Freezing average salary

[ishi]

Consider a final average pay, defined benefit plan. Assume the benefit formula is 1% of 5-year average compensation, times years of credited service. A client wishes to freeze the average compensation portion of the formula (say as of March 31, 2015), but still allow participants to earn credited service. Is this permissible? The future formula would then be 1% of 5-year average compensation as of March 31, 2015, times years of credited service at subsequent termination of employment. Are there any other considerations? Thank you in advance for your help.

[John Feldt ERPA CPC QPA]

They'll need to provide an advance 204(h) notice with an explanation of the magnittude of reduction. 15 days advance notice for under 100 participants, 45 days for 100 and up.

I think this type of accrual might cause the plan to change from its nice old "safe harbor - no testing needed" accrual formula and cause the accruals to be general-tested under 401(a)(4). This is because the averaging period now used will no longer end on the date of determination for the accrual of benefits.

How will benefits be provided to new entrants? What will be used for their average compensation?

[ishi]

Thanks John. In my particular case, the plan has been closed for over 5 years, so no new entrants, and all actives have at least 5 years of salary for the average.

I do not have much experience with 401(a)(4) testing (as most of my plans are safe harbor), but you bring up a good point. If the plan does need to be 401(a)(4) tested, would this design necessarily cause problems, or do the potential problems depend on the plan demographics?

The client is foreign owned, and this type of design is permissible for the parent, hence the question. It is attractive to the client from an ASC expense perspective, since the PBO effectively becomes the ABO upon change, reducing the Interest Cost and loss amortization.

I know a prospective career-average formula would provide similar results, but wanted to see if there was an immediate "NO" about freezing the average salary while letting the service continue to grow.

[John Feldt ERPA CPC QPA]

It depends on the demographics. As a larger plan, I assume, it may not have problems passing, but you'll have to run the test to know for certain. Closed for 5 years - probably still okay for coverage and 401(a)(26). If this is a smaller plan, those issues would probably have already affected the plan.

[AndyH]

If the HCEs are at the 401(17) limit, assuming a low inflation rate their benefits to average comp ratio might not change quickly since their pay increase will effectively be at the inflation rate.

If NHCE pay increases exceed CPI, the ratio of NHCE benefts to comp could quickly become lower than the ratio of HCE benefits to comp.

Add to that the preclusion of new participants, and you could see testing difficulties in the near future.

[ishi]

Thanks John and Andy. After some research, this type of design is NOT permitted.

Regulation 1.401(a)(4)-3(e) defines rules relating to average compensation. Subsection (2)(i) further defines the allowable compensation history as "An employee's compensation history may begin at any time, but must be continuous, be no shorter than the averaging period, and end in the current plan year." (the averaging period is at least 3 years). The last part about ending in the current year is the death knell.

[John Feldt ERPA CPC QPA]

Under 401(a)(4), the compensation required for testing benefits is defined in 1.401(a)(4)-3(e), but the plan's definition can differ for purposes of accruing benefits.

[ishi]

OK - I'm retracting my pronouncement from post #6. As John points out, the 401(a)(4) section does refer to compensation for testing purposes, and need not apply to compensation used for accrual purposes.

Having spent more time researching, I am comfortable with freezing the average salary, but allowing benefit service to continue. The plan will need to be tested every year, comparing the actual accruals (using frozen salary) to the then current compensation, and hope for then best from a demographics perspective.

Thanks again.