non discrimination

[Gary]

owner has 12/31/10 AB = 9,000 per month payable at 62

owner has 12/31/11 AB = 11,000 at 62

plan amended to change early ret at 55 from act equiv to fully subisdized unreduced AB at 55.

12/31/10 AB payable at 55 after amendment = 7,000 (limited by 415)

12/31/10 AB act equiv at 55 before amendment = 6,000

12/31/11 AB at 55 after amendment = 9,000

So for annual accrual for 401a4 testing at age 55 which is more correct?

1) 9,000 - 7,000 = 2,000 payable at 55 prior to normalizing

2) 9,000 - 6,000 = 3,000 at 55 prior to normalizing

thanks

[AndyH]

You need to test both the (NAR) 12/31/2011 Accrued benefit at NRA less the 12/31/2010 Accrued benefit at NRA. Sounds like $3 K.

and then for the MVAR you need to test the QJSA equivalent of each benefit difference at age 55 12/31/2011 less 12/31/2010, each normalized.

And then you can try accrued to date.

Fun stuff.

Reminds me of what I heard a prominent speaker at a conference say once (not my advice) - "just use what your software spits out for the MVAR since nobody can figure it out anyways".

[Gary]

to put my question another way.

if a plan makes an amendment that impacts past service and future service as in my ecample.

The question is:

should the increase in the benefit for past service be part of the accrual for a4 testing?

or should a4 test only on the accrual related to the one year of service when using annual method of testing?

so another numerical example could be:

AB 12/31/10 = 1000 before amendment

AB 12/31/10 after amendment = 1200

AB 12/31/11 after amendment = 1500

so should accrual be 1500-1200 = 300 or current year

or 1500 - 1000 = 500 which includes impact of amendment on past service as well?

obviously this can have a huge impact re: testing.

the thought being that the past service amendment is determined on its own merits as either discriminatory or not

and the current year accrual (ignoring past service increase in benefit) is tested separately as well.

thanks

[AndyH]

The benefit that is tested is the change in benefit during the measurement period. The measurement period is the option, not the methodology.

If the measurement period is the current plan year, then the eoy AB must be compared to the prior EOY AB, and a major amendment such as a past service amendment can easily make the test fail. Yes, it must be included.

If the measurement period is the current year and all prior years that can lessen the impact of a retroactive benefit increase, or any benefit increase for that matter, provided that you can justify the testing service used in the denominator.

The 5 year past service rule is a separate matter.

[Gary]

yes accd to date is another option and of course that can result in the accrual being perhaps just what the one year accrual without re: to past svc piece of amendment as mentioned above.

say in my example the AB was 1000 after 5 years of svc.

and after amendment it is 1500 after 6 years of svc

for the one year the accrual is a massive 1500-1000 = 500

however using accd to date it is now 1500/6 = 250 (which is even less than the 1500- 1200 =300 above)

interesting; accd to date based on years of participation will generate the best results.

cool!