined contribution rules apply to them and all of the defined benefit rules apply (i.e., PBGC premiums, no lump sums without spousal consent, no last day rule, minimum contributions that may be greater or less than the "principal credit" - there is no necessary correlation between the assets held by the plan and the total of the account balances, etc.). And don't forget that an annual valuation by an enrolled actuary is always required! And if the enrolled actuary certifies an AFTAP below 80%, lump sum payments are subject to restriction under IRC Section 436.

$275,000 is too big, even for the limitation year beginning in 2015. The 401(a)(17) compensation limit for the 2014 limitation year was only $260,000. Nothing above that could have been recognized for plan accrual purposes. The compensation limit for 2015 was $265,000. How did the plan determine plan accruals? Final year earnings, final 5, highest 5 ever? To get to a lump sum of $2 million, the plan had to have been in existence for 10 years or so, right? For Section 415 compliance (10 years of participation?). Why, if the plan is not well enough funded to pay the benefit, did the participant decide to terminate the plan? And if the lump sum is $2 million and the assets $1 million, how can you possibly have a $300,000 limit on the deductible? Sorry, just not understanding how this is all possible.

I am not involved with defined contribution testing, but would it work this way: a. Person A takes the entire bonus as an addition to their 401(k) plan. 402(g) must be satisfied and ADP testing must be satisfied. b. Person B takes the entire bonus as cash. No 402(g) testing and Person B counts for all tests as someone who received no annual additions (at least with respect to the bonus). Wouldn't that tend to make it harder for the plan to pass testing, since the lower-paid employees are more likely to take cash and the higher-paid employees are more likely to opt for the money to be paid into their 401(k) accounts?

Alas! Trying to cover things up only makes them worse.

I don't know about you, but I can remember "412i" but not "412(e)(3)"!

I thought so. How do you make a "withdrawal" from an annuity? Wouldn't you just start to receive it (or, if permitted, cash the entire annuity out)? But this doesn't help you resolve your issue. I don't work with 412i plans and don't know if there are any special regulations governing them. In a funded defined benefit plan (the kind subject to sections 430 and 436), the plan is permitted to allow benefits to commence after attainment of normal retirement age or after attainment of age 62 while still in service.

If one were talking about a straight-forward defined benefit plan, how could one handle an RMD without requiring that the participant be treated as having retired, include having to choose one of the plan's benefit options (with spousal consent, to the extent necessary)? In a defined contribution situation, carving out a portion of the benefit could make sense but how could that work in a defined benefit situation? Even if something must be paid, the spousal consent rules must be satisfied. If one were trying to pay the bare minimum under the RMD rules (supposing that the distribution was, in fact, limited to the RMD by in-service distribution rules of the plan), how would you approach paying it out as a QJSA? That has to be the default method.

Would "withdrawal" and "his account" be meaningful in the context of a [fully-insured, defined benefit] 412i plan?

Am I missing something here? Why would the investment manager/broker involve his own assets when the liability for the error (if any) belongs to the bank? And how would executing some sort of transaction in an account unrelated to the account affected by the error "fix the loss"?

Perhaps I am misunderstanding the point being made in the first paragraph of the above post. When I looked at 1.401(a)(9)-5 Q&A 1, which concerns RMDs under defined contribution plans, it did appear to me to specifically state that under no circumstances can the RMD be greater than the entire account balance as of the date of the required distribution. Suppose that the life expectancy for a year was 10 years, and that the balance at the end of the prior year was $500,000. This year's RMD would then be $50,000, but if $475,000 was invested in Enron or Bernie Madoff's fund and the value as of this 12/31 was only $25,000, then per the regulation, a payment of $25,000 would suffice to meet the RMD for the year. After all, it would be impossible to pay out more than is there, wouldn't it? As the regulation says: "A-1. (a) General rule. If an employee's accrued benefit is in the form of an individual account under a defined contribution plan, the minimum amount required to be distributed for each distribution calendar year, as defined in paragraph (b) of this A-1, is equal to the quotient obtained by dividing the account (determined under A-3 of this section) by the applicable distribution period (determined under A-4 or A-5 of this section, whichever is applicable). However, the required minimum distribution amount will never exceed the entire account balance on the date of the distribution."

Not really seeing this as an ambiguity. The minimum distribution is an amount calculated under a specific formula, but to the extent that the current account balance is lower than the calculated amount, the RMD must be cut back accordingly. You cannot pay out more than what is there. If the payment is larger than the RMD, it may well be acceptable, but it is not all RMD in that case. For most defined benefit plans, the plan language would normally support allowing the entire accrued benefit to go into pay status, even if that means paying more than just the RMD. If the participant is being paid a lump sum, first one strips off the actual RMD (prior year and current year) and then what remains can be rolled into an IRA. Few, if any, defined benefit plans(other than small defined benefit plans) have language that limits the in-service distribution for someone who has reached the required beginning date to the minimum amount required as an RMD.

This is easy. Any modifications of the fully generational RP14/MP14 require (per auditors) you to prove that this is the best of the best. Somehow the fully generational RP14/MP14 is the best of the best without any proof needed. If the calculations are for purposes of meeting accounting requirements under either ASC-715 or ASC-960, isn't it the sponsor who has the primary responsibility for selecting the mortality assumption (presumably to be carried out in a way that is acceptable to the accountants)? That selection for such purposes is not the ultimate responsibility of the plan's actuary?

OK, I'll do it. The answer given by the IRS, verbatim, is "Yes".

As a suggested starting point for this inquiry, what does the plan say?

The majority of the defined benefit plans I have seen accommodate commencement of benefits in any of the plan's annuity options (or, if permitted under the plan, a lump sum option) upon attainment of the required beginning date as though the participant had retired. Normal spousal consent would be required, and normal withholding would apply as it would for non-RMD benefit commencements/payments. The plan would typically say that benefit payments would not be made until the participant had retired or, if earlier, the required beginning date, so, in effect, the plans seem to except RMD situations from restrictions on in-service distributions otherwise applicable. Default withholding would be 20% if paid as a lump sum (and don't forget to carve out the actual RMD from the part subject to rollover!) and whatever would be the default (unless they return the W-4P) for the annuity payments to be made if the benefit is to be paid as an annuity.

Is it not now standard practice, when any transaction occurs that does not come from the address on record, to send a confirming email or letter to the prior address? Is it even possible that sending a blank form out could result in fiduciary liability? It was the new company that accepted it, apparently without question. How could the old company have any liability for sending out a blank beneficiary form? The whole thing sounds like it might make for a good movie. Family member of the former beneficiary? New beneficiary's girlfriend? Forgery, handwriting experts? Has it been verified that the participant died of natural causes?

I agree that the use of the SOA tables, however much the accounting profession may prefer their use, makes no sense in your situation. If the national office comes back with a demand that you use the SOA tables for their 12/31/14 accounting, tell your client that you think that they are mistaken but that you are going to use whatever your client instructs you to use. You have spoken up and (if you are in fact putting yourself out to pasture) have little stomach for battling over such things, so if your client does not want to argue the point, just use what the accountants (with the backing of their national office) specify for the purpose. At worst, when the settlement is reflected in 2015, there will be a gain because the lump sums will be less than the liability measured as of 12/31/14. I cannot say to what use FAS35 (aka ASC960) values are put besides those for which they are intended, but I can say that if the DOL has issued any guidance concerning the use of the SOA tables, I have not heard about it. Does the DOL have jurisdiction over such things as ASC960, beyond a keen interest that they be done right?

If your software just bumps up the table each year by another year's projection of RP-2000 using Scale AA, there should be no difference between the static tables for 2016 and those shown in the IRS notice. Just another year of more of the same. The current rules allow use of Scale AA to produce a fully generational table. It is likely that if the IRS decides to go to the SOA RP-2014 basis effective in 2017, whether they will require fully generational application of the new mortality improvement scale or permit the use of static tables reflecting the new scale on a more limited basis, they are likely to use the new scale in some fashion. Time having gotten so short, it appears that the IRS has deferred any material changes to the mortality status quo to 2017. Whatever they do, material changes will require them to offer a comment period. Don't know about you, but some accountants are already requiring use of the SOA tables with generational projection, so at the very least, prepare your systems to handle them soon if they cannot already. Just don't have to have them in place for minimum funding or 417(e) purposes just yet. It would not surprise me at this point if the IRS decides to allow static tables each year and (for minimum funding but not 417(e)) permit but not require the use of fully generational tables. Static tables may be preferable for 417(e) purposes.

But don't forget - once they go over 120, they would need to get down below 100 to use the SF again! If between 80 and 120, get to continue using same form as year before. Unlikely that sponsors would really want to continue using 5500 if fell below 100 and otherwise qualified for the 5500-SF. But could if they wanted to if stayed over 80.

If it ever did have a possible effect on being chosen for an audit, it probably hasn't since the IRS changed it from an actual request requiring justification and IRS approval to the automatic "send it in by the deadline and you're all set" process it is today.

What possible disadvantages would there be?

Haven't quite hit the deadline yet!

But the plan, having sent out payment, would issue a 1099-R and then the participant would, as a separate action, have to claim a casualty loss offsetting the taxation as reported by the plan. Right? So the plan would issue a 1099-R in any event, forcing the participant to take some sort of action (either pay the taxes or pursue a casualty loss claim).