Kern’s Welfare-Funding Repair Plan Names One Quarter; the State’s Instruction Reaches Further

A narrow promise beside a broader correction instruction

Kern County’s newly filed welfare-funding corrective plan contains a small but consequential difference between what state reviewers requested and what the county expressly promised. California’s Department of Social Services told Kern Human Services to correct a misclassified expense in the December 2025 quarter and any subsequent quarters. The county’s response names only a December 2025 adjustment claim, with an anticipated January 31, 2027 completion date.

What’s the Scoop With Broach compared the finding, requirement, proposed action and deadline columns in the full plan filed with the October 13 Board of Supervisors packet. The mismatch does not prove that later claims contain the same error or that the county intends to leave one uncorrected. It establishes that the written response does not explicitly commit to the entire period covered by the state’s instruction. That makes the eventual adjustment’s coverage, as well as its completion date, a public accountability question.

The expense was disability legal advocacy under a contract. It had been claimed under a program identifier ending in expense code 005, housing assistance; reviewers said code 031, contracted activities for unemployed people, was the appropriate classification. The document does not give its dollar amount, establish that the advocacy never occurred, or demand a repayment in this finding. A coding correction should not be recast as a proven loss of housing aid.

Seven missing attachments became seven supplied checks

The same column-by-column comparison changes the meaning of another finding. Seven named purchase-order selections initially arrived without the required federal exclusion-system queries: samples 9, 10, 11, 17, 18, 19 and 24. The plan says the queries were later supplied upon request. Counting these seven as seven vendors proven never to have been screened would erase the finding’s own counterevidence.

There were four named suppliers across those seven selections: Advanced Data Storage, Tel-tec Security Systems Incorporated, American Business Machines and Trans-West Security Services. Multiple selections involving the last supplier are not additional distinct vendors. The document supplies neither a total sample denominator nor a quantified ineligible-payment amount, so these counts do not support a departmentwide screening-failure percentage.

A separate contract selection, involving Bakersfield Drug Testing, lacked suspension-and-debarment language. Kern responded that this was a countywide agreement and promised to work with County Purchasing on contract language while continuing annual SAM queries. The state’s requested control also calls for keeping the queries with purchase orders of $25,000 or more. Finding the records on request and routinely retaining them alongside the relevant purchase order are different control tests. None of these passages identifies a supplier as actually debarred.

Training knowledge and written controls gave different signals

The fiscal monitoring occurred May 11–15, 2026 and reviewed a random sample of timesheets and time studies for the December 2025 quarter. Two caseworkers and two supervisors interviewed by the state accurately understood time studying. Their answers differed about how often formal training occurred, and reviewers found the supplied written procedures insufficiently detailed.

Kern promised updated induction and annual training and a detailed manual with task and code descriptions by December 31, 2026. Those are responses to documentation and succession risks; the four favorable interviews are material evidence against describing the finding as proof that staff universally did not understand their work. The public plan does not disclose enough sampled time-study results to calculate a billing-error rate.

A repeat accounting finding remains a promised repair

The plan also carries forward a revenue-recognition material weakness from the June 2024 Single Audit, itself identified as repeating the 2023 finding. In the state’s account, revenue was understated in the General Fund and overstated in Behavioral Health and Human Services funds, requiring material ledger corrections. The plan does not quantify those corrections, and a fund-level misstatement is not evidence that cash disappeared.

Human Services acknowledged the repeat finding and said its corrective work had begun. It committed to accrual and revenue-recognition training, desk-manual procedures, communication with administering departments and mock-claim functionality in CECRIS by December 31, 2026. A begun plan and a completed, tested control are different statuses. The department’s September 18 certification records approval of its plan; the CDSS approval table has no entered approval date in the reviewed copy.

This report is original public-document analysis, not a new audit or an interview with county employees. We reconciled the plan’s four types of issue, counted the named selections without treating them as distinct suppliers, and compared the correction instruction with the county’s stated scope. As of October 15, the reviewed records do not establish completion, later-quarter miscoding, federal exclusion violations or a dollar loss. The strongest finding is the specific gap between a correction instruction and the written commitment to carry it out, with the supplied records and favorable interviews preserved as counterevidence.