Coverage testing on defined benefit plan

[Guest forum4]

I am in the HR department of an organization that is part of a controlled group of employers. We exclude several of our affiliates from participation in our defined benefit plan. These excluded affiliates are composed primarily of home health care workers that usually work < 1000 hrs/yr. There are no HCEs employed with the excluded affiliates. We also exclude people that are categorized as "per diem" employees - if you are expected to work less than 80 hours a month then you are categorized as per diem within our organization (per diem employees general receive a very limited benefits package). Otherwise, eligibility is age 21 and one year of service based on elapsed time, and you enter 12 months after your hire date (or attainment of age 21 if later). There are no additional accrual requirements once eligibility is met

Our actuaries use a "snap shot testing date" for coverage testing purposes with the snap shot date being the last day of the plan year. As such, the payroll data supplied for coverage testing purposes is only those people employed on 12/31 of any given year. My interpretation from the ERISA Outline Book is that if you use the snap shot testing date, you shouldn't use the last day of the year so as to avoid unfairly carving out people who terminated within the year.

My opinion is that we are materially skewing our testing results to pass coverage testing. We are carving out all terminated people regardless of hours worked, and the majority of the people that do terminate within the year are with the excluded affiliates or the per diem employees. Are our actuaries performing coverage testing correctly?

[AndyH]

Before your second to last sentence, everything sounded fine. Since you have no accrual requirement, terminated participants are still benefiting. But the snapshot day, if I recall correctly, needs to be reasonably representative of the year. Sounds to me like that is where they are going wrong.

What is the passing margin? If the issue is 95% versus 90%, I wouldn't worry about it. If it is close to 70%, or close to the safe harbor percentage if the ABPT is used, then I would raise the issue.

[Guest forum4]

Use this as an example:

We have 2,400 non-union employees who have had a year of service and over age 21. 100 are HCEs (and all benefit). Another 800 belong to an excluded affiliate or are per diem employees (and excluded from the pension plan). That gives us a ratio of roughly 65%, which fails (2,300 NHCE, 1,500 benefiting).

Now assume that we have 200 people terminate throughout the year, and all of the 200 are per diem or employed with excluded affiliates. Based on the last day of the year snap shot date (looking at only those employed on the last day of the year) my ratio is now 71% and passing (2,100 NHCE, 1,500 benefiting).

These numbers closely parallel our situation (rounded to the nearest 100) and in fact when terminated people are brought into testing (or use a snap shot date of the beginning of the year), we fail 410(b) coverage. So, to rephrased my question, can you in fact use the snap shot testing date of 12/31 (end of the year), exclude terminated people and therefore, in our case, pass coverage (although barely)? As mentioned, the ERISA Outline Book says that the snap shot date should be close to the beginning of the year to have a reasonable representation of the work force and actual coverage. It is our actuaries that state that you can use the last day of the year as the snapshot date. My background is with 401K plans and I have had only limited exposure with DB plans (over a decade ago).

[AndyH]

Revenue Procedure 93-42 contains the rules governing snapshot dates. I believe that the requirement is as I described. I see nothing that prohibits a specific day from being available. My view is it would be ok if the last day would be reasonably representative, which does not seem to be true.

Please be aware that dropping below 70% does not mean that you failed coverage. It means that you failed the ratio/percentage test. You can still use the ABPT, and if you have a 401(k) plan covering most of these people, you will most likely pass the ABPT easily by testing the plans on a whole on a benefits basis.

[WDIK]

While not related to your question, you might want to consider the following.

You state that the service requirement for eligibility is one year of service based on elapsed time. You also state that the employees of the exluded affiliates generally work less than 1000 hours during the plan year.

If the plan's service requirement were one year of service based on 1000 hours, these employees would not be included in the coverage test.

Of course there may be some other factors involved in choosing the eligibility requirements that are not elaborated here.

[david rigby]

Do those working less than 1000 hours all work over 500 hours?

[Guest forum4]

These excluded individuals from the plan work varied hours. Some work over 40 hours a week. Management is not changing the document. I can't get a clear answer why eligibility wasn't 1000 hours with one year of service because vesting is based on 1000 in the plan year... We have "passed" 410(b) coverage in the past because the actuary has used a snap shot date of 12/31 (includes only active employees as of 12/31). This then carves out the terminated "per diem" classification of employees and excluded affiliate organization employees from even being considered in the coverage testing (even though some work more than 1000 hours/year), because this group represents the highest turnover within the company. To me, we are severely skewing coverage-testing results. However, based on the prior response from AndyH, this isn't an issue because we can use ABPT to easily pass coverage (don't know what ABPT is but I'm assuming the actuary would...). I would be curious to know if the ABPT testing would then be the numbers to report on the 5500, schedule H relative to coverage.

[AndyH]

By ABPT I mean the Average Benefits Percentage Test.

To pass coverage, you need to pass either the Ratio Percentage Test or the Average Benefits Percentage Test.

If your Ratio/Percentage falls below 70%, you have not failed coverage. You have simply failed the Ratio Percentage Test.

My educated guess, and that is all that it is, is that if you have a Ratio/Percentage anywhere near 70%, and you are testing a DB plan, and you also have a 401(k) plan, then you will amost certainly be able to pass the Average Benefits Percentage Test. This is because you can include the 401(k) plan in the ABPT if you wish. And it is less likely to be skewed towards HCEs. And you can, if you wish, convert 401(k) contributions to benefits and add them to the DB benefits in the test. And that almost always helps a great deal.

So it sounds like you have legitimate concerns about how the coverage requirements are being tested, and reported. But I doubt that you have a coverage failure.

You have more of a potential problem than a problem. But I think that you deserve kudos and congrats for noticing this before it was too late.

I'd hire somebody to get a second opinion if I were you. Sounds like sloppy work at best.

[Effen]

"I would be curious to know if the ABPT testing would then be the numbers to report on the 5500, schedule H relative to coverage. "

5500 Schedule T should have this detail.

[Guest forum4]

Thanks for everyone's comments and help.