fractional accrual rule

[david rigby]

Plan defines service by 1000 hour rule and accrued benefit by fractional rule; no special definitions.

Situation: EE hired at (say 25) works full- time for 6 years (total projected service is 40). Terminates employment with more than 500 but less than 1000 hours, is rehired two years later and again works 1000+ each year. when he finally terminates, is the denominator of his fraction still 40? Or is it reduced by the "missing" years in which he did not work (or may have worked a fraction of the year, getting between 500 and 1000 hours)? My point is that the definition of the fraction refers (numerator and denominator) to the defined (and capitalized) term "Years of Service".

We have a difference of opinion in our office on this. Some say that because the EE did not work 1000+ hours, it does not meet the definition of "Year of Service" and should be ignored. Others say that this does not make sense because it means that an accrued benefit can increase merely because an EE worked parttime, thus reducing the denominator by one each year. Taken to an extreme, this means that an EE who works several years full time and then changes to partime for the next 30 years will end up with a fraction of 1 at NRD.

Any comments? Thanks.

[mwyatt]

I've often thought of this situation. Generally, I've seen most plans interpret the denominator as service to date plus expected future service from date of calculation to retirement date (which would lean in interpretation that yes, you do get some bonus accrual for stepping out and then returning to work). However, if you take the position that the missed service still stays in the denominator, how do you reconcile the benefit if the participant stays until retirement? Consider a benefit of 50% of average reduced by YoS less than 25 at NRD with service accrual. A participant starts employment at age 30, works full time until 40, leaves for 10 years, returns to work at 50, and stays until 65 (NRA).

First, let's ignore the missing service in the denominator. At age 40 termination, his fraction is 10/35. At rehire at 50, this fraction balloons to 10/25 (good for him) but then continues increasing at traditional rate to 25/25 at age 65.

Second, let's count all theoretical service from initial date of hire to retirement. At age 40 termination, his fraction is 10/35. So far so good. He returns at age 50, fraction is still 10/35. At age 64 his fraction is 24/35. At age 65, however, the fraction balloons to 1.0 (remember the normal retirement benefit takes over here and he does have the requisite 25 YoS in the NRB formula.)

I know that either interpretation doesn't provide for equitable or orderly accrual. Any comments on this analysis? I guess that this is also a good reason for moving towards 133 1/3 accrual (i.e., recouch the benefit as 2% per YoS max 25).

[This message has been edited by mwyatt (edited 10-04-98).]

[This message has been edited by mwyatt (edited 10-04-98).]

[david rigby]

i think you have captured my situation. It seems to me ridiculous that someone's accrued benefit could change (up or down) because of partime employment, or a break in employment. I ceratinly would not want to explain to Mr. CEO that situation.

I would love to get other responses on this to see if I/we can reach a consensus.

[Guest Keith N]

I tend to agree w/ mwyatt. In general I would look at it as YOS to date in numerator and YOS + potential future service in demonimator. Your concern about the increase in accrual due to a decreasing denominator of the person working more than 500 but less than 1000 should be offset by the YOS required for full benefits. If the formula is 40% of comp reduced for less than 25 YOS, then eventually his projected benefit will decrease since at NRD he won't have the 25 YOS. And if he did have his 25 YOS at NRD, he should be entitled to the full benefit. Actually, now that I think about it, I believe at NRD his accrual fraction must be 1.0, thus this is the only acceptable method. Otherwise you could have a situation where a participant has attained NRA, completed the required service, but not be entitled to the full benefit since his accrual fraction is

[david rigby]

OK Keith, what is your conclusion? Did you just come down on both sides of the fence? Let me try a specific example:

Comp is 10,000. Benefit is 30% of comp, prorated for less than 25 years of service at NRD. The Accrued benefit is that, multiplied by the accrual fraction, which is actual service divided by projected total service.

If EE is hired at 35, works 5 years (so vesting is not an issue), the accrued benefit is: .30 x Comp x 5/30 = $500. The EE terminates and is later rehired at age 42. Thus the potential future service at rehire is 23. Does the fraction become 5/28, immediately changing the AB to $535? I contend that this is ridiculous, since it means the EE has received an increase in AB for NOT WORKING. To me, the AB cannot have changed. In fact, I believe that the denominator of the fraction must always be 30.

Comments?

[richard]

I ran into this situation a number of years ago, and it is very tricky. By the way, this approach can also apply to rehired employees (rhe period of time between initial termination of employment and date of rehire is analogous to the period of time working between 500 and 1000 hours per year), time an employee workes for a noncovered division, or time an employee worked for the employer as a member of a union.

The situation was resolved as follows:

An employee who has 25 or more POTENTIAL years of service at NRD will receive a 50% benefit multiplied by his "actual years of service" at NRD and divided by his "potential years of service" at NRD. Actual years of service were years in which the employee worked at least 1000 hours. Potential years of service were measured from his original hire date.

So, in MWYATT's example of an employee hired at age 30, terminated at age 40, rehired at age 50 and retired at age 65 would have an accrued benefit of 10/35 x 50% at age 40, 10/35 x 50% at age 50, 24/35 x 50% at age 64, and 25/35 x 50% at age 65. This reduction factor at age 65 is because he did not work for the company between ages 40 and 50 (or equivalently, if he had worked between 500 and 1000 hours between ages 40 and 50).

Stated another way, for an employee ORIGINALLY HIRED at age 30, his benefit is (50/35)% per year of service (where year of service is based on 1000 hours).

For employees (with at least 25 years of potential service at NRD) who terminate before NRD, their benefit is 50% multiplied by his "actual years of service" at termination of employment and divided by his "potential years of service" at NRD.

Finally, for employees with less than 25 years of potential service at NRD, the 50% described above is proportunately reduced.

The way this mess was written in the plan document was actually quite succinct.

Accrued Benefit = 50% x

[min(potential credited service at NRD, 25) / 25]

x min [actual credited service at termination / potential credited service at NRD , 1 ].

I'll not go into the details of how employee who worked past NRD were treated; it's ugly.

[david rigby]

I'm starting to see the other side. For example, what do you do with the following?

two 40 year old EEs with identical comp, each has 5 years of service. The only difference is that Employee A worked from age 30 to 35, and was then rehired at 40; Employee B was hired at 35 and has not severed employment.

I think we may have a discrimination problem if our plan treats these two EEs differently. In other words, that may be the primary reason that such plans often have minimum service requirements. Am I just being picky? Any comments?

[richard]

David, there is no more discrimination than any other fractional accrual situation.

Employee A (the 30 year old hire) is accruing his 50% projected pension at the rate of 1.43% (50/35) per year; Employee B (the 35 year old hire) is accuring his 50% projected pension at the rate of 1.667% (50/30) per year. After 5 "Years of Credited Service (at whatever ages these "Years of Credited Service"), Employee A has a 7.15% benefit (50%x5/35 or 5 x 1.43%); Employee B has a 8.33% benefit (50%x5/30 or 5 x 1.667%).

If the rehire issue concerns you, let's ignore rehires completely. Compare the age 40 benefit of Employee B (hired at age 35) to the age 35 benefit of Employee C (hired at age 30). Both have 5 years of service, but Employee B's accrued benefit is 8.33% (50%x5/30) and Employee C's accrued benefit is 7.15% (50%x 5/35). They are different, even with the same service.

It makes sense (I think), but it gets tricky.

[mwyatt]

I still think that you have a problem at normal retirement by counting the missed service in the accrual fraction (first off, the benefit at NRD is the "Normal Retirement Benefit", not the accrued benefit). How would you justify two employees at Normal Retirement having completed the required 25 YOS for the full benefit, but saying that one has a lesser benefit because of missed service (in fact he might have more service than the other employee). Let's say Employee A had 5 YOS, then left 5 years, came back, worked until 65 and completed an additional 25 YOS. Employee B started at age 40 and worked until 65, completing only 25 YOS and gets full benefit. I think you would have a real problem saying that Employee A at NRD would not get the full benefit.

[richard]

I would define the benefit at NRD (ignoring the 25 year minimum) to be 50% multiplied by years of credited service at NRD divided by potential years of credited service at NRD. I believe that takes care of the problem at NRD of two different employees hired at different ages but have the the same total service at NRD due to missing intervening years.

If this rationale isn't satisfactory, you can define the benefit formula using a unit credit approach rather than the fractional approach, with different unit credit accrual rates based on original hire age. For example, the benefit formula for employees hired at age 40 or later would be 2% X years of service; the benefit formula for employees hired at age 39 would be 1.92% X years of service (1.92% = 50% / 26), etc.

While the "standard" fractional accrual formula is cleaner to meet the IRS safe harbor tests, the "unit credit approach" is equivalent, so it should meet the safe harbor as well (of course, it might have to be explained to the IRS reviewer). Regardless, the 401(a)(4) general test is easy to pass anyway.

(By the way, what I don't like about defining the benefit using the unit credit approach is that it is explicitly clear to young hires that they are accruing their benefit more slowly than old hires; at least with the fractional rule, that is "hidden." Some HR folks don't mind hiding these things.)

[Guest Ken Newhouse]

Although you can probably do anything you want as long as pass 401(a)(4), Rev Rul 81-11 provides guidance on minimum accrual in this situation. It basically says you have two fractions for each period and then you combine them into one fraction with a specific formula. It would be way too complicated to explain to a participant, so our documents have been written to define the denominator as credited service at termination plus potential future years to NRD. This always give a fraction greater than the minimum required in 81-11.