Are 75% of Walmart employees on welfare?
Executive summary
No — the claim that 75% of Walmart employees are “on welfare” is not supported by the available reporting; federal and media sources show thousands of Walmart workers received SNAP or Medicaid in sample states, but those counts are far from proving three‑quarters of the company’s workforce rely on public benefits and the data cited are limited in scope [1] [2]. The measurement most often referenced is counts of beneficiaries by employer in selected states, not a nationwide percentage of all Walmart employees enrolled in public programs [1] [2].
1. What the core data actually say, and what they don’t
A Government Accountability Office analysis commissioned in 2020 produced employer‑level counts of SNAP and Medicaid recipients in a subset of states, and Walmart appeared among the top four employers of program beneficiaries in every surveyed jurisdiction, with figures such as roughly 14,541 Walmart employees on SNAP across nine states and about 10,350 on Medicaid in six states cited by advocates and reports [1] [2]. Those raw counts show Walmart employs many people who use public assistance, but the GAO data reported employer rankings and case counts in selected states rather than a nationally representative share of Walmart’s total workforce, so they cannot be extrapolated to support a 75% national prevalence claim [1] [2].
2. Why “75%” is a red flag: mismatched numerators and denominators
The jump from state‑level beneficiary counts to a sweeping 75% figure requires a nationwide denominator (total number of Walmart employees) and a validated numerator (how many of those employees are enrolled in public programs); the sources provided do not supply that nationwide crosswalk. Media and advocacy pieces emphasize high counts and proportionally prominent employer rankings, but those items do not equate to three‑quarters of all Walmart workers being on government assistance [2] [3]. Even politically charged summaries that highlight millions in taxpayer subsidies to low‑wage workers rely on modeled estimates and limited samples rather than a direct, company‑wide welfare rate [4] [1].
3. How advocates and critics use the findings
Progressive advocates and critics of low wages point to the GAO numbers to argue that large employers effectively shift labor costs onto taxpayers — a contention reflected in Sen. Bernie Sanders’ framing and in reports that estimate billions in public assistance tied to low‑paid workers [1] [4]. Conversely, Walmart and some media outlets stress context: Walmart is the nation’s largest private employer, so absolute counts of employees on benefits will be large even if only a small share of its workforce participates, and company spokespeople describe beneficiaries as a “small percentage” of employees [5] [2] [6]. Those competing frames reflect different agendas: systemic wage critique versus corporate defense emphasizing employment access.
4. Additional important patterns in the reporting
The GAO and follow‑ups also reveal that a large portion of SNAP and Medicaid recipients are employed full‑time and work in private‑sector firms, which bolsters the argument that many working households rely on public assistance despite regular employment [7] [2]. Reporting underscores that retail and food service firms — not Walmart alone — are prominent employers of beneficiaries, with chains such as McDonald’s, Dollar Tree, Amazon and others repeatedly named in state lists [1] [2].
5. Bottom line and limits of current evidence
There is no credible evidence in the cited materials to support the assertion that 75% of Walmart employees are on welfare; available GAO counts show thousands of Walmart workers enrolled in SNAP or Medicaid in sampled states but do not provide a nationwide welfare rate for the company, and Walmart’s own spokespersons dispute the implications by stressing relative scale and employment access [1] [5]. Because the studies and media reports rely on partial state data, headline extrapolations to a 75% figure are unsubstantiated given the documented limits of the data [1] [2].