Changing ADP/ACP Testing Methods

It's been a long time since I've worked on a non-safe harbor 401(k) plan, so a refamiliarization is in order.  A plan that has been using current year testing for >5 years fails both the ADP and ACP tests for 2022 and switching to the prior year method is being considered.  The plan was amended in mid-2022 to eliminate the eligibility requirements, effectively permitting about 20% more employees to become participants when compared to the prior year.  Is it safe to presume that this would not be considered a plan coverage change pursuant to Treas. Reg. 1.401(k)-2(c)(4) and that it would be OK to change to the prior year testing method and just use the actual NHCE ADP/ACP results from 2021?

Also, my understanding is that earnings for any refunds made by 3/15/23 should only be calculated through 12/31/22 using any reasonable method - is this correct?  Thanks in advance for all assistance. 

I don't think that an amendment to change the testing method can be adopted retroactively. In other words, if they wanted to change from current year to prior year testing for 2022, they would have had to do so by December 31, 2022.

For determining the earnings on excess contributions that are being refunded, the regulations allow a plan to use any reasonable method, but you need to check your plan document to see what method it specifies.

With the expanded eligibility have you tested otherwise excludeable employees separately? If not it may help your testing results. Or you may have already tried and it still fails.

If it's too late to change the testing method for 2022, perhaps doing it now for 2023 testing is in order, as they will probably have the same problem.

The plan hasn't yet been tested for otherwise excludable employees (aka permissive disaggregation?).  None of the extra employees who entered due to the amendment are HCEs, so the OEE part of the plan would automatically pass if I understand correctly.  However, the rest of the plan (the non-excludable employees) would still fail, but by a smaller margin.  Is it all or nothing in such a case, i.e., can you just refund the excess deferrals/match based on the results of the disaggregated testing, or are you forced to make the refunds based on the original test before disaggregation?           

Unlike the current year / prior year testing options that must be stated in your document, you can change your testing for otherwise exludable from year to year and use which ever results give you the best case scenario. Though we usually find throwing out the folks who have not met the statutory exclusions tends to give the best results as there are often a lot of NHCEs at 0% in the that group.

Test it as hard as you can first, then issue refunds once you're out of testing options and have your best results.  If the Plan previously passed, and the eligibility change is what's mostly affecting this year, then you can likely get a passing test again.

What's wrong with refunds?

To me, the only thing prior-year testing does is have some of the HCEs put in less than they could.  It rarely works as intended, unless ALL the HCEs put in exactly the amount needed to pass the test.  If one HCE goes below that, then the others can go just a little bit higher.  But who keeps track of that?

Say the max HCE average is 4% and I have 4 HCE.  All HCE can put away exactly 4%.  But what if one lady lowers hers to 2% midway through the year?  Now, at teh end of the year, the HCE combined % is 3.75.  Thus the other 3 could've gone a little higher.  Do yu want to be the one explaining to the HCEs they did not get the maximum they could?

I tell all my clients "refunds aren't bad."  It's guaranteeing they get the most into the plan they are allowed.  It's only delaying the taxation on the amounts by a year.

As an HCE I'd much rather put in $8,500 and get a $300 refund that only put in $8,000 only to find out I could have done $8,200