Immigrant public benefit participation: Our estimates

Introduction

Multiple efforts are underway to restrict safety net access for immigrants, often based on rhetorical arguments that immigrants, non-citizens, or undocumented immigrants overuse benefits. How much do immigrants participate in public benefit programs, and how does their participation compare to that of U.S.-born citizens? These questions have high political relevance; their answers shape public perception of immigration and government assistance, affect estimates of the fiscal impact of immigration, and inform ongoing policy debates concerning immigrant access to these programs. Yet, there is no clear consensus among researchers on how to measure such participation.

As a result, researchers tend to describe immigrant benefit use relative to that of U.S.-born citizens quite variably—some finding that immigrants use more benefits, others that they use less. In this report’s companion methodological guide, we define the key choices researchers face when constructing these estimates and recommend our preferred approach. Here, we present the numbers produced by that method. The estimates produced in this piece reflect program participation prior to the implementation of recent legislative and executive actions aimed at restricting program access. Recent and pending changes will almost certainly reduce non-citizen participation in the safety net further.

Our analysis examines participation in the following programs: the Supplemental Nutrition Assistance Program (SNAP, formerly the Food Stamp Program), Temporary Assistance for Needy Families (TANF), Supplemental Security Income (SSI), federal housing assistance, Medicaid, the Women, Infants, and Children (WIC) nutrition program, the free or reduced-price school lunch program, and the Earned Income Tax Credit (EITC).

Eligibility for these programs varies by immigration status. Unauthorized immigrants are generally ineligible for SNAP, TANF, SSI, federal housing assistance, and Medicaid, though some states use their own funds to extend support to federally ineligible groups. Lawfully present non-citizens are divided into “qualified” and “non-qualified” categories under the Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) of 1996.¹ Most “qualified” immigrants such as lawful permanent residents are subject to a five-year ban before becoming eligible for federally funded means-tested public benefits programs, though some exceptions for refugees, asylees, and other humanitarian entrants apply. Other programs, including the WIC nutrition program and the school lunch program are available regardless of immigration status in almost all states.² The Earned Income Tax Credit (EITC) is only available if all filers have a work-authorized Social Security number.

Our estimation approach uses nationally representative survey data and adjusts for factors like under-reporting, take-up rates, and program eligibility rules. Prior analyses have typically measured receipt at only the individual level or only the household level across all programs. This can be misleading because it obscures intra-household dynamics—it may count a citizen child’s receipt of Medicaid as receipt by a non-citizen parent, or ignore that a program like SNAP is intended for and administered to the household group. Our method distinguishes the unit of analysis on a program-by-program basis. For programs that are not administered at the individual level, we take particular care to distinguish between receipt that would exist regardless of the non-citizen’s presence and receipt that is higher due to the non-citizen’s presence.

These choices that went into our methodological approach are discussed more fully in our companion piece.

We find that, after isolating the benefits attributable to non-citizens themselves rather than the citizens in their households, non-citizens use public benefits at lower rates than citizens across nearly every major program for which we have data, with free or reduced-price school meals being the main exception. When we instead count all benefits received by households containing non-citizens—including those flowing to their citizen members—participation often matches or exceeds that of all-citizen households (a difference that reflects benefits reaching citizens, most often U.S.-born children, rather than non-citizens).

We reproduce this analysis across a variety of restricted samples, including breakdowns by income distribution, household composition, length of U.S. residence, and age group. Consistently, immigrants tend to use benefits at lower rates than citizens, despite being more concentrated at the bottom of the income distribution, where these programs are targeted. In a limited number of cases—school meals among children, and Medicaid among non-citizens 65 and older—individual non-citizen usage is comparable to or higher than that of citizens.

The persistence of lower benefit use among low-income non-citizens suggests that existing eligibility restrictions meaningfully limit program access for immigrants. Many non-citizens are excluded not because they fall outside the income range these programs target, but because of restrictions tied specifically to immigration status—restrictions that, as discussed below, will expand further under recent policy changes.

Our estimates

We use the 2023 American Community Survey (ACS) 1-year estimates, produced annually by the U.S. Census Bureau, to construct statistics on the usage rates of certain benefits programs across groups. We use the 2023 survey due to data availability constraints in more recent datasets.³ The 2023 ACS includes a representative sample of approximately 3,400,000 individuals in approximately 1,500,000 households, corresponding to about 1% of the population and including unauthorized immigrants. Figure 1 shows the weighted national totals from the 2023 ACS sample, broken down by citizenship and nativity at both the individual and household levels.

Our analysis draws on two types of program participation data. For SNAP, TANF, SSI, Medicaid, Social Security, and Medicare, we use self-reported survey responses—respondents report whether they or any other household member received the benefit currently or at some point in the past 12 months. For EITC, housing vouchers, WIC, and free or reduced-price school meals, we use Census Bureau-modeled estimates from the Supplemental Poverty Measure (SPM) research files, which are imputed from household characteristics and income, rather than self-reported. Census SPM imputations do not consider immigrant eligibility restrictions.

In Table 1, we present our estimates of non-citizen versus citizen usage of safety net programs we consider reasonable to include under the “public benefit” umbrella.

We distinguish between household-level and individual-level programs. If eligibility and benefits are determined for and provided to a single person, we measure the program at the individual level. If eligibility and benefits are determined for a household, family unit, or a defined subset of a household such as an assistance unit, we measure the program at the household level.

For individual-level programs, we show participation rates for citizens and non-citizens. For household-level programs, we report three rates. The first is the participation rate for all-citizen households. The remaining two columns report rates for households with at least one non-citizen member (these households often also include citizen members). The second column reports the share of these households that receive a benefit through any member, which includes benefits received by citizen members and non-citizen members. The third column reports the share of these households whose benefit receipt we determine is specifically attributable to the non-citizen member; this third column is effectively a subset of the second. More detail on how we determine non-citizen attribution is provided in Table 2 below.

Non-citizens use SSI and Medicaid at lower rates than citizens and receive free or reduced-price school lunch at a slightly higher rate—likely reflecting both the absence of eligibility restrictions for school lunch and lower family incomes among the non-citizen population.

Households with at least one non-citizen show higher usage than all-citizen households for SNAP and TANF; usage rates for housing vouchers are roughly equivalent between the two groups. Notably, the higher usage rates largely reflect benefit receipt by citizen members of mixed-citizenship households. When we count only receipt attributable to non-citizen members, the relationship inverts, with non-citizen households showing lower rates than all-citizen households across all four benefits, including EITC.

Measured over individuals rather than households (see Appendix Table 1), the same household programs show non-citizens using every program at lower rates than citizens. This confirms that the higher household-level participation in SNAP and TANF in households with at least one non-citizen reflects program participation by the citizen members of these households (since the non-citizen individuals within those households participate at consistently lower rates).

Our method

Our method relies on applying program-specific eligibility rules and adjusting for over-imputation and under-reporting to ensure that estimates reflect actual receipt. For more information on the decisions that went into determining our approach, see our methodological guide.

Table 2 illustrates how we count non-citizen participation for programs that are not measured at the individual level. We differentiate between receipt by any member of the household and receipt by the non-citizen member. For each program, this requires considering whether the non-citizen in the participating household is both eligible for the program (and therefore adding to the amount of the benefit that unit receives) and part of the specific assistance unit or subset of the household to which the program is targeted.

To identify which non-citizens are likely ineligible for status-restricted programs, we use a probabilistic model that scores each adult non-citizen based on characteristics observable in the ACS. Non-citizen children inherit the classification of adult non-citizens in their household. For programs where the benefit accrues to a defined sub-unit rather than the whole household, we additionally apply program-specific rules to determine whether the non-citizen specifically is the relevant recipient.

The process for assigning non-citizen receipt varies by program. Consider a low-income household with one unauthorized parent, one citizen parent, and one citizen child. This household could be eligible for and receive SNAP benefits. In determining the benefit, some or all of the unauthorized parent’s income would be counted, but the maximum benefit would be calculated based on only the two eligible members. If the unauthorized parent has any income, the household will necessarily receive less in SNAP benefits than it would if that parent were not present. This is because the parent’s income counts against the benefit calculation, while the parent themself is excluded from the household size used to set the maximum benefit—so the unauthorized non-citizen’s presence can only lower the benefit, never raise it. We therefore do not count that household in the “receipt attributed to non-citizen member” column. The same logic applies to housing vouchers, which have similar eligibility restrictions: the household is not disqualified from receipt if it includes an ineligible member, but that member does not add to (and may subtract from) the benefit size.⁴

Although states have the option to impose immigration-related eligibility restrictions on WIC, most do not; we do not adjust for likely unauthorized status when counting non-citizen WIC use. In the case of EITC, the entire tax filing unit is ineligible for benefits if one member lacks a valid work-authorized Social Security Number, so if we deem one member of the household to be likely ineligible, we do not count that household as a recipient in the second or third column estimates shown in Table 1.

TANF benefits are not provided to every person living in a household—instead, they are calculated and distributed to a specific assistance unit within it. This is typically the parent and any dependent children, excluding any household member who does not qualify for assistance. Commonly, TANF cases are considered “child-only” if the parent is ineligible due to immigration status. We count a TANF household in the “receipt attributed to non-citizen member” column only if the assistance unit itself includes an eligible non-citizen member.

For programs measured through self-reported survey questions, we apply corrections for the tendency of survey respondents to under-report program participation relative to administrative records. Programs measured using Census Bureau-modeled SPM estimates do not require this correction. Instead, we adjust estimates using program–specific take-up rates to account for eligible individuals who do not participate in the program.

Table 3 summarizes all adjustments by program and sample grouping.

Age and income profiles

Figure 2 shows the age distribution of U.S.-born citizens, naturalized citizens, and non-citizens. Non-citizens are more concentrated in working-age years than native-born citizens. Naturalized citizens are the oldest of the three groups, with over a quarter aged 65 or above and very few under 18. These differences matter for interpreting program usage: a larger share of U.S.-born citizens are children or elderly, making them more likely to qualify for age-based programs like school meals, and for entitlement programs such as Social Security and Medicare, where eligibility is tied to work history and immigration-related restrictions apply.

Figure 3 shows the distribution of household income as a share of the federal poverty line for U.S.-born citizens, naturalized citizens, and non-citizens. Non-citizens are substantially more concentrated at the bottom of the income distribution. About 17% of non-citizens fall below the poverty line, compared to roughly 11% of native-born citizens and 10% of naturalized citizens. The gap persists well above the poverty threshold—non-citizens are overrepresented at nearly every income level up to about 300% of the poverty line. Means-tested programs typically target households below these income ranges. Naturalized citizens are the group most likely to be in the high-income range.

Additional samples 

Aggregate comparisons between citizens and non-citizens can obscure as much as they reveal. As Figures 2 and 3 show, the two populations differ systematically along age and income dimensions that themselves affect benefit eligibility and participation. A single headline rate folds these compositional differences together with any underlying difference in how likely otherwise-similar people are to use benefits. Restricting the sample—by income, household composition, length of residence, and age—allows for more of an apples-to-apples comparison. Which restriction is most informative depends on the question being asked. Below, we present estimates across a range of subsamples, each offering a different view of immigrant benefit participation.

Table 4 restricts the sample to households and individuals below the federal poverty line—a group whose income would generally satisfy program eligibility thresholds. The participation rate attributable to non-citizens falls below the all-citizen-household rate for every household program, and at the individual level, non-citizens use SSI and Medicaid at far lower rates. The exception is free or reduced-price school lunch, where non-citizen children participate at a slightly higher rate, consistent with the absence of immigration-status restrictions on that program. In the “households with at least one non-citizen member” columns, “receipt by any household member” comes much closer to the all-citizen rates.

Table 5 restricts the sample to households with children. Here, the two measures of non-citizen household participation diverge sharply. On a “receipt by any household member” basis, households with at least one non-citizen member participate in SNAP, TANF, and housing vouchers at higher rates than all-citizen households. This likely reflects the family structure of the households in this sample: most children in non-citizen households are U.S.-born citizens, and the benefits flowing to these eligible children lift the household’s “receipt by any household member” rate. The third column rates, which isolate receipt attributable to the non-citizen members themselves, fall below the all-citizen household rate for every program. The gap between the two measures captures household benefit use that runs through citizen members rather than directly to the non-citizens. EITC is the exception to this pattern: households with at least one non-citizen claim it at roughly half the all-citizen rate because the credit attaches to the tax unit and a single member without a valid work-authorized SSN disqualifies the whole unit.

Table 6 breaks the results down by how long foreign-born members have lived in the U.S., splitting non-citizens into recent arrivals (five years or fewer) and longer-tenured residents (six or more). Households are assigned to the longer-tenured group if any non-citizen member has lived in the U.S. for more than five years, and to the recent-arrival group only if all non-citizen members arrived within the past five years.

Comparing the two tenure groups, longer residence is associated with greater non-citizen participation in several programs. Among the individual programs, SSI use and school meal participation are both higher for longer-tenured residents, while Medicaid is similar across the two groups and WIC participation declines slightly with years of residence. On the household side, SNAP and EITC participation are higher among longer-tenured households, whereas TANF and housing vouchers are roughly stable.

Medicaid is a notable exception to the tenure pattern: rather than being lower among recent arrivals, as the five-year bar facing most qualified immigrants would suggest, it sits close to parity for both arrival groups. Likely explanations are that recent arrivals include a substantial share of refugees and asylees, who are exempt from PRWORA’s eligibility restrictions, and that most states extend coverage during the five-year wait to lawfully residing immigrant children and pregnant women.⁵

Table 7 breaks individual-level programs down by age group and adds Social Security and Medicare alongside the means-tested programs. We consider Social Security and Medicare to be distinct from the public benefits discussed elsewhere in this piece—they are entitlements earned through work history and payroll contributions, not means-tested transfers (for further discussion, see our methodological guide). SSI rates are shown only for adults: the ACS does not allow us to identify SSI receipt for children under 15, so we exclude the under-18 category rather than report a partial figure. Medicaid and Medicare are not mutually exclusive—individuals can be enrolled in both, so the two rows should not be read as competing alternatives.

Among those 65 and older, non-citizens enroll in Social Security and Medicare at far lower rates than citizens, while their Medicaid rate is more than double the citizen rate. One likely reason is that some older non-citizens with shorter histories in the U.S. have not accumulated enough quarters of covered employment to qualify for Social Security or Medicare, leaving Medicaid as an important source of health coverage in old age.

The policy stakes

As the Trump administration moves to redefine which programs count as federal public benefits and to expand the scope of public charge considerations, reliable estimates of how much immigrants participate in public benefits are pertinent.

In July 2025, multiple federal agencies—including the Department of Health and Human Services and the Department of Education—issued parallel notices reinterpreting what constitutes a “federal public benefit” under PRWORA. These reinterpretations would newly classify programs like Head Start, Federally Qualified Health Centers, and adult education programs as federal public benefits, restricting access for non-qualified immigrants. The notices have since been challenged in court, and as of July 2026, preliminary injunctions have blocked their enforcement in over twenty states.

The Department of the Treasury also announced at the end of 2025 that it intends to reclassify the refundable portions of credits like the EITC, the Child Tax Credit, and the American Opportunity Tax Credit as federal public benefits under PRWORA, which would similarly limit eligibility for many lawfully present non-citizens. The Treasury’s proposed reinterpretation departs from longstanding congressional practice codified across multiple statutes, including the Tax Cuts and Jobs Act of 2017, the CARES Act of 2020, and the One Big Beautiful Bill Act of 2025. These legislative acts consistently distinguish between the limited categories of “eligible aliens” entitled to federal public benefits and the broader population permitted to claim refundable tax credits based on a valid Social Security number.

The administration has also issued a final rule changing the approach to public charge consideration in immigration decisions, which will be in effect on September 18, 2026. If immigrants are determined likely to become a public charge because they have received or are expected to receive certain benefits, immigration officials have grounds to deny their application for adjustment of status. The new rule expands which programs might qualify under the doctrine and places the final decision-making within the subjective discretion of individual immigration officers.

Conclusion

As discussed in our companion piece, there are a variety of methodological decisions required to characterize the benefit use of immigrants, which depend in part on the question being asked. When we isolate benefits attributable to non-citizens themselves, rather than citizens in their households, we find they utilize public benefits at lower rates than citizens. This remains true even after adjusting for factors like income, age, and household composition. While households containing non-citizens often show participation rates equal to or higher than all-citizen households, these benefits are often solely due to the eligibility of citizen members of the household, not the non-citizens.

Political rhetoric commonly characterizes immigrant benefit use as excessive. While overuse is in the eye of the beholder, our preferred estimation method makes it clear that non-citizens do not use safety net benefits disproportionately compared to citizens, and low-income non-citizens are much less likely to participate than low-income citizens. In part, this is because most newly arrived legal immigrants and most unauthorized immigrants are excluded from most major programs. New restrictions will further reduce non-citizen participation relative to these levels by excluding certain categories of “qualified” immigrants from receiving benefits and by expanding the use of immigration restrictions related to public charge.