South Carolina Accepts 92% of a Medicaid Refund Finding; Its Earlier Correction Covered Only Part of the Gap

The dispute is not over the whole $108.6 million

South Carolina’s response to a new federal Medicaid audit accepts a $99,983,506 refund recommendation while disputing a separate $8,592,050 calculation. A public-record reconciliation by What’s the Scoop With Broach finds that the accepted portion is 92.1% of the combined $108,575,556 finding. Calling the entire amount contested would obscure the much larger reporting error the state has acknowledged.

The September 30, 2026 HHS inspector-general report examines federal shares of Medicaid and Children’s Health Insurance Program collections from January 2020 through March 2025. Collections—such as drug rebates, recovered overpayments and receipts from other responsible payers—reduce the net program costs submitted for federal reimbursement. Underreporting the federal share therefore leaves the state with more federal money than it should receive, even when the underlying collections are present in state accounting records.

The central accountability question is whether a prior state correction resolved the federal reporting mismatch. The audit’s quarterly table and the state’s response show that it did not. But the same records also demonstrate real corrective activity and narrow the remaining disagreement to the first pandemic quarter, rather than the full multiyear total.

A catch-up entry offset about a quarter of the COVID-share gap

Appendix B reports $104,959,324 in net variances attributable to the increased COVID-era federal matching share. Its December 2023 row contains a negative $34,285,572 COVID-share variance: South Carolina had identified earlier underreporting and included some of that share in its federal report for the quarter. The table’s note explicitly says the correction did not include the full amount.

We added that negative catch-up amount back to the net COVID-share variance to calculate $139,244,896 in positive COVID-share variances across the other displayed quarters. The December entry offsets 24.6% of that aggregate. This is a table reconciliation, not a finding that a quarter of every transaction was corrected or an exact allocation of the catch-up entry to individual earlier quarters.

The report also lists a net $4,975,820 offset from unrelated variances and an explicit $99,983,506 recovery recommendation. Subtracting the two printed variance components produces $99,983,504, a $2 discrepancy; we use the report’s stated recovery amount rather than invent an explanation for that small inconsistency. The offsets still matter: the audit says some other mistakes, including duplicating collections in reporting worksheets, overreported the federal share. Its recommended recovery is net of countervailing errors. Adding the gross COVID-share variance to the refund recommendation would double count dollars.

The accounting error can survive a seemingly correct match rate

Table 1 reconstructs a March 2021 drug-rebate example. State records held $70,990,015 in total collections and $54,216,716 in federal shares, including the temporary pandemic increase. The federal report instead showed $64,730,603 in total collections and $49,732,522 in federal shares. The difference in the federal column was $4,484,194.

The report explains why the federal form could appear to apply the enhanced match while still omitting money. The state took its recorded regular federal share, divided it by the enhanced matching rate to reconstruct an understated total collection base, then applied the regular and enhanced rates to that smaller base. The resulting total federal share equaled only the originally recorded regular share. A correct-looking percentage did not repair an incorrect starting amount.

That mechanism makes reconciliation to actual accounting receipts—not simply checking that a rate field includes the COVID increase—the relevant control. Auditors attributed the errors to missing written preparation and reconciliation procedures. The state said it had begun updating those procedures, was engaging a third-party vendor to review them and develop a training guide, and had developed a report comparing actual and historical matching rates.

What South Carolina still disputes

South Carolina said it would work with CMS to adjust future reports and refund the $99,983,506. That is a commitment recorded in a response, not evidence that the payment has already been made. It rejected the separate first-quarter $8,592,050 recommendation, arguing that early 2020 collections related to expenditures incurred before or during the quarter and should use the original, non-enhanced rate.

The inspector general maintained its calculation. It said the state received enhanced reimbursement for that quarter and had consistently used the rate in effect when collections arrived for the categories included in the calculation. To be conservative, auditors excluded categories for which the state traditionally attempted to match collections to the original expenditure period. The $8.6 million is not already included in Appendix B’s $99,983,506 net variance; Appendix C calculates it separately.

The federal-share rules and temporary enhancement apply to government financing, not a new charge to patients. The audit does not establish that Medicaid enrollees committed wrongdoing, received improper care or should lose coverage. The dispute concerns how the state credited federal shares of recovered money.

Method and remaining verification boundary

We read the full audit, its March 2021 example, both calculation appendices and the complete state response. We recomputed the accepted share and the December correction’s relationship to the other COVID-share variances, keeping negative offsets and the separately calculated first-quarter amount distinct. This original analysis establishes the scale of the acknowledged reporting error and the limited reach of the earlier catch-up entry; it does not independently audit South Carolina’s transaction ledger.

As of October 14, 2026, the records reviewed here document recommendations, concurrence and planned control changes, not a final CMS recovery determination or a verified completed refund. The response supplies meaningful counterevidence of correction and disagreement, but neither cancels the audit’s remaining net finding. A later reporting adjustment or payment record would be needed to establish actual recovery.