Calculating final average earnings

[Guest Jensen]

I've got a plan that defines final average earnings based on the highest 5 consecutive calendar years within the last ten years of employment. The employee only worked partial years during the first and last years of the 5 consecutive year period. Can we pro-rate those years (e.g. salary earned/actual number of months worked x 12 = annual salary) or do we have to just include the actual annual salary for her two short years? The plan is silent as to this issue -- technically, the lack of any provision allowing for a pro-ration suggests that we must use whatever annual salary she actually had. But the intent of the plan in determining the "highest" average suggests that the intent is to give the participant the benefit of pro-rating her earnings rather than skewing her average by using less than full years of employment. Any thoughts?

[SoCalActuary]

Fairness to the participant is what you seek.

If your plan only counts full years of service for benefits, then you should not prorate the compensation,

because they get a full year of benefit on all past years.

If your plan counts months of service for benefits, then you should compute a monthly average pay based on actual months worked.

Try the math and you will see what I mean.

[david rigby]

"The plan is silent as to this issue -- technically, the lack of any provision allowing for a pro-ration suggests that we must use whatever annual salary she actually had. But the intent of the plan in determining the "highest" average suggests that the intent is to give the participant the benefit of pro-rating her earnings rather than skewing her average by using less than full years of employment."

I agree with the first part of the quote, but disagree with the comment about "intent". The answer should be to do what the plan says. IMHO, inferring a proration could violate the terms of the plan.

[Guest Jensen]

Thanks for your thoughts!

The plan does not count months of service, just "years of service" (which is defined as a consecutive calendar year with at least 1,000 hours worked.) Since the participant worked just slightly more than 1,000 hours in years 1 and 5, those years count as "service years" even though her salary was roughly 1/2 of what it was in years 2-4.

I think I agree with you, Pax, that we must follow the Plan language, even though the outcome does not seem fair to the participant. Another option I considered was just dropping years 1 and 5 and calculating the benefit based on only the 3 full years, but that seems to be even more of a deviation from the plan language, which calls for five consecutive calendar years. A final option was to include previous years of service (this participant was employed for about 3 years, left, returned, and has now left again.) We could come up with 5 full years of employment by using the 2 full years she was employed the first time around plus the 3 full years in her latest stint, but they would not be consecutive calendar years as required by the Plan language. Again, while I think that would give a more fair outcome to the participant, it is a deviation from the plan. So, I think we will just use the 5 consecutive years, including the two partial years to calculate the benefit. Anyone disagree with this approach?

[AndyH]

Yes. A safe harbor plan (I presume) can't base average comp on an averageing period greater than the period of time worked. It sounds like that is whay you are proposing.

Did the person in totality work more than 60 months? If not I would take total comp divided by total period of time worked. If yes I would take the last 60 months and compare that to the first 60 and give the greater.

If instead you are saying that the person was part time some years and full time others, then I would say tough luck-follow the document. IMHO.

[david rigby]

You care about the plan provisions.

If the plan defines service as years, using the 1000-hour rule, you don't care about months. You don't care about termination and rehire; you only care about the plan provisions.

BTW, you also don't care about perceived "fairness". In most cases, such subjective analysis is used to develop plan design, not to interpret the plan.

Did I mention that you care about the plan provisions?

[Guest Jensen]

The question gets messier!

We had been assuming that this participant had worked the minimum 1,000 in both years 1 and 5, but I told my client that they needed to confirm this. Turns out, she didn't quite make the 1,000 hours in 2002 (i.e. 2002 isn't a "year of service" so is not eligible for inclusion in the final average earnings). So now we have a situation where the participant has five years of service for vesting purposes (combining both the first stint and the second stint of employment), but does not have five consecutive years of employment, even if we include the partial year. The plan provides no alternative for such a situation. We have to calculate a benefit for her, and can't do so without a final average earnings (there are a couple of alternatives for calculating the benefit, but they both require the final avearge earnings figure). Without five consecutive calendar years of service, I can't calculate a final average earnings. I feel like I'm missing something here, but I've been through the plan forwards and backwards and can't come up with anything!

[WDIK]

So now we have a situation where the participant . . . does not have five consecutive years of employment, even if we include the partial year. The plan provides no alternative for such a situation.

I feel like I'm missing something here, but I've been through the plan forwards and backwards and can't come up with anything!

I would be very, very, very surprised if the document did not address the issue. If that is truly the case, QDROphile has a colorful description that may apply.

[AndyH]

LOL

Touche to our resident archivist!

BTW, what does pax think of "fairness"?

[Larry M]

If there is a concern about "fairness", and if the document is silent as to these anomalies (which are not unusual), why not discuss the option with and get the intent from the client and then amend the document to arrive at whatever result the client wants?

On the other hand, it is more fun to go back and forth among those of us wh o are not directly related to the situation.

[SoCalActuary]

So your fact pattern is this:

2002 less than 1000 hours

2003-2006 1,000 hours or more.

We find in our documents (see Accudraft) that a year with no service credit would be a year excluded from the compensation averaging. So you would have four years of service and four compensations to average.

What does your document say about years that don't count?

For what my opinion is worth, there is nothing unfair that the benefit was based on the 3+ years of actual compensation. The person worked 3 full and one partial year receiving compensation and providing services to the employer. But they received four full years of pension credit.

[Guest Jensen]

Well, I guess I'm stuck somewhere between being paranoid that I'm just missing something in the Plan and resigning myself to the fact that this plan sucks -- though "sucks" may be a little strong, perhaps it is just "annoyingly deficient"?!?

The plan does allow for disregarding a break in service of less than 12 months, but this employee was gone from 1999 until 2002 when she returned. So I have no problems with her being credited with the pre-1999 service, but I don't find any provision allowing us to just exclude the 3+ year absence, and treat 1999, 2003, 2004, 2005, 2006 as a "consecutive" period.

I know that we need to stick with the plan provisions (yes Pax, you did mention that!!!), but I don't see anyway to calculate the final average earnings in this case without deviating to some extent from the plan. Either I don't use five years or I don't use consecutive years. I'm trying to find anything else in the plan that would give some guidance on which way to deviate. The only thing I can find is a reference (under the provisions on limitations of benefits) to calculating the final average earnings using three consecutive years, which is consistent with Code sec. 401(l). Maybe this is apples and oranges, but I'm leaning towards sticking with the "consecutive" requirement, and just using the four consecutive years that we've got.

SoCalActuary -- you make a good point about "fairness" in the calculation. That was one of my client's big concerns -- though I had already advised them that we had to follow the plan, even if the result wasn't what they (or the employee) perceived as "fair", but that I would look at the document with an eye towards being as fair to the employee as allowed.

[TPApril]

This is along the same theme of calculating final average earnings based on five consecutive years where the plan document leaves a lot to be desired (or interpreted) for the first and last years of calculation. In this case, FAE is defined as the average of ee's Monthly Comp. Monthly Comp is defined as 1/12 of W-2 pay while participating in the plan. In this plan, participant can get partial year credit for months worked during the year.

The way the prior actuaries calculated FAE was to count actual months of first year (when under 5 yrs) and then no consistency for last year. If the total was less than 60 months, then sometimes they counted actual months in last year, sometimes they counted 12. If there were greater than 5 years worked, then they always counted 12 months for the final year (if comp was greater than 5 years earlier). Makes sense for those longer employees, but I have no clue which is the best approach for the shorter term employees. I'm dizzy and will apparently be correcting for about half of those who have less than five full years of comp. Any thoughts?