Maryland’s legislative auditors reported that a resident who collected a $2 million lottery prize kept drawing food assistance for more than two years afterward, and that at least 1,858 payments went to incarcerated people who did not qualify for them. The finding, released in July, is the sharpest illustration this month of a pattern that recurs across state audit reports: eligibility data exists, the computer matches fire, and nobody acts on them.
This month’s State Watch reviews accountability developments in eight states, drawing on state auditor reports, pension fund disclosures, credit rating actions and court filings. Where a matter is an allegation rather than a finding, it is identified as such.
Maryland: nine findings, a lottery winner, and a $240 million federal exposure
The Maryland General Assembly’s Office of Legislative Audits released a July 15 report on the Department of Human Services’ Family Investment Administration covering June 1, 2021 through Feb. 28, 2025. Auditors outlined nine findings, centered on the administration’s failure to conduct timely follow-up on questionable recipient eligibility flagged by its own computer matches.
According to the report, a recipient who won a $2 million Maryland Lottery prize continued receiving benefits and collected roughly $9,000 after the win was flagged. Auditors also identified at least 1,858 instances in which incarcerated individuals — categorically ineligible for the Supplemental Nutrition Assistance Program and Temporary Cash Assistance — received those benefits. In dozens of additional cases, local departments of social services could not produce documents verifying that recipients met employment, income or immigration-status requirements.
The audit also examined a 2024 procurement for “tap and go” electronic benefit transfer cards. Auditors wrote that the winning vendor’s bid omitted required pricing information, which allowed it to come in $11.1 million below the competing bid. The contract was later terminated after the losing bidder appealed. The department disputed that finding, with Acting Secretary Stacy Rogers writing that labeling it a failure to follow request-for-proposal terms “significantly oversimplifies the procurement process.” The department acknowledged a mistake in evaluating the two financial proposals.
The fiscal stakes are not abstract. Federal data cited by Maryland Matters shows the state cut its SNAP payment error rate from 13.64 in fiscal 2024 to 13.08 in fiscal 2025. Under the cost-sharing formula enacted in H.R. 1, states above a 13.30 error rate receive a one-year grace period; states between 6.00 and 13.30 pay a share of benefits. By improving just past the threshold, Maryland becomes liable for more than $240 million in SNAP costs. Auditors noted they received a hotline referral alleging that department management may have intentionally manipulated the error rate to defer that liability. They were unable to substantiate the allegation and said the review “did not identify any matters that warranted a referral to the Office of the Attorney General’s Criminal Division.” That allegation remains unproven and should be treated as such.
Ohio: a 15.6 percent Medicaid eligibility error rate in tested samples
The Ohio Auditor of State’s 2025 State of Ohio Single Audit covered $45 billion flowing through 371 federally funded programs, with $28.5 billion going to Medicaid. Auditors identified 18 findings and questioned costs of nearly $6.8 million across seven agencies, including more than $6.5 million in summer food assistance provided to students who were age-ineligible.
The more consequential number is a projection. Auditors reported a Medicaid eligibility error rate of roughly 15.6 percent in tested samples, including payments for services to residents who had died. Auditor of State Keith Faber stated that if the same parameters held across the full Medicaid population, potential unallowable costs would fall between $800 million and $4.4 billion. That is an extrapolation from a sample, not a confirmed loss, and Ohio’s report presents it that way.
The audit also documented a plausible mechanism: 16.9 million alerts were routed to county Job and Family Services caseworkers for review. Auditors wrote that “the volume of incoming alerts/matches being sent to county caseworkers results in an increased workload and ineffective application of the alert process.” Two findings for recovery were issued — $50,000 against a business owner who received fraudulent TechCred reimbursements, and $6,862.33 against an employee paid by two employers for identical hours. The office’s Aug. 4 release advisory lists a further finding for recovery attached to the City of Lorain’s 2023 financial audit.
Washington: $37 million in questionable child care subsidies
The Office of the Washington State Auditor reviewed $23.7 billion in federal funds across 28 programs and reported an estimated $37 million in questionable child care subsidy payments for 2025 — $27.2 million from the federal Child Care Development Fund and $9.9 million from Temporary Assistance for Needy Families. Auditors found providers who did not respond to requests for attendance records, overbilled for services unsupported by those records, or lacked required parent signatures.
The structural finding is the important one. The Department of Children, Youth and Families does not review supporting documentation before paying providers; it relies on post-payment audits. Reviewing a year of the department’s own audits, the State Auditor’s Office found that 67 percent identified overpayments and that 22 percent of audited payments were overpayments, amounting to $2.2 million. State Auditor Pat McCarthy said the state “should take additional steps to detect and prevent improper payments.” In context, Washington’s report showed improvement: 50 findings, down from 82 the prior year.
Hawaii: lieutenant governor indicted in COVID contract probe
An Oahu grand jury returned a 12-count indictment on July 24 charging Lt. Gov. Sylvia Luke with bribery, criminal conspiracy to commit bribery and falsifying campaign committee reports, according to the state Department of the Attorney General. Four others were charged, including a former state representative, a lobbyist, a former Public Utilities Commission chair and a state transportation official. Prosecutors allege the conduct relates to efforts to secure state funding for COVID-19 community testing in 2022, when Luke chaired the House Finance Committee.
Attorney General Anne Lopez said her office had “a commitment to following the facts to where they lead.” A supervising deputy attorney general stressed the presumption of innocence. These are charges, not findings. All five defendants are presumed innocent, and TIJ will report the disposition. Luke has not responded to a request for comment; the record on right of reply will be updated as it develops.
New York: Albany downgraded, New York City warned
S&P Global Ratings lowered the City of Albany’s long-term rating to A from A+ with a negative outlook on July 15, citing “reduced financial flexibility following its large, anticipated drawdowns in fiscal years 2025 and 2026 and its already thin available general fund reserves.” The agency noted the city “has depleted its fund balance, limiting its ability to respond to current and future operating pressures,” even after a $60 million state aid commitment in May. Mayor Dorcey Applyrs, who previously served as city auditor, said the assessment “reflects conditions that developed before my Administration took office” and pointed to a multi-year stabilization plan.
At the end of July, Fitch and Moody’s declined to downgrade New York City but maintained a negative outlook on roughly $53 billion of general obligation debt, warning that the city must narrow projected deficits. On the pension side, Comptroller Thomas DiNapoli reported that the New York State Common Retirement Fund returned an estimated 11.94 percent for the fiscal year ended March 31, closing at a record $295.4 billion against a 5.9 percent assumed rate of return and a 92.2 percent funded status as of March 2025.
California: strong returns, unresolved eligibility risk
CalPERS reported a preliminary 14.8 percent net return for the fiscal year ended June 30, its best in five years, lifting the Public Employees’ Retirement Fund to $637.1 billion and a funded status of 85 percent, up from 79 percent. Those figures are preliminary and subject to revision.
Investment performance does not resolve administrative controls. The California State Auditor’s investigations report substantiated more than $5 million in waste and improper activity, including $4.6 million the Employment Development Department spent on monthly service fees for more than 6,200 mobile devices that sat unused — some in storage for two years or more — a $171,446 overpayment to an Air Resources Board employee across 15 months of extended leave, and roughly $400,000 in taxable housing benefits the Yountville Veterans Home did not report. Separately, the auditor’s forthcoming work includes a state high-risk audit of Department of Health Care Services Medi-Cal eligibility, currently estimated for release in fall 2026.
Illinois, New Jersey and the national pension picture
Equable Institute’s State of Pensions 2026, published July 23, puts the projected national average funded status at 85.0 percent, up from 81.2 percent, with total unfunded liabilities of $1.13 trillion. Forty-five states improved. Two remain below 60 percent funded: New Jersey at 56.7 percent and Illinois at 56.4 percent. Six states combine sub-80 percent funding with pension debt exceeding 10 percent of state GDP — Illinois, Kentucky, New Jersey, Mississippi, Hawaii and New Mexico — a combination the report describes as beyond what market returns alone can resolve.
Two risk signals in that report deserve attention. Employer contributions have reached 31.83 percent of payroll, a historic high, but only 9.42 cents of each dollar funds new benefits; the remaining 22.42 cents services accumulated debt. And the share of pension assets priced by valuation models rather than market prices has risen to 27.1 percent, roughly triple the 2001–2007 average. Equable estimates direct exposure to a basket of artificial-intelligence-related companies at 8 percent to 10 percent of assets, or $513 billion to $642 billion. If those valuations are overstated, reported funded ratios are overstated with them.
Pennsylvania and Texas: routine compliance work, cumulative signal
Auditor General Timothy DeFoor announced on July 31 the release of compliance audits of 25 municipal and police pension plans across 16 counties, including Harrisburg’s non-uniformed plan. These audits verify that plans use state aid consistent with law, and their volume makes them a useful running index of local pension administration quality.
The Texas State Auditor’s Office released an audit of educational loan repayment programs at the Higher Education Coordinating Board in July, alongside a follow-up audit of cybersecurity controls at the Texas School for the Deaf.
The recurring theme
Across Maryland, Ohio and Washington, the failure is not detection. All three states run the data matches that identify ineligible recipients. Maryland’s system flagged the lottery winner. Ohio generated 16.9 million alerts. Washington’s own post-payment audits catch overpayments two-thirds of the time. The failure is that detection is decoupled from action — alerts outrun caseworker capacity, and payment controls sit downstream of the payment. States that pay first and audit later are structurally guaranteed to recover less than they lose.
What warrants deeper investigation
Four threads merit sustained reporting. First, whether Maryland’s SNAP error-rate improvement was driven by genuine process change or by sampling and casework decisions — auditors could not substantiate the hotline allegation, but the $240 million consequence of landing just under the 13.30 threshold makes the question a legitimate one for the record. Second, whether Ohio’s $800 million to $4.4 billion Medicaid extrapolation holds up against a full-population review, which no state has yet published. Third, whether pension funded ratios reported at record levels survive an independent look at valuation-priced assets, given that 27.1 percent of holdings are not market-priced. Fourth, whether the wave of thin-reserve municipal downgrades in New York — Albany is not an isolated case, by S&P’s own account — reflects one-time pandemic-aid cliffs or a durable structural gap.
Every figure above is drawn from a published audit, an official pension disclosure, a rating agency action or a court filing. Where a claim is an allegation or a projection, it is labeled. TIJ will follow each of these matters as documents become available.