MGB

Note that the IRS changed the address this year for Form 5558 from Atlanta to Ogden, Utah. So, if you are looking at old instructions, it will not have the correct address.

Tom mentioned the grandfathering. There was a short period in the '80s whereby a public employer could set one up. Then they took that away again, but those that had set it up could keep them. That would have included all state, local and instrumentalities (e.g., port authorities and other transportation quasi-governmental organizations). I assume that is when California put one in. Just last week I was reading a letter ruling that allowed a new one, but that had to do with very specific circumstances of the parent municipality having one and letting an instrumentality of them set up a new one (I don't recall the exact situation). As for NYC, do you have more information? Are you sure it is a 401(k) and is new for that unit? In general, they cannot set up new ones.

If you merge a MP into a PS; or if you change the MP into a PS (no former PS), do you always have to do 100% vesting? Is there any way to do this with retention of the MP vesting schedule?

When this idea first surfaced about 15 years ago, I was involved with a number of analyses of different insurance products being sold like this. None of them were worth it. However, certain situations would make it worth it, but they would not be able to buy the insurance due to underwriting. Remember, in order for the insurance to be sold, there is overhead (sales expense, administration, profits, commissions). None of these are involved in the qualified plan payments. Also, most J&S factors are subsidized within the plan in relation to the cost of insurance producing the same amounts.

The law is very clear that ANY restriction will trigger the catch-up contributions. So, in your example, as long as the person puts in 15%, then they can also put in a catch-up on top of that. However, the plan must allow it. Once a plan allows it, ALL plans of the controlled group must also allow it. That is ALL plans, not just 401(k)'s. This seems a little absurd, because it includes defined benefit plans. Discussions with Treasury indicate that they will give guidance that it should only apply to plans that can have elective deferrals (note that the catch-up contributions are not elective deferrals under 402, they are a new 414 section). Here's a hypothetical: A plan only allows 2% elective deferrals. Anyone that puts in 2% and is age 50 or over is now eligible for catch ups. Note that other restrictions can also apply, such as ADP testing or 415 limits. Hitting either one of these will also trigger the eligibility for catch ups. Some of the messier issues are with regard to new participants or participants with two or more jobs. If the person hits the 402(g) limit in aggregate, they are eligible for the catch ups. However, any one of the plans wouldn't know that. There will have to be some type of verification/substantiation from the participant that they are eligible.

Pax's reference is to my earlier explanation on the subject. I'd like to add one more thing: The "10 years" that is used to score a reconciliation bill is not the whole story. Actually, changes can only be within the 10 years, it does not have to stretch out to the full 10 years. If they had used the full 10 years, the sunset would have been in 2012, not 2011. Why did they sunset it a year earlier? The numbers were too big. In 2011, the full effects of the estate tax repeal will kick in (along with a lot of other items that are phased in). The total tax reduction would have been closer to $2 trillion over 10 years instead of the reported $1.35 trillion. That would have even been bigger than what Bush was asking for at $1.6 trillion. So, to keep the number small enough to sound good to the press, the sunset was moved forward one year. That means that the current 10-year scoring (a calculation of the change in revenue) includes a huge amount of tax revenue in the 10th year when all provisions revert...offsetting the first 9 years tax reduction which is actually larger than $1.35 trillion. There is absolutely no doubt in my mind many reductions/provisions will be rescinded before 2011 or else we will be in bigger deficits than what we had in the past. That could include some pension provisions (although they were peanuts compared to other provisions).

Janie, The normal form of payment must be a life annuity of some type. Most plans use a single life annuity. This means the defined monthly amount will be paid until the participant dies. No payments are made after that. If the plan allows an optional lump sum, the lump sum must be calculated in a particular way. The theory is that it will be equivalent to the payment stream in the life annuity. If you were to only use interest to discount each future payment, there would be no way to incorporate the fact that the payments will not go on forever. Although it is mathematically possible to have a lump sum equal to an annuity that never ends (a perpetuity), we know that people don't live forever. The probability that a payment will be made when they are 100 yrs old, given that they are 65 or less now, is very small. The probability that they will still be alive to be paid next year or the year after is very high. These probabilities are based on the mortality table and that is how the mortality table enters into the calculation. A lump sum is the sum of present value of each individual possible payment out to the end of the mortality table. Each individual present value is the probability that the person will live until the payment is made times the present value of the payment (how much would need to be invested right now to produce that single payment in the future) based on a compounded discount using the stated interest rate. A way to get an approximation to a lump sum is to know the life expectancy for the person under the mortality table being used. For example, assume the life expectancy of the person getting the lump sum at retirement is 20 years. If you assumed that the payments would last for 20 years and discounted these payments with interest only, you would get a very close approximation of the more complicated correct present value calculation described above. Although you are only using interest in this calculation, the mortality table was used to get the life expectancy. The above assumes the lump sum is equivalent to an immediate payment of the annuity. Most lump sums are deferred annuities. I.e., the person is terminating before the normal retirement age and the annuity is assumed to not start until then. In this case, the calculation as of the normal retirement age (in the future) is the same as above. To discount from there to the actual age of payment, interest is again required. However, in this part of the calculation, a plan may or may not use mortality. Generally, if there is no death benefit prior to the assumed starting date of the annuity, mortality is again used in the discount. Plans with large death benefits may discount with interest only. This will create a larger lump sum to incorporate the value of the death benefit being given up when the lump sum is taken. Another way to view the lump sum is from a grouped approach. Assume a 1000 people, the same age, took the same lump sum and pooled their money into a fund that received a fixed rate of interest that equals the interest rate used in calculating the lump sum. Have the fund pay out the annuities to each person. Some will only receive benefits for a short time, others for a very long time. The expectation is that the money in the fund will run out just as the last person finally dies.

Well, there is the elusive maximum exclusion allowance. Yes, it seems to be gone starting in 2002. However, there is a sunset date on this "repeal". So, in 2011, it officially comes back. Therefore, data will need to be retained in the meantime until the law is changed to do away with the sunset. That is not from my wandering thoughts...that comes from discussions with Treasury and may be in the guidance they will be issuing later this summer.

Actually, for some people (within 3 yrs of NRA) it is a $34,000 opportunity, not just $22,000. And, for others (age 50 or over), it is a $24,000 (or more) opportunity, not just $22,000. And it will rise fast over the next few years, because each individual piece will increase $1,000 per year. In the three years prior to retirement, the 457 limit is double the otherwise applicable limit (previously this was $15,000). So, you can do $22,000 without worrying about coordinating the $11,000 in another plan. The law says you cannot have a catch-up contribution (for those 50 and over) in the 457 in the same year as the double-limit. However, it appears you CAN have it in another plan in the same year. So, there is an additional $1,000 that can go to another plan. Once we get to 2006, this will be $15,000 in 401(k) or 403(B), plus $5,000 catch-up, plus $30,000 in 457 for someone within 3 years of NRA. Note that the catch-up can be in any plan. The law does not restrict the number of catch-ups you can have (although regs. probably will, or a technical correction). So, for those that are not within 3 yrs of NRA, they can have (in 2002), $11,000+$1,000 in 457, $11,000 in 403(B) or 401(k) plus $1,000 (or $1,000 in each if the plan sponsor has both).

This was completely intended...very specific, not a glitch As far as fairness...whoever has the loudest squeaky wheel gets the grease. Senators and Representatives listen to their constituents in governments back home more than anyone else. They wanted it...they got it.

Although the intent of the drafters was to have it apply to limitation years ending in 2002, the language came out ambiguous. The IRS has indicated unofficially in meetings over the past couple of weeks that they intend to clarify that it will apply to limitation years beginning in 2002.

Sorry, I interpreted your "defrosting" as meaning you no longer were doing FIL. With staying on FIL, my same comments as before: If you recognize the change in assumptions first, you cannot have a negative amount because that would be recognizing a negative UAL, which is not allowed. Thus, you do not have a base for the change in assumptions. The amendment base needs to be adjusted for the credit balance. Again, you could alternatively do the amendment first, and then the assumption change. But, the AL associated with the amendment will probably be different because it is based on old assumptions. I have a different problem with this. If you had zero UAL before with a credit balance, then you shouldn't have been using FIL. This is one of the situations in Rev.Proc. 2000-40 that states you must revert to another method because the FIL is not acceptable at that point. That is what I thought you were referring to when you said "defrosting".

The one year at a time approach comes from the proposed regulations in 1988 implementing the OBRA'86 changes to require continued accrual. 4/11/88, 1.411(B)-2 Accruals and allocations after a specified age. The only reason there are two schools of thought is that there are people that stubbornly ignore proposed regulations that have not been finalized. There is very good theoretical justification for the year-by-year approach whether or not the regulations say it. The point is that you must preserve the value of the accrued benefit that a person could receive if they retired. This does not mean "retired at NRA", it is retired at any time. That is what is driving the year-by-year approach.

There was actually a third base (done first) for a change in funding methods. By the numbers you stated, I assume at that point you were in negative UAL territory, so no base is established. The change in assumptions is next. It appears that you would still have been in negative UAL territory, so no base should have been recognized. The third change (plan amendment) brought you out of negative territory and it should only recognize the amount of UAL that puts you into balance, i.e., adjusted for the credit balance. That only leaves you with one positive base. Alternatively, steps two and three could be reversed. I.e., recognizing the plan amendment before recognizing the assumption change. At the time of recognizing the plan amendment, you would only recognize the amount that puts you into balance, which is going to be less than the actual amount of the AL of the plan amendment because some of it gets eaten up as you come out of negative territory. Then the assumption change is an actual change in AL difference due to assumptions, given the new plan provisions. I assume this would be a very different number than what you are now using. I do not like the second approach, above. From a theoretical purist point of view, I think the ordering should be: G/L Method change Assumption change Plan amendment In each step, the amount of change recognized has to limit the prior step to zero (adjusted for the credit balance) if the net at that point is a negative UAL. However, there is no actual rule from the IRS on the ordering, so other ordering is allowable.

There are no federal guidelines. Federal laws are generally not applicable to state or local governments (separation of powers issues). There may be state laws concerning this, but it is doubtful. Most states do not have many laws concerning benefits at the local level. In designing the provision, the county probably adhered to any state laws. However, it is still a possibility they didn't. Finding out the state laws in your state is not an easy task. A local lawyer familiar with state laws would be your best bet.

The regs have only been amended in Dec. 81; just minor changes in language (primarily multiemployer-oriented). Yes, a notice would have been required in 1981 for someone over the NRA at that time. You could have suspended previous to the effective date without a notice. If a notice was not given in 1981, you could have a suspension up until 1981, and then would be required to give an actuarial increase after that date.

ERISA section 203(a)(3)(B) (original passage in 1974, not a later amendment) provides for the ability to suspend, with a direction to DOL to issue regulations to implement the rule. December 19, 1978, DOL issues proposed regulation which includes the requirement to provide a notice in order to suspend. Final regulation issued January 27, 1981, 2530.203-3 "Suspension of pension benefits upon reemployment of retirees." Note that even though the title says "reemployment", this includes continued employment. There were also minor amendments on December 1, 1981, following a written comment period. Preamble to the final regs: "Plans which provide for suspension of benefits will be required to comply with all relevant aspects of the regulation. To the extent that this regulation imposes specific requirements not provided for in the Act, it will have only a prospective effect on the operation of plans and the rights of employees. Suspension of benefit payments by plans prior to the effective date of the regulation will be governed by section 203(a)(3)(B) of the Act without reference to the regulation." Effective date: May 27, 1981. So, in the period when one hour of service is worked after the effective date, a notice had to be given to suspend. You could have suspended prior to that without a notice because the notice was only in the regulation, not the Act. Pax's mentioning of 1988 is a different, related subject. Under ERISA, you could stop benefit accruals at NRA. All suspension meant was that you were forfeiting the receipt of benefits that you could otherwise have received if you terminated (suspension of benefits is an exception to the vesting requirements). Just stopping the payment is not suspension...you have to lose value to have a suspension. If you did not provide the suspension notice, you need to preserve the value of the benefits they could have received, i.e., give an actuarial increase. Under OBRA'86, amendments to the Code, the Act and ADEA (passed within a month or two of TRA'86), plans were mandated to give accruals after NRA, whether or not you are suspending benefits. However, if you are not suspending benefits, you can offset the new accrual and the actuarial increase, effectively giving the larger of the two. The IRS issued proposed regulations under the new OBRA'86 rules on April 11, 1988. They have never been finalized. Note that nothing in OBRA'86 (or the '88 proposed regs.) changed the suspension of benefits notice...it only changed the accrual requirement.

Yes.

This section of EGTRRA is targeted at mergers, not ongoing plan situations. If you read the conference report, they give examples of situations that the regulations are to address. This is not meant to allow the removal of subsidies in a plan...it is to remove one of two alternate subsidies that are nearly equal to each other and create administrative complications in trying to administer both. The de minimus issue comes in during the comparison of the two types of subsidy. A single subsidy is not a de minimus difference from another by definition (it has nothing to compare to).

Joe, I'd like to know which analyses from accounting firms, consulting firms and law firms you HAVE been reading. I have 27 of them (7 consulting firms, 4 accounting firms, 12 law firms and 4 professional associations) that clearly state this in their analysis, including our own (I know, because I write ours). Also having been involved in the crafting of the legislation, I KNOW this was the intent and there is no ambiguity whatsoever.

No need to get opinions. The law is very clear that the new schedule only applies to new matching "contributions for plan years beginning after December 31, 2001." (With a delayed effective date for collectively bargained plans.)

The intent was for the $40,000 to apply to limitation years ending in 2002. However, the language is not really clear on that. It will now fall in the hands of the IRS making a decision. The confusion is that the DB limit specifically references limitation years, whereas the DC does not. The reason the DB limit was done that way was because it originally was set up as a phase-in over multiple years. The words limitation year were added in conference to clarify that phase-in application. Once the numbers changed to be a one-time increase, the word "year" would have been sufficient, in my opinion (which is what the DC limit was worded as). The single "years" reference (with no mention of limitation years) is similar to earlier law changes in 415 limits. Note that "limitation years" is not a concept in the law, it is only in regulations. The regulations say you can use the limitation in effect in the "year" that the limitation year ends. It is the year in quotes that the law refers to in its effective date. An argument could be made that this is no different than the CPI adjustments being effective in a "year," but may be used in the limitation year ending in that "year."

The catch up is available to anyone that reaches a limitation in that one year (even a plan limitation like 10% of pay, or an ADP failure). Nothing from the past affects this. The reference to "catch up" was because this was introduced as a "wife" provision. Presumably, women are more likely to have been out of the workforce for child upbringing. Although that is what the provision was targeted at, they will not be the ones that actually use it. In the end, it is just higher limits as you get older. If you were basing it on prior hitting of limits, there would be no way to integrate the information from before the person was an employee. Note that the French have had this system for years. The amount of deductible contributions is based on your age.

Moe, I do not think he was referring to a specific payment from the church for housing. Practitioners have interpreted 107 to apply the housing allowance exception to retirement plan distributions (that come from a church plan) that are used for housing. I presume that is what the question is referring to. I would think that as long as the person is receiving distributions from the church plan, it makes no difference what his current status is as far as being disassociated or not...just opinion, no references. Is he still considered a priest for any other reason under the tax code? I.e., can he still avoid FICA if he had earned income?

Ray, You must pay FICA on the present value of the benefits at the earlier of the date of accrual or the date of nonforfeitability. You may not delay the FICA until the actual payment unless there is still a substantial risk of forfeiture. Even if there is, you are allowed to pay FICA on the PV. If there is any uncertainty in the amounts to be paid (e.g., it is adjusted each year based on an underlying qualified plan's payments' increasing due to 415 limits), then you must reconcile the payments with the PV and possibly pay more FICA in the future when the payments are made. However, you cannot get a refund if the payments turn out to be less. Why would you want to pay up front? Easy...the wagebase limitation. In this example, the FICA (both er and ee) would be 15.3% on the $60,000 payments each year. However, if you pay on the PV, you probably only need to pay 2.9% because the person is already over the wagebase (assuming we are not at the beginning of the year). You always want to have the termination late enough in the year to avoid the 12.4% up to the wagebase. However, even if you are at the beginning of the year, you only pay 12.4% up to the wagebase, not the entire amount (about $500,000). Keith, The rules are in Regulation 31.3121(v)(2)-1 (quite lengthy with numerous examples). Note that if there is a unilateral provision of the compensation AFTER termination of employment, it is not considered compensation subject to FICA. FICA only applies if the agreement (which could be unilateral) is reached prior to the termination of employment (see section (B)(4)(vi) 'Benefits established after temination'). Section ©(2)(ii) 'Present value defined': "...using actuarial assumptions and methods that are reasonable as of that date." So, no, there is no guidance...do what is reasonable and could stand up to scrutiny when audited. Note that they do give numerous examples, such as using a 7% rate and GAM83 male for a male calculation in 2001, but that isn't too usable because the regulation was finalized in 1999.