When is a compensation test required?

[Guest MES]

Compensation Test

If a plan uses a definition of compensation that is not a safe harbor definition, does it need to run a comp test (average percentage of total comp included for the HCEs as a group does not exceed the average percentage of total comp for NHCEs by more than a de minimis amount - 1.414(s)-1(d)) if the plan only has deferrals and matching contributions? For example, a plan excludes bonuses and overtime from compensation for purposes of deferrals and match, but when testing the plan a 414(s) definition of compensation is used. If the testing comp is safe harbor is the comp test necessary?

I know a test is necessary if a profit sharing contribution is to be allocated on the basis of that compensation definition, but I am unsure if it is required in the absence of a p/s contribution.

If a plan with only deferrals and match fails the comp test, would the correction call for a QNEC in the amount of the deferral percentage on the excluded comp? and it's corresponding match?

[Guest]

I vote yes.

consider a plan that excludes commisions.

you make 110,000, but 100,000 in commisions. you will not get a large match.

but your boss makes a straight 150,000. he will get a full match.

now, ignoring the fact that the ACP test might fail, you have definitely been discriminated against based on comp. and that is why you test.

by the way, a 3% difference in testing has been considered de minims, though there is nothing locked in concrete.

obviously you do not have a safe harbor 401(k) because you are required to use a definition of 414(s) that passes.

I'm not sure what you do to correct - a QNEC might go to all ees and so that wouldn't work. maybe thumb screws or a note from mom saying you won't do it again.

[Guest mo again]

I think I vote no. Check out 1.401(a)(4)-1(B)(2)(ii)(B), let me know if you agree.

[MWeddell]

In general, the answer is no (in my opinion) for reasons explained at

http://www.benefitslink.com/boards/index.php?showtopic=4313

[Guest]

I will agree with your comments with the caveat that the plan is not intended to be safe harbor since the definition of comp is not 414(s). if the plan was intended to be safe harbor then running the comp test could 'prove' (tounge in cheek) the plan has a safe harbor definition of comp.