I have a client with a Safe Harbor 401(k) plan where we're still testing a select population of employees. I've spoken with a friend in the industry who thinks there's away around this testing, and wanted to bounce it off of people here to see what the reaction was...

Plus, it's a fun little thought exercise.

Plan Facts:
401(k) Deferral: immediate eligibility
Match eligibility; 1 year of service
Match Formula - Assume statutory minimum safe harbor amounts, 100% of first 3%, 50% of next 2%
Eligibility service measured by elapsed time.
Matching Contributions calculated on a payroll basis

Current Testing Treatment
Group 1: Satisfied 1-year eligibility requirement prior to commencement of current plan year. They are elgiible for the safe harbor match for the entirety of the current plan year. Thus, no testing is performed on this group.
Group 2: Hired during the current plan year. They will not satisfy the 1-year service requirement by the end of the plan year, thus are statutorily excludible. These employees are tested, albeit separately in their own group with no HCEs (because none have any prior year compensation).
Group 3: Hired during prior plan year. These employees will attain 1-year of service during the plan year. Thus, by the end of the year, they are non-excludible and must satisfy testing. Further, because their matching contributions were made for only part of the plan year, they do not satisfy the safe harbor design and must be tested.

Each year, we test group 3 separately, and invariably there's always one or two employees who were hired early enough in the prior plan year to earn enough to reach HCE status. This testing regime was built into the plan document, as well, so any deviation on a go-forward basis would require a plan amendment.

The Proposition
The fix proposed by my friend is to, for the sake of testing each participant's matching contribution against the safe harbor standard, to look at deferrals of compensation where compensation is defined as only compensation while elgibible for the match. So, a participant hired 10/1/07, for example, would be excludible for 2007. In 2008, he would receive a matching contribuiton starting 10/1/08. When we look at matching made on deferrals of compensation for the period 10/1/08-12/31/08, the rate would meet the Safe Harbor formula, and no testing would be required.

The Authority
The closest thing I can find on this topic is Treas. Reg. 1.401(k)-3(b)(2) which, for safe harbor purposes, allows you to define compensation as compensation only during an employee's period of participation. The potential problems I see with this are as follows:
1) 1.401(k)-3(b) pertains specifically to nonelective contributions, and I'm unsure if we could apply it in the case of this plan which uses a matching safe harbor contribution.
2) The language merely says "period of participation." My initial reading of that provision (and all examples I've seen on the subject) interprets this to mean participation in the plan as a whole, which in this case would be triggered at hire as eligibility for deferral purposes is immediate. This provision seems to be addressing the case where all participants are eligible for the contribution, but participation in the plan requires 1-year of service. In those scenarios, paricipation in the safe harbor contribution and the plan as a whole occur at the same time. By contrast, in the plan at issue, participation in the plan, generally, precedes the match participation by a year. I haven't seen anything specifically permitting carving up participation on a benefit-by-benefit basis for this purpose.

Any thoughts? Does this proposed interpretation make sense, or is my friend halucinating regulations that don't exist? Would it be possible to treat this transition group as satisfying the Safe Harbor requirements, or is a plan being punished for extending participation for deferral purposes to excludible employees?

Thanks for the feedback.

I'm not a TPA, so I'll leave it for others to analyze your testing questions. But, I would point out that the definiton of "safe harbor compensation" in Treas. Reg. Section 1.401(k)-3(b)(2) is for "this section" of the regs--i.e., it applies to all safe harbor plans whenever "safe harbor compensation" is used, such as Treas. Reg. Section 1.401(k)-3(c)(2) (safe harbor basic matching contribution description).

I am a bit puzzled, so probably didn't understand something. The regs say you can have immediate eligibility and have a 1 year wait for match. (thus the ability to test otherwise excludables seperately) so if someone enters mid year, and the match is done on a payroll basis, I thought that still fell within the boundaries of satisfying the safe harbor (e.g. match cam be from date of entry) even though they only receive a match for 1/2 the year

Sieve - Thanks for the feedback. On re-reading, I agree with your analysis. Given how and where the term is used in the regs, I don't think there's any other reading that makes sense (as -3(c) uses the same terminology as -3(b) without providing an alternate definition).

As for clarifying the rest of the issue, perhaps I could re-state it more simply...

If a nonexcludible employee is eligible under the plan (for deferral purposes) for the entire year, but is eligible for the safe harbor contribution only for a portion of the plan year, does that partial year of Safe Harbor eligibility fully satisfy the nondiscrimination requirements with respect to that individual for the plan year?

TMFT --

I've not seen this plan design. But, be that as it may, how do you explain Treas. Reg. Section 1.401(k)-3(c)(5)(ii)? That allows the SH match to be based on payroll period calculations. Certainly what you are doing (basing the SH match on 1/2-yr period) is no worse than basing it on payroll periods for that 1/2 year during which the employee was eligible for the match, so it should meet the match ADP SH.

That reg. language could be read to require a retroactive contribution of the SH match even for payroll periods during the year when there was a deferral prior to eligibility for the match, although that would be somewhat difficult--but doable--from a timing perspective in light of the requirement to make such contributions no later than the end of the calendar quarter following the calendar quarter of the deferral.

All this assumes, of course, that the Plan doc. doesn't require some other measuring period (such as the full year) for the match SH (e.g., as required by the Corbel prototyye).