In Nearly Every State, Immigrant Households Use Welfare at Substantially Higher Rates. The Data Is Real. The Debate About What It Means Is Also Real.

A new analysis by the Center for Immigration Studies, a Washington-based immigration-restriction advocacy group, finds that households headed by non-citizens use one or more means-tested welfare programs at substantially higher rates than households headed by U.S.-born Americans in virtually every state. The analysis draws on Census Bureau data covering 2021 through 2025 and measures program participation across food stamps, Medicaid, WIC, housing assistance, school meals, TANF, and SSI.

The headline numbers are notable: 47 percent of non-citizen-headed households use at least one program, compared with 28 percent of U.S.-born-headed households — a 19-percentage-point gap. In New York, the spread reaches 28 points. Those figures, however, come with a significant methodological caveat that any honest reading of this data requires stating plainly: welfare use is measured at the household level, and 38 percent of non-citizen households contain at least one U.S.-born minor child whose citizen eligibility drives part of the household's benefit use.

The Cato Institute, using the same underlying Census data but measuring consumption per capita rather than per household, reaches the opposite conclusion: immigrants consume 24 percent less in welfare and entitlement benefits than native-born Americans on a per-person basis. Both sets of researchers are working with real numbers. The dispute is entirely about what the right unit of measurement is — and that question has direct implications for how to think about the fiscal costs of immigration.

• 47% — non-citizen-headed households using at least one means-tested program — vs. 28% of U.S.-born households · Source: CIS analysis of Census CPS-ASEC 2021–2025
• 38% — of non-citizen-headed households contain at least one U.S.-born minor child whose citizen eligibility counts toward household welfare use · Source: Washington Times / CIS
• 87.5% — of non-citizen households include at least one worker — higher than the 70% rate for U.S.-born households · Source: CIS / Census CPS
• 24% less — welfare and entitlement consumption per capita among immigrants compared to native-born, per Cato Institute analysis of the same Census data — the competing measurement
• $177 billion — in mandatory federal spending attributable to immigrants from the recent surge and their U.S.-born children projected over 2024–2034 · Source: Congressional Budget Office

The CIS report, authored by researchers Steven A. Camarota and Karen Zeigler, examines five years of the Census Bureau’s Current Population Survey Annual Social and Economic Supplement (CPS-ASEC), combining data from 2021 through 2025 to build state-level sample sizes large enough to be statistically meaningful. The programs counted — SNAP (food stamps), Medicaid, WIC, housing assistance, school meal subsidies, TANF, and SSI — are standard means-tested anti-poverty programs. Social Security and Medicare, which are not means-tested, are excluded.

The central finding is that non-citizen households use at least one of those programs at higher rates than U.S.-born households in virtually every state they examined. The national gap is 19 percentage points — 47 vs. 28 percent. The state spread is wide: New York shows a 28-point gap (61% vs. 33%); Massachusetts is at 55% for non-citizens; Nebraska, at 54% vs. 21%, shows the largest relative disparity in the sample. Arizona (53%), California (54%), and Maryland (50%) all exceed the national non-citizen figure.

Medicaid and food programs drive the largest program-specific gaps. The CIS 2024 SIPP analysis — a companion report using a separate Census survey with more detailed program data — shows immigrant households using Medicaid at 39% compared to 27% for U.S.-born households, and food programs at 35% vs. 22%. The Earned Income Tax Credit, a wage subsidy delivered through the tax code, also shows a gap: 15% of immigrant households vs. 10% of U.S.-born.

Importantly, Camarota and Zeigler note that non-citizen households are not less likely to be working. Fully 87.5% of non-citizen households include at least one employed person, compared with 70% of U.S.-born households. The gap in welfare use therefore does not reflect labor force non-participation; it primarily reflects the income levels and household composition of lower-wage immigrant workers and the eligibility of their U.S.-born children.

Keep Reading

Economy · Tax & Fraud

Banks Flagged $2.5 Billion in Payroll Tax Fraud in 2025. Treasury Says Illegal Hiring Schemes Built It.

Read →

The single most important thing to understand about the household-level finding is this: when a non-citizen heads a household, all benefits received by anyone in that household count toward the household’s welfare “use.” In 38 percent of non-citizen-headed households, at least one member is a U.S.-born minor child — a full American citizen whose Medicaid card or SNAP enrollment counts in this measure as “immigrant household welfare use.”

Federal law generally bars undocumented immigrants from SNAP, non-emergency Medicaid, TANF, SSI, CHIP, and HUD housing programs. Lawful permanent residents face an additional hurdle: a five-year waiting period before becoming eligible for most means-tested programs (with exceptions for refugees, asylees, and trafficking survivors). DACA recipients are not classified as “qualified immigrants” and are ineligible for most federal benefits. The programs that do reach non-citizens directly are limited: emergency Medicaid for stabilizing conditions, WIC for pregnant women and young children regardless of status in most states, and K-12 education.

This eligibility structure means that in many “immigrant households” counted as welfare-using by CIS, the actual benefit recipients are citizen spouses, citizen children, lawful permanent residents past the five-year bar, or refugees with full eligibility — not undocumented non-citizens. The SNAP program enrollment data confirms this: of approximately 1.764 million non-citizens receiving SNAP in FY 2023, the Migration Policy Institute’s Julia Gelatt notes that “noncitizens use SNAP at lower rates than U.S. citizens” once poor-household participation rates are compared directly.

The legal framework for immigrant benefit eligibility is worth stating clearly because it is frequently misunderstood in both directions — overstated by those who claim all immigrants freely collect federal benefits, and understated by those who imply immigration has no fiscal footprint at all.

The Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (PRWORA) remains the baseline. It bars undocumented immigrants from virtually all federal means-tested benefits and imposes a five-year waiting period on most qualified immigrants (green card holders) for SNAP, Medicaid, TANF, CHIP, and SSI. Refugees and asylees are exempt from the five-year bar and receive a fuller benefit package. DACA recipients — roughly 580,000 people — fall outside the “qualified immigrant” category and are ineligible for most federal programs, though several states offer state-funded Medicaid to DACA holders.

The 2025 One Big Beautiful Bill Act tightened restrictions further, adding new limits on SNAP, Medicaid, CHIP, Medicare, and ACA marketplace premium subsidies for non-citizens, with implementation phased in through 2026. The Congressional Research Service’s R47318 and R46510 reports remain the authoritative reference for the current legal landscape.

The practical implication is that when CIS reports high Medicaid use in immigrant households, the Medicaid enrollment often belongs to U.S.-born children, refugee family members with full eligibility, or lawful permanent residents past the five-year bar — not to undocumented immigrants, who are ineligible for non-emergency Medicaid under federal law. States vary: 14 states offer Medicaid to all low-income children regardless of immigration status, and additional states cover pregnant women regardless of status. Those state-funded expansions are real program costs, but they are state choices, not federal defaults.

Keep Reading

Economy · Energy

Albany Found a Villain for Your Electric Bill. It Picked the Wrong One.

Read →

CIS and Cato debate use rates and per-capita consumption, but the Congressional Budget Office provides the most authoritative estimate of the aggregate fiscal picture. In its 2024 report on the effects of the immigration surge, the CBO projected that benefits received by immigrants who arrived in the recent surge — including their U.S.-born children — would account for approximately $177 billion in mandatory federal spending over the 2024–2034 period. The largest components are premium tax credits, other refundable tax credits, and Medicaid; SNAP and other cash programs are a smaller share.

The CBO report also projects that the same surge adds to federal revenues through income taxes, payroll taxes, and excise taxes — and that the net effect on the federal budget over the decade is projected to be a reduction in the deficit, primarily because younger immigrant workers tend to pay more into Social Security and Medicare than they initially draw out. The fiscal picture is not simply “immigrants cost money” or “immigrants pay their way.” It is more granular: near-term benefit costs are real and concentrated in Medicaid and tax credits; longer-term revenue contributions are also real and concentrated in labor income taxes and payroll.

Cato Institute researchers have argued, in their per-capita work, that immigrants on average consume less in means-tested benefits than the U.S.-born and that working-age immigrants contribute substantial payroll taxes before they are old enough to draw Social Security and Medicare. Long-run deficit projections built on that dynamic are real but sensitive to assumptions about future earnings, and a multi-decade horizon is less useful for evaluating current-year program costs — which is exactly where the CIS household figures bite.

“Non-citizen households use one or more means-tested programs at substantially higher rates than the U.S.-born in virtually every state.”

Steven A. Camarota & Karen Zeigler · Center for Immigration Studies · 2026 CPS analysis

One of the more striking individual program findings in the CIS data is the Earned Income Tax Credit. Averaging the 2023 and 2024 SIPP results, CIS finds that 19.1 percent of non-citizen households reported receiving the EITC — nearly double the 10.3 percent rate for U.S.-born households.

The EITC is technically a refundable tax credit, not a traditional welfare program, but it functions as an income supplement for low-wage workers — and by dollar volume it is one of the largest anti-poverty programs in the federal budget, distributing roughly $70 billion annually. Eligibility for the EITC requires earned income and a valid Social Security number for the claimant and any qualifying children. Undocumented immigrants are barred; the recipients are workers with valid SSNs, which primarily means lawful permanent residents, refugees, naturalized citizens, and certain visa holders.

The EITC gap is directionally consistent with the overall CIS finding and is less affected by the mixed-household methodological critique, since EITC is claimed by individual filers with qualifying earnings. The 2025 tax law changes (One Big Beautiful Bill Act) restricted EITC eligibility for undocumented parents even with citizen children by requiring valid SSNs to claim the Additional Child Tax Credit, which will reduce some of this program’s immigrant use rate going forward.

“Most of these households have a worker. The high use rate reflects low wages and the eligibility of their U.S.-born children — not an unwillingness to work. But it still raises real fiscal questions about admitting large numbers of less-educated workers.”

Steven A. Camarota, CIS Director of Research, paraphrasing his public explanation of the findings

The CIS household-use-rate finding is real. More households headed by non-citizens have at least one member using a means-tested program than comparable households headed by U.S.-born Americans. That finding holds across nearly every state, across multiple program types, and across multiple years of Census Bureau data. It is not a fabrication, and the underlying CPS-ASEC and SIPP datasets are public and replicable.

What the finding does not resolve is causation, policy prescription, or net fiscal impact. Non-citizen households are disproportionately low-wage working households — 87.5% have at least one employed person — in industries where wages fall below benefit eligibility thresholds. The same economic conditions that make a U.S.-born working-poor family eligible for Medicaid and SNAP make an immigrant working-poor family eligible for those programs through their citizen children and eligible household members. The program rules work the same way for both groups; the income distribution differs.

The Cato per-capita finding is also real. When you measure dollars consumed per individual rather than the share of households touching any program, immigrants as a group consume less than native-born Americans — because they are younger on average, because they tend not to draw on Social Security and Medicare yet, and because many are individually ineligible for programs that their citizen household members can access. That finding is not a rhetorical escape from the CIS numbers; it reflects a genuinely different way of measuring a genuinely complex fiscal relationship.

The honest accounting requires holding both in view: near-term, household-level welfare program participation among immigrant households is substantially higher. Longer-term, per-capita and lifetime fiscal contributions from immigrant workers are also substantially positive. Whether the balance tips toward fiscal gain or fiscal cost depends on the immigration category (refugees draw more benefits; high-skilled workers contribute more in taxes), the time horizon, and the assumptions about wage growth and program use over a lifetime. None of that complexity should be used to wave away the CIS numbers — and none of the CIS numbers should be used to claim the fiscal case against immigration is settled.

More From Civic Intelligence

Economy · Federal Spending · Entitlements · Social Security

Fiscal Doomsday: $193.6 Trillion in Unfunded Medicare and Social Security Promises

The FY2025 Financial Report of the U.S. Government (released March 19, 2026, signed by Treasury…

Economy · Jobs · Labor Market · BLS

The Numbers We Argue About: What BLS Jobs Data Really Shows

A data-first look at the U.S. labor market through Bureau of Labor Statistics primary data — and the…

Last updated June 13, 2026