Kansas Audit: Free Lunch Count Doubled, $38-53M Overpaid
Summary
Kansas's nonpartisan legislative audit office found the free-lunch student count that decides $488 million a year in school funding is more than double what Census data says should be eligible, with most of the applications auditors could actually verify turning out ineligible. The likely result, in the 2023-24 school year alone: Kansas overpaid an estimated $38 million to $53 million in at-risk school funding, and the federal government overpaid $10 million to $14 million in meal reimbursements -- on a program auditors say runs almost entirely on self-reported income nobody checks.
Report R-25-013⧉, published July 2025, found the free-lunch count has drifted far from its original purpose: about 198,000 students qualified for a free lunch in 2023-24, more than double the roughly 92,000 the audit estimates were actually living in poverty that year using Census data. The likely result is real money misdirected -- an estimated $38 million to $53 million in state at-risk funding, and $10 million to $14 million in federal meal reimbursements, paid out in the 2023-24 school year for students who probably did not qualify.
A formula that used to track poverty, and mostly stopped
The mechanics explain why this happened. A student can get a free lunch by having a parent submit an income application, or -- far more commonly now -- by being directly certified: automatically qualified because they, or someone in their household, already receives , , or Medicaid, with no separate income check required. Direct certification covered 46% of free-lunch students in 2020-21; by 2023-24 it covered 79%, a jump the audit traces almost entirely to one policy change -- Kansas started auto-qualifying Medicaid recipients for free lunch starting in the 2022-23 school year. Because Medicaid's income ceiling runs higher than the free-lunch threshold in many cases, and because auditors could not verify whether the state agencies granting , , and Medicaid themselves made accurate eligibility calls, the free-lunch count grew on a foundation nobody in this audit was able to check.
View data as table
| Census-estimated income-eligible, 2023 | 92,000 | 20% of students |
|---|---|---|
| Free-lunch count, 2023-24 | 198,000 | 41% of students |
Where auditors could check, most of what they checked failed
Auditors couldn't verify direct certification -- that depends on other agencies' own records -- but they could verify the smaller share of students who qualify by submitting an application or income survey, because those documents list a household's reported income directly. They pulled income records from the Kansas Department of Revenue and Department of Labor for a random, statistically projectable sample and compared it to what families had reported. Of 108 application-based students they could locate in those systems, 68 -- 63% -- reported income above the eligibility threshold; two-thirds of those were at least 50% over the line, and about half of that group had household incomes above $100,000, including two above $200,000. Projected statewide, auditors estimate 54% to 72% of the roughly 34,000 students who qualified by application were likely ineligible -- 18,400 to 24,600 students. A parallel spot-check of income surveys at schools with universal free lunch found a similar pattern: 9 of the 16 verifiable surveys, 56%, also looked ineligible.
View data as table
| Eligible | 40 | 37% of 108 verified |
|---|---|---|
| Ineligible | 68 | 63% of 108 verified |
| -- within 10% of threshold | 7 | 10% of the 68 |
| -- 10%-50% over threshold | 16 | 24% of the 68 |
| -- 50%+ over threshold | 45 | 66% of the 68; about half had income over $100,000, two over $200,000 |
Why the state can't catch it
This isn't a story about families lying en masse; it's a story about a verification system built not to catch them. Federal rules require districts to accept an application's self-reported income at face value, bar them from cross-checking it against tax or wage records, and cap formal verification at 3% of approved applications -- in 2023-24, 80% of Kansas's 286 districts had to verify five applications or fewer. A 2019 GAO report⧉ the Kansas audit cites found never built a process to assess fraud risk in the school lunch program at all, despite historically high error rates the Kansas audit describes as evidence the program may be 'inherently vulnerable to fraud.' Layer the district-level gap on top of the unverifiable 79% who are directly certified through other agencies, and the audit's conclusion is that no one -- not the district, not the state education department, not -- can currently vouch for the number that decides who gets a free lunch and how much at-risk money follows them.
View data as table
| 2020-21 | 77,000 | 46% of free-lunch students |
|---|---|---|
| 2023-24 | 157,000 | 79% of free-lunch students |
The Legislature's move, and KSDE's pushback
The audit's one formal recommendation goes to the people who built the formula, not the agency that runs it: because the free-lunch count may no longer accurately measure which students are at risk of academic failure, the Legislature should reconsider whether at-risk funding should keep being allocated on that basis at all. It set no deadline and named no agency to fix the underlying verification gap -- that decision now sits with lawmakers, not KSDE. The Kansas Department of Education agreed with the audit's findings in its formal written response, but pushed back on the framing: free-lunch eligibility is measured at a single point in time (September 20) while the audit's income checks used full-year tax records, department officials noted, and state law never defined poverty as the at-risk qualifier in the first place -- foster students, for instance, are automatically at-risk by statute regardless of a family's income. Both of those points are about definitions; neither one explains away the $38 million to $53 million gap between what the state paid and what the audit's own sampling says it should have paid.
- Kansas's 2023-24 free-lunch count (198,000 students) is more than double the roughly 92,000 the audit estimates were actually in poverty, using the same year's Census data -- a gap the audit's own comparison surfaces for the first time.
- Of the applications auditors could actually verify against tax and wage records, 63% (68 of 108) looked ineligible, projecting to an estimated 18,400-24,600 ineligible students statewide -- while the other 79% of free-lunch students, who are auto-qualified through other agencies, could not be checked at all.
- The likely result: $38 million to $53 million in state at-risk funding and $10 million to $14 million in federal meal reimbursements overpaid in 2023-24 alone, against $488 million in total at-risk funding paid that year.
- The audit's fix goes to the Legislature, not KSDE, and carries no deadline -- reconsider whether at-risk funding should keep being tied to the free-lunch count at all; KSDE agreed with the findings but disputed how directly they translate into overpayment.
The $38-53 million state figure and the $10-14 million federal figure are two separate estimates from two different formulas -- state at-risk funding (about $2,500 per qualifying student) and federal meal reimbursement (about $4.25-$4.35 per lunch served) -- both drawn from the same inflated 2023-24 free-lunch count but calculated independently by the audit. This piece keeps them distinct rather than reporting a single combined total, matching how the audit itself presents them; a simple sum of the two ranges ($48-67 million) is shown in the underlying analysis data as order-of-magnitude context only, not an audit-stated figure.
The audit's ineligibility estimates (54%-72% of application-based qualifiers) are projectable only to the 16% of free-lunch students who qualified by submitting an application or survey -- about 34,000 of the 198,000 total. For the 79% who were directly certified through , , Medicaid, or foster-care status, auditors verified that paperwork existed (99% of a 615-student sample had supporting documentation) but could not independently confirm that the underlying , , or Medicaid eligibility determinations made by other Kansas agencies were themselves correct -- that question was outside this audit's scope. The true statewide ineligibility rate, across the full free-lunch population, remains unknown.
Sources(2) ▾
- Kansas Legislative Division of Post Audit, Reviewing Free Lunch Student Counts Used as the Basis for At-Risk Funding (Report R-25-013) (2025-07-01) — A performance audit presented to the Legislative Post Audit Committee, published July 2025 (Report Number R-25-013; the cover page states only the month, not an exact day -- 'date' above is set to the 1st of that month as a placeholder for sorting). The Committee authorized the audit at its April 24, 2024 meeting; the topic was suggested by LPA staff. Audit objective: does the number of free lunch students used for at-risk funding accurately reflect the number of students eligible for the program? Auditors reviewed documentation for a random, projectable sample of 769 (of about 198,000) students who qualified for a free lunch in the 2023-24 school year, cross-checking income against Kansas Department of Revenue tax records and Kansas Department of Labor wage records. Read directly from the publisher's PDF via page-by-page extraction after direct download (no login required); charts (a pie chart on p.9 and a scatter plot on p.11) were additionally read from the rendered page images to confirm the numeric labels. kslpa.gov · original document
- U.S. Government Accountability Office, School Meals Programs: USDA Has Reported Taking Some Steps to Reduce Improper Payments but Should Comprehensively Assess Fraud Risks (GAO-19-389) (2019-06-04) — Cited in the Kansas audit's Appendix A (p.24) as evidence the school lunch program's self-reported, largely unverified income model carries an inherent fraud-risk gap. This piece links directly to 's product page for the report rather than treating the Kansas audit's characterization of it as independently verified wording -- the gao-fraud-risk-citation fact below is sourced to the Kansas audit's own text (doc-kslpa-r25-013), which is where the 'inherently vulnerable to fraud' phrase in quotation marks appears; this office could not retrieve the underlying PDF directly (gao.gov blocks direct download) to confirm that exact phrasing against 's original text. gao.gov · original document
Comments
Always open. Logged-in readers can annotate paragraphs in place.
The Kansas Legislative Division of Post Audit⧉ -- the nonpartisan legislative-branch office lawmakers use to check whether state programs work as intended -- reviewed whether the state's free-lunch student count still measures what it's supposed to measure. In Kansas, that count does double duty: it determines which students get a free school lunch under the federal National School Lunch Program, and it is the near-total basis for at-risk funding, the roughly $488 million a year the state gives school districts to serve students at risk of falling behind academically.