Bad information used in testing

[Zorro1k]

Client failed testing. Refunds issued and cashed. Discovered bad data was used. Tests rerun and client still failed but refund required for HCEs was less than what was issued?

As to issue of refund, can they self correct?

[Tom Poje]

how large a difference? (yes, supposedly you request ees to return the $ to the plan as a correction)

arguably 6.02© under EPCRS - recovery of small overpayments could be invoked (amounts less than $100)

so lets say the first time I ran the test I used comp from date of participation. Now I rerun using total comp so I fail more than I did using the bad data, but now due to the fact I am using corrected data it comes closer to my original results - . or I run the test not splitting out otherwise excludables, or one of the other optional methods of testing.

[hr for me]

Many many many years (prior to EPCRS), a coworker at a large HR consulting firm that shall not be named ran a test that failed with very very very bad data. To correct it, we re-ran the testing a few different ways and compared the results by person. And then chose the one that was least different from the incorrect one. And argued successfully to the DOL/IRS that the differences (both to different people AND different amounts) were small enough to overall be basically ignored. It took a lot of time/analysis and (my) wages. The HR consulting firm ate that cost which was very high, but it saved the client. Don't know if it would work today, but does sound like that is what Tom (above) is suggesting too.