Corrective Amendment

[Archimage]

I have a plan that just gave me the correct birthday for a participant after everything has been done for the 2003 plan year including the tax return. Now the 401(a)(4) test fails. Am I correct that I can do a 11(g) amendment to give a contribution to one participant (100% vested) in order to pass the test?

[SoCalActuary]

You are headed for an argument with IRS over substantiation quality data.

Is this a small plan, where the employer was expected to have accurate info on each person? If so, why was the birth date record revised? If it occurred in 2004 after an employee was found to lie on their employment information, you may have an argument to ignore it for 2003, but you may have to use the new date in 2004 administration. There is some legal history that the employer can rely on the original birth date for employee benefit amounts.

This might involve consulting a good ERISA attorney.

If it is a large plan, then you can probably ignore it.

On the other hand, if the birth error was due to careless data collection, then you should probably fix 2003, since either the employer or you are at risk.

[Archimage]

This is due to careless data collection by the client.

[SoCalActuary]

Do you feel lucky? Clint Eastwood had some great lines.

If you don't correct, then you face audit roulette. Not my recommendation, but also not my call. If the employer tax return is already complete, then deduct it for 2004. Otherwise, you should amend 5500 and admin.

[AndyH]

I'm not sure I follow this logic. It seems to me that if a test fails, a test fails, and you have until the following 10/15 to correct it, provided that you correct it by providing something extra, and fully vest it.

What does the reason for a data error have to do with this?

Under what circumstances would you not correct it if you are within the 11-(g) period?

[Tom Poje]

I don't even see where you are getting 'fully vest it' from.

You have to provide a meaningful benefit. If the contribution was going to a terminated ee who was zero vesting, then I could see fully vesting it but if it was to active ee then I don't think that is necessary, as long as the contribution isn't just a few cents.

[AndyH]

Agreed.

[Archimage]

When I said 100% vested I meant the participant was already 100% vested. I was trying to alleviate 5 or 6 posts that would go into great detail about the additional contribution having to be a meaningful benefit. Unfortunately, I failed to do so.

[SoCalActuary]

To Andy H:

When I worked on large plans, 1000+ records, we would routinely get 80 to 90 pct accurate data. There were always data corrections in the next plan year.

In small plans, you have plan years already closed with completed admin, benefit reports, tax forms, and maybe even distributions made with incorrect data.

In the large plans, you just forget about last year's errors. In the small plans, you sometimes make the participant live with their incorrect date if they caused the problem, or you just correct the admin going forward. If distributions were made that underpaid a participant, I recommend the plan sponsor make additional payments. If overpayment was made, we would look to see if HCEs benefited unfairly and make a request to return funds.

In this case and the facts presented, the plan sponsor appears to be at fault for the error. Since the error would cause disqualification, I would push to have it corrected immediately, using the appropriate self-correction method. However, in other fact patterns, I might counsel that the error be corrected in the next year plan administration. Of course, since I work in db plans, that is often easier than with dc plans.

[AndyH]

Interesting comments, thank you.

I work on DB plans mostly as well, and the data correction issues are obviously much different there. Nobody is harmed by less than 100% accurate information because there are no account balances to build on, no contributions to divy up, and no assets to individually allocate (precisely).

But I would submit that if a cross tested DC plan is within the 11-(g) period, then use of a wrong date of birth is not a disqualifying failure. In the small plan world, census is often completed by people named "Bambi" who have trouble grasping the concept of census.

Sorry to any Bambi's out there, but I could not resist because that is a true current story.

[Archimage]

Does it matter at all which participant gets the additional contribution under the 11(g) amendment?

[SoCalActuary]

Good question. The 11g issue allows you to correct the discrimination problem. IMHO, you can amend in any reasonable way to pass the test, and you are not required to give the additional amt to the age-confused participant.