Slushy. Accruals continue, but no new entrants after [date]

[John Feldt ERPA CPC QPA]

A prospect with a 32 participant DB plan indicated to us that:

1. Their plan does not allow any more new entrants hired on or after [date].

2. The usual benefit accrual formula continues for all who entered the plan before [same date].

To me, it seems likely that this plan will eventually fail the ratio percent test when testing all employees who would normally have have entered without regard to #1 above (it might already fail, I have yet to see the data). When that happens, even if we run an average benefits test, how will such a plan continue to satisfy 410? I'm not sure that a 'hired after' date works as a reasonable business classification...

Can we just ignore all of the employees hired after [date] when we do the coverage tests? If so, where's the code/reg or other cite for that? If the benefit accruals were frozen, sure, but that's not the case here. Am I missing something?

[Effen]

No, it will eventually fail either 410(b) or 401(a)(26) or both.

[QDROphile]

Maybe no 410(b) failure if new hires are the right mix of HCEs and non HCEs.

[himt4]

Just thinking...hmmm... And you make sure the plan formula is based on years of participation.

And then you put fail safe language into the plan that says that if plan fails 410(b)/ 401(a)(26) then you add to the plan NHCEs from the "excluded class" (i.e. those hired after [date]) in the order of when they were hired until such point that you pass 410(b)/401(a)(26).

It looks ok at first look. You've defined an excluded class (i.e those hired after [date]) that does not seem discriminatory, and failsafe language that does not seem discriminatory . But when you think about it, you've essentially changed the statutory one year waiting period to an "as long as it takes until we need you" waiting period. Perhaps you send the plan to the IRS for a determination letter.

[Blinky the 3-eyed Fish]

Excluding people after x date is a reasonable classification. I don't have the cite handy but it's specific in the regs. You should continue to run the tests and once you reach failure, add back under an -11(g) amendment. If you have a safe harbor design in the DB plan it very well may work out for coverage testing that the ratio test and average benefit test yield the same results.

[John Feldt ERPA CPC QPA]

Blinky,

Here's all I was able to dig up from the regs:

1.410(b)-4(b) Reasonable classification established by the employer. --A classification is established by the employer in accordance with this paragraph (b) if and only if, based on all the facts and circumstances, the classification is reasonable and is established under objective business criteria that identify the category of employees who benefit under the plan. Reasonable classifications generally include specified job categories, nature of compensation (i.e., salaried or hourly), geographic location, and similar bona fide business criteria. An enumeration of employees by name or other specific criteria having substantially the same effect as an enumeration by name is not considered a reasonable classification.

and

1.410(b)-4(c )(3)(ii) Factual determination. --A classification satisfies this paragraph (c )(3)(ii) if and only if, based on all the relevant facts and circumstances, the Commissioner finds that the classification is nondiscriminatory. No one particular fact is determinative. Included among the facts and circumstances relevant in determining whether a classification is nondiscriminatory are the following --

(A) The underlying business reason for the classification. The greater the business reason for the classification, the more likely the classification is to be nondiscriminatory. Reducing the employer's cost of providing retirement benefits is not a relevant business reason.

(B) The percentage of the employer's employees benefiting under the plan. The higher the percentage, the more likely the classification is to be nondiscriminatory.

(C ) Whether the number of employees benefiting under the plan in each salary range is representative of the number of employees in each salary range of the employer's workforce. In general, the more representative the percentages of employees benefiting under the plan in each salary range, the more likely the classification is to be nondiscriminatory.

(D) The difference between the plan's ratio percentage and the employer's safe harbor percentage. The smaller the difference, the more likely the classification is to be nondiscriminatory.

Is the cite your suggesting from subparagraph (A) above? If so, I'm still not convinced that a hire date works.

[Blinky the 3-eyed Fish]

After looking up the cite a better description is that it doesn't violate the age and service conditions. As for a reasonable classification, I don't know how you argue it's not being that you are applying it uniformly and objectively.

See 1.410(a)-3(e)(1) example 6 to back up my first sentence.

But as my last sentence said in my first post, this all could be moot (with a safe harbor formula and unit accrual).

[AndyH]

I agree with the responses. 401(a)(26) will eventually kill the plan if nothing else. For 410(b), typically this plan might be aggregated with a new PS plan for coverage testing if need be, but that will not help 401(a)(26).

[John Feldt ERPA CPC QPA]

Thanks, your help is very much appreciated.