Relius Cross testing - Accrued to Date Method

[bcmom]

I have a straight profit sharing only plan where the prior TPA used the Accrued to Date method for cross testing. I don't see where Relius offers that method. If it does, where do I code it?

[Bill Presson]

I believe that's just a testing method and not an allocation method. See if it's an option when you run the a4 testing.

[BG5150]

I don't think Relius does it.  I went down that path (or tried to) a couple years back.

Try Relius customer service.  They are really good in my experiences with them.

[bcmom]

Yes, its a testing method and not an allocation method. I submitted an incident to Relius to see what they say. My guess is that it can't be done. I've never used it and wasn't aware it existed. I don't recall this method discussed at any ASPPA sessions.

[Mike Preston]

Then you didn't go to the right sessions.

[Tom Poje]

1. I think in all the years I attended ASPPA Conferences the accrued to date method for DC plans was available once maybe twice. I have a copy of notes from 1998 Conference [Annual vs Accrued to date]

Also have a copy of notes in which the examples are as of 1995 which I obtained from somewhere.

but I don't recall the topic discussed at any of the conferences I attended, so I'm not sure I would say "you didn't attend the right sessions" is quite accurate, though I have gone in recent years. but I have seen the list of topics discussed and I don't remember that being on the list. (unless it was only briefly mentioned as part of the general discussion)

2. correct, Relius does not perform accrued to date on DC plans. of course, even if it did, it could only test on years you have on the system.You say the prior TPA ran the test that way, it implies you have a takeover case, so without historical data on the system you would be limited running accrued to date even if the system did that. (as I recall you could pick any year as a starting point), or as the notes I have say "you request the 12/31/1998 balances and years of participation for all employees" -which shows how old the notes are!

Just because accrued to date was used previously does not mean you have to continue to test that way.

[BG5150]

47 minutes ago, Tom Poje said:

Just because accrued to date was used previously does not mean you have to continue to test that way.

Yeah, but the bind you can get into is if the accrued to date method gives better results to the client.  I had that very situation (though I do not think the previous outfit did the testing correctly).  Their testing had the owners maxing out and the staff getting just the gateway.  When I tested it without accrued to date, the staff was getting somewhere around 7%.  We had to explain to the client the previous method of calculation wasn't something we could readily do.  I eventually got it down to around 5 1/2% with component testing, but the client wasn't thrilled.

[Mike Preston]

I can't believe that Relius doesn't support accrued-to-date.  Without permitted disparity isn't it as simple as taking a ratio of A/B where B = the annual testing benefit divided by compensation and A is the accrued to date benefit divided by average compensation and further divided by years benefitting?  In other words, give me the results of the annual method and a simple spreadsheet can generate accrued-to-date testing with just a couple of inputs.  Imputing permitted disparity is, I agree, a  bit tougher.

[Tom Poje]

BG5150 -agreed that can happen, which is why people go accrued to date method

 

my comment was simply meant to imply that one is not locked into a particular method that was done in the past.

but you raise a valid point that it could save some on the overall contribution.