Economic Impact of Undocumented Immigration in the US

The economic impact of undocumented immigration in the US includes workers’ production, consumer purchases and tax payments, alongside public-service costs and pressure on particular labor and housing markets. Those effects fall on different people and different levels of government. A larger economy does not automatically mean higher income for every resident, and taxes paid are not the same as a net fiscal gain.

Research checked November 3, 2026. This guide compares dated research rather than presenting a single current-dollar “cost” or “benefit.” It distinguishes undocumented-population estimates from studies of all immigrants or a recent immigration surge, and observed evidence from policy simulations.

Economic impact: compare the population and year before the number

The following findings answer different questions. Read across each row before comparing down the column; adding or subtracting these figures would not produce a meaningful national balance.

What does “undocumented immigration” include?

Undocumented, unauthorized and illegal immigration are common search terms, but a study’s classification determines what its numbers mean. Some demographic definitions include people with temporary protections or pending proceedings. Work permission, a pending case and permanent immigration status are separate concepts.

CBO’s immigration-surge category includes people permitted to enter or remain and people without that permission. Its federal analysis measures the additional surge relative to a counterfactual, not the economic contribution of every undocumented resident. Our US immigration pathways and agencies guide explains the wider categories; an economic research label cannot determine an individual’s legal situation.

For population comparisons and estimation methods, see our guide to US unauthorized immigrant population estimates. Do not combine one institution’s population estimate with another institution’s average tax figure unless their definitions, reference years and methods are compatible.

Undocumented immigrants, GDP and consumer spending

The Bureau of Economic Analysis measures GDP as a measure of domestic production. To understand immigration and economic growth, separate the size of that production from the income and purchasing power of particular people.

A useful economic analysis asks three questions: how production changes, how income is distributed, and what happens to government finances. A dollar of business sales, a dollar of workers’ earnings and a dollar of tax revenue cannot be added as three independent contributions to GDP.

Consider a hypothetical neighborhood restaurant. Additional workers might let it serve more customers; those workers also buy groceries and pay rent. The owner, customers, workers and suppliers can experience different changes. That example illustrates possible channels, not a measured result or an assumption about anyone’s status.

Total GDP and GDP per capita answer different questions. An economy can produce more while output per resident changes little or falls. Average wages can also change because the workforce’s composition changes, even if the same existing workers do not receive a pay cut. Compare outcomes for incumbent residents separately from outcomes for new arrivals.

Brookings’ March 2026 analysis discusses labor supply, consumer demand and population aging across immigration more broadly. Those channels help explain the subject, but its all-immigrant findings should not be relabeled an undocumented-only GDP estimate. There is no undocumented-only GDP number established by the evidence table above.

Jobs and wages: who gains, who faces competition?

Pew’s 2023 industry estimates put unauthorized workers at 15% of construction employment, 14% of agriculture and 8% of leisure and hospitality. Industry shares differ from occupation shares; neither tells you an individual worker’s status. The figures describe an earlier labor market and should not be presented as current hiring totals.

The National Academies’ 2017 consensus report found small overall long-term effects of immigration on native-born wages and employment, with possible negative effects concentrated among prior immigrants or native-born workers without a high-school diploma. That is a broad immigration finding, not proof that every local worker is unaffected by undocumented immigration.

A newer, narrower study is the Dallas Fed’s March 23, 2026 working paper. Using administrative microdata and local-market variation, it finds approximately one additional local job per additional unauthorized immigrant worker during the boom episode, without a significant local wage decline. It also reports lower labor income per capita, consistent with a change in workforce composition.

This is a working paper, not an official Federal Reserve policy conclusion. Its setting and identification assumptions matter. “No statistically significant decline” does not prove an exactly zero effect, and the result should not be extrapolated automatically to every industry, later year or enforcement policy.

When evaluating an “immigrants take jobs” or “immigration raises wages” claim, ask which workers are being compared. Existing workers who perform similar tasks may face different pressures from workers whose tasks complement newcomers. Employment, hours, hourly wages and annual earnings are separate outcomes. A national average cannot settle every local question.

Do undocumented immigrants pay taxes?

Yes. ITEP’s tax-incidence model estimates that undocumented immigrants paid $59.4 billion federally and $37.3 billion to state and local governments in 2022. Its total includes estimated income, payroll, consumption and property-tax burdens. Property taxes can be paid directly by owners or borne indirectly through rent.

ITEP is a research organization, not the IRS. Its estimate relies on population classification, incomes and tax-compliance assumptions; it is not an administrative count of every taxpayer’s immigration status. The study estimates $25.7 billion in Social Security taxes and $6.4 billion in Medicare taxes for the same year. These contributions do not establish an individual’s benefit eligibility.

The IRS’s ITIN guidance explains that people who need a federal taxpayer identification number but are ineligible for an SSN can use an Individual Taxpayer Identification Number. An ITIN serves a tax purpose; it does not authorize work, confer immigration status or entitle its holder to Social Security benefits.

Do not count all ITIN users as undocumented immigrants. Tax identification and immigration categories are different, and a tax estimate is not advice about a person’s filing obligations or eligibility. Equally, the absence of work authorization does not mean that purchases or payroll produce no tax revenue.

Why federal and state/local fiscal effects can differ

Fiscal impact means changes in public revenue and public spending. It is narrower than the entire economic impact. A useful comparison keeps the population, government level, accounting period and included programs consistent.

In its July 2024 scenario, CBO projected $1.2 trillion in additional federal revenue and $0.3 trillion in additional mandatory spending and net interest over 2024–2034. Its headline deficit reduction excludes a detailed estimate of discretionary funding changes and does not include state/local budgets. It is a dated projection under the report’s assumptions, not a measurement of realized savings through 2026.

The separate CBO state/local report estimated $10.1 billion in direct additional revenues and $19.3 billion in direct spending in 2023. The direct net cost equaled about 0.3% of state/local spending excluding federal grants. Education, shelter-related services and border security were major components. An alternative measure including broader revenue effects and nonbudgetary service pressures produced a potential $9.8 billion net cost.

A federal revenue gain can coexist with a city’s short-term service costs. Ask who receives the revenue, who funds the service, whether transfers compensate the locality and whether the analysis measures one year or a generation. Neither result automatically disproves the other.

Do not subtract ITEP’s 2022 tax estimate from CBO’s 2023 local cost, or compare either directly with an eleven-year federal projection. That would mix populations, years and budget boundaries, creating a number none of the researchers estimated.

Housing, schools and health services need separate treatment

The Dallas Fed working paper also finds higher local house prices and rents without a corresponding short-run housing-supply expansion. This is an important distributional issue: renters and property owners can experience the same market change differently. Its finding concerns the studied locations and episode, not a claim that immigration explains all US housing inflation.

For a local housing comparison, ask whether the study tracks new construction, vacancy, financing costs and other demand changes. A correlation between population growth and rent does not by itself isolate the effect of immigration status.

For schools or health spending, distinguish an undocumented resident’s costs from those of an entire mixed-status household. Ask whether US-citizen children are included, whether the estimate allocates average or marginal costs, and whether federal reimbursements have been deducted. These are accounting questions; they do not make the services costless or establish a person’s eligibility.

HealthCare.gov says undocumented immigrants cannot obtain federal Marketplace coverage for themselves, while they may apply for eligible documented family members. State-funded coverage and emergency-care arrangements require separate checks. Our health insurance guide for undocumented immigrants explains those coverage distinctions; a national economic estimate cannot substitute for an enrollment decision.

Mass deportation and work authorization: scenarios are not outcomes

The Penn Wharton Budget Model’s July 25, 2025 analysis compares a four-year removal policy with a permanent ten-year policy. It projects 2034 GDP below baseline by 1.0% and 3.3%, respectively. It also estimates different wage outcomes across skill groups, including potential gains for authorized lower-skilled workers under sustained removal and losses for higher-skilled workers.

Those are model results conditional on removal rates, future immigration, investment and fiscal assumptions. They are not observed changes attributable to an enacted policy. The analysis’s broader budget calculation includes enforcement outlays and revenue losses; quoting a GDP change alone does not describe the full scenario.

Work-authorization proposals need the same discipline. Ask which residents qualify, how earnings and tax compliance are assumed to change, what happens to program eligibility, and which administrative costs are included. An assumed percentage wage gain is a model input or estimate, not an automatic pay increase for every person.

Economic scenarios can inform a policy discussion without settling legal, humanitarian or enforcement questions. Compare the full alternatives rather than treating the removal of a worker as the disappearance of only wages, or treating authorization as a change with no fiscal or administrative consequences.

How to compare claims about immigration costs and benefits

An older Migration Policy Institute report from December 2009 illustrates how long these debates have existed. Historical research can explain an argument, but its conclusions are not a current estimate simply because the page still ranks in search. Publication date, data year and policy setting all matter.

Use this checklist when a headline offers a single “net cost” or “economic benefit”:

1. Identify the people: all immigrants, undocumented residents, recent arrivals or their households?
2. Record both dates: publication and the year or years measured.
3. Name the outcome: taxes, public spending, GDP, wages, rents or household purchasing power?
4. Choose the boundary: federal, state/local, all governments or private economic activity?
5. Check the comparison: observed change, estimated causal effect or a model relative to a hypothetical baseline?
6. Read the distribution: which workers, consumers, businesses, generations and places gain or lose?
7. Inspect uncertainty: population adjustments, compliance assumptions, housing response and alternative specifications.
8. Avoid double counting: sales, earnings, tax revenue and transfers can overlap rather than represent separate new production.

For notes or a school assignment, save one evidence row: source → population → data period → metric → comparison → included costs → limits. Leave a field unresolved when the source does not specify it.

For example, a hypothetical claim that a population “paid $10 billion, so it generated a $10 billion government profit” fails at the included-costs field. A second claim that a city faced a service shortfall therefore national GDP fell fails at the metric and government-boundary fields. Neither calculation becomes valid by attaching a precise dollar sign.

The most useful answer is therefore a dated, clearly bounded finding. Keep production, distribution and public finance visible together, and leave room for uncertainty where the evidence does not establish a single national result.

Frequently asked questions

Do undocumented immigrants contribute to the US economy?

They contribute through work, purchases and taxes. A complete assessment also examines public-service costs, labor-market distribution and housing pressures. Those effects vary by place, period and the population a study measures.

Do undocumented immigrants pay more in taxes than they receive in benefits?

A gross tax estimate alone cannot answer that. A valid net fiscal comparison must use compatible populations, years and government budgets, specify included services and explain how mixed-status households are treated.

Does immigration lower wages for American workers?

Research does not support one effect for every worker. Overall averages, competing and complementary occupations, earlier immigrants, education groups and local conditions can produce different results. Broad immigration research is not automatically undocumented-only evidence.

Does a higher GDP mean everyone is better off?

No. Total output, output per resident, wages, prices and after-tax household income are different measures. A larger economy can distribute gains unevenly, so check whose outcome the research measures.

Are deportation cost estimates actual government spending?

Not necessarily. Many published figures are modeled scenarios that assume removal rates, enforcement costs and economic responses. Check the report’s baseline and time horizon before treating a projection as an observed expense.