Rate Group Test

[Chippy]

I have a 401(k) plan with a new comp formula for the ER non-elective. When running the nondiscrimination testing on Relius, the plan is passing ABT and Gateway. The plan is failing the ratio percentage test. If I run the test and check the rate banding box on relius then it passes. Has anyone used this as an option to get the plan to pass? If it can be used, can you explain what it does? I tried the stat. exlcusion and impute disparity options, but the plan still wouldn't pass.

[Tom Poje]

as an explanation I will start with an example.

for rate banding the regs say you can use a 5% difference above and below a midpoint.

so lets suppose my midpoint is 10%.

5% either side of that is .5, so anyone between 9.5 (lo rate) and 10.5 (hi rate) is treated as being '10'. this is greta for getting other people into the rate group.

the problem (?) with Relius is that it will starte with the HCE at the hi rate and back into the midpoint and the lo rate.

the regs are pretty specific that rate banding is fine as long as it doesn't overly favors the HCEs. well, I don't remember the exact language, but something to that effect.

In other words, if you put the HCE at the midpoint and bring NHCEs up to him, and NHCEs down to that rate, then that would be ok.

but to start with the HCE at the HI rate would not seem to pass the 'smell test'.

in other words lets say you had NHCE 1 = 10.47 HCE = 10 and NHCE2 9.75.

you could group around a midpoint of 10. you are bringing an NHCE up and bringing an NHCE down.

but if you had HCE = 10.47 NHCE1 = 10 and NHCE2 9.75 you are now bringing an NHCE up and an HCE down, which does not seem to be in the spirit or intent of how the regs are worded.

so I personally don't recomend using the Relius default. since you can set the rate bands (rather than using the default) you might be able to accomplish what you desire anyway and still meet the regulation guidelines.

It is also possible that using a pre interest rate of 8.5% and post rate of 7.5% will pass if you are that close.

or if you didn't use it, a mortality table of 1983 IAF

[ERISAnut]

Assuming the ABT was run correctly, what would be the issue? Does the employer have other plans; or is this the only plan and the ABT test was ran to include all sources of this one plan in order to pass.

The key here is that once you pass a correctly run ABT, then you have proven non-discrimination for 410(b) for all plans of the employer. This may be a compelling argument not to test any further.

[Chippy]

This is the only plan the ER has, I ran the ABT test to include all sources, EE and profit sharing. There are 126 NHCE's and 32 hce's. The NHCE's are 14.387% and 70% of the HCE's is 5.779. The plan is passing the ABPT but not the ABT.

410b is passing. 93 out of 126 nhce's benefit and 18 out of 32 hce's benefit.

When I run the 401a4 summary on Relius (which I know Tom doesn't like) It says the plan is failing the ratio percentage test and it is failing the nondiscrimination classification test. It is passing the ABPT.

the rate group that is failing has 1 hce and 1 nhce. When I check the rate banding option on Relius, it brings 4 additional NHCE's into the group. Tom, the ebar on the abt for this hce is 24.705 and the nhce is 21.622 but when I print out the rate group, without banding, it has the ben pct (ebar) as 16.196 for the hce and 16.914 for the nhce. How does it calcuate the ebar for the rate group? With rate banding, it shows the hce at 14.654 in the rate group. The nhce is still at 16.914 with the rate banding. I guess I don't understand the percentages used for the rate banding.

[Tom Poje]

before trying to explain the rate group report, lets see what you have

126 Total NHCE

32 total hces

158 total ees

126/158 = 79.747% NHCE concentratio %

always round down, so you have 79%

if you calculate / look up on table / whatever translates to a mid point of 30.75.

in the rate group you have 1/126 for the NHCEs

and 1/32 for the HCEs

dividing these two numbers = 25.397 which is less than 30.75, so fails.

the good news is that you only need 1 more NHCE to pass nondiscrim classification testing.

since you already pass the avg ben % test, then everything works if you get just one more NHCE.

I am going to assume when you said HCE = 24.705 e bar, you mean for the avg ben % test. those e bars are used for that test only and not the nondiscrim test under 401(a)(4).

the problem with the rate group report is it will only show you ees who are in the rate group and not someone who is real close.

now, you said the HCE was at 16.196. and then when you rate banded it shows him at 14.654

here is what has taken place: the system has determined a midpoint of 15.425.

5% on either side is .771 so 15.425 + .771 = 16.196 Hi rate and 15.425 - .771 = 14.654 lo rate.

what the most important question is: what is the E-Bar for the NHCE next in line . who is the closest to the HCE rate of 16.196?

lets suppose I set the HCE as the midpoint = 16.196. 5% either side is .8089, so this gives a Hi rate of 16.196 + .8090 = 17.006 which would include the NHCE at 16.914.

the lo rate would be 16.196 - .8090 = 15.387. so if you have an NHCE close to that you are probabaly ok. (and can generate the report to indicate so.

there are some other possible testing assumption that might work.

If the HCE was born in the 2nd half of the year then using age last will help if the NHCE was born in the first half of the year.

[Chippy]

I ran the test with rate banding and I set the bands manually for that group. 17.01 for the high and 15.39 for the low. (relius would only take 2 decimals) With those, my hce is at 15.39. and it brought 4 nhce's into the group at 15.589%. So it is now passing with flying colors.

I have never used the rate banding option on Relius, but for future to be sure I understand, if I have a group that is failing, I use the HCE as the midpoint and manually input the high and low for the rate bands based on the 5% above and below? If there is more than 1 hce in a group that is failing, do I average their percentages to come up with a mid point?

[Tom Poje]

first - good job and congrats, especially that you could even follow my notes and translate them! It can be quite a complicated subject

now as for rate bands, I suppose setting the HCE at the midpoint is a good starting point. If the HCE ended up a little above the midpoint, its probably ok, but I certainly wouldn't go with the HCE at the top and bring NHCEs up. (fair is fair - that could be problematic if the plan ever was audited)

the 5% is only a maximim range. there is no reason you can't use something else. so with more than 1 hce you could have a rate band between 4.8 and 5.2 and another between 5.21 and 5.3 - as long as the rate band don't overlap. the range can be different from one rate band to another. (e.g. you could use 5% for one band and 1% for another.)

or you could do like you said with 2 HCEs and combine them. I simply frown upon having one HCE at the top.

see 1.401(a)(4)-3(d)(3)(ii) which says "...accrual rate may not be grouped if the accrual rates of HCEs within the range generally are significantly higher than the accrual rates of the NHCEs within the range."

prior to Relius 13 I could only enter one rate band, with Relius 13 I can enter more than one. either they fixed something or we had a glitch on our system.

[Chippy]

Thank you very very much. This was a big help and much appreciated!

[BG5150]

so I personally don't recomend using the Relius default. since you can set the rate bands (rather than using the default) you might be able to accomplish what you desire anyway and still meet the regulation guidelines.

Has anyone taken flack from the IRS or an auditor by using Relius's aggressive rate banding mid-point selection?

[Tom Poje]

I think a more important question would be 'has an auditor even noticed?'