The $14 Trillion Question: What Immigration Actually Costs and Creates in the American Economy

The United States is home to 47.8 million foreign-born residents — 14.3 percent of the population, the highest share in over a century [1]. That figure encompasses tech entrepreneurs in Silicon Valley, farmworkers in California's Central Valley, refugees from Afghanistan, H-1B holders at Amazon, and undocumented construction workers in Texas. Treating them as a single economic unit produces misleading conclusions in every direction. The fiscal impact of a software engineer from Hyderabad on an employment visa has almost nothing in common with that of an asylum seeker from Guatemala with a sixth-grade education. Policy debates that ignore this heterogeneity are not serious.

What follows is an attempt to present the strongest available evidence on both sides of a question that has become the defining domestic policy fight of this era: Is immigration making America richer or poorer, and for whom?

Source: U.S. Census Bureau, American Community Survey

Data as of Mar 22, 2026CSV

The Fiscal Ledger: Who Pays and Who Receives

The most rigorous attempt to quantify immigration's fiscal impact remains the National Academies of Sciences 2017 report, which modeled lifetime net contributions by education level and age at arrival. Its central finding: an immigrant arriving before age 25 with a bachelor's degree generates a net fiscal contribution of roughly $334,000 over a lifetime, while one arriving without a high school diploma costs approximately $314,000 [2]. Immigrants with graduate degrees who arrive young reduce the budget deficit by more than $1 million each over their lifetimes [2]. The average new immigrant — legal or unauthorized — produced a modest net positive of roughly $10,000 [3].

At the federal level, the math tilts positive. The Congressional Budget Office estimated that the post-2021 immigration surge would add $1.2 trillion in federal revenue from 2024 to 2034 [4]. Undocumented immigrants alone paid $96.7 billion in federal, state, and local taxes in 2022, according to the Institute on Taxation and Economic Policy [5]. Many of these workers pay into Social Security and Medicare through payroll taxes but are ineligible to collect benefits — a dynamic that effectively subsidizes the programs for native-born retirees.

The picture reverses at the state and local level. The CBO found that the immigration surge imposed net costs on state and local budgets in 2023, driven primarily by public education expenditures averaging over $15,000 per student annually [6]. Immigrant families have more children on average than native-born families, amplifying these costs. The Manhattan Institute's 2025 update confirmed this split: immigration is a net fiscal positive federally but a net negative at the state and local level in the short run [7].

The Welfare Debate: CIS vs. Cato

How much welfare immigrants consume depends entirely on how you measure it, and two organizations have spent years reaching opposite conclusions using the same underlying Census data.

The Center for Immigration Studies, which favors lower immigration levels, reports that 54 percent of non-citizen households use at least one welfare program or qualify for the Earned Income Tax Credit [8]. CIS measures at the household level, meaning if a U.S.-citizen child in a mixed-status household receives Medicaid, the entire household is counted as using welfare. By this measure, immigrant households have higher usage rates than native-born households.

The Cato Institute, which favors more open immigration, measures per capita individual consumption and finds immigrants consume 24 percent less in welfare benefits than native-born Americans [9]. Native-born Americans used an average of $7,134 in old-age entitlements and $3,638 in means-tested benefits in 2023; immigrants used $4,864 and $3,370 respectively [9]. Noncitizens, including undocumented individuals, consumed 53 percent less per capita [9]. A 2026 Cato study found that immigrants reduced federal deficits by $14.5 trillion cumulatively since 1994 [10].

Both methodologies are defensible. CIS captures the real cost that mixed-status households impose on benefit programs — costs that are paid by taxpayers regardless of which family member is the technical recipient. Cato captures the fact that immigrants as individuals draw less from the system than native-born Americans, particularly from the most expensive programs like Social Security and Medicare. The disagreement is fundamentally about unit of analysis, not data fabrication.

Fiscal Impact by Country of Origin

CIS's analysis of Census Current Population Survey data reveals enormous variation in welfare usage by immigrants' country of origin. Non-citizen households from Central America use welfare at 73 percent, the Caribbean at 51 percent, and the Middle East at 51 percent, compared to South Asia at 19 percent and Europe at 34 percent [8]. At the country level: Afghanistan 87 percent, Dominican Republic 78 percent, Guatemala 77 percent, Honduras 75 percent, and Mexico 67 percent — versus Korea 30 percent, the United Kingdom 25 percent, Canada 21 percent, and India 16 percent [8].

This variation is driven primarily by education level and legal status. Immigrants from Central America disproportionately arrived through humanitarian or family channels with low formal education. Immigrants from India and South Korea arrived disproportionately through employment-based visas requiring advanced degrees. The correlation between welfare usage and educational attainment is strong; the correlation with ethnicity per se is weak once education is controlled for [8].

A 2024 IZA study by van de Beek, Hartog, Kreffer, and Roodenburg found strikingly similar patterns in the Netherlands: labor migrants contributed a net positive of more than €100,000 each, while asylum seekers imposed net costs of approximately €400,000 and family migrants approximately €200,000 [11]. The Dutch figures do not translate directly to the U.S. due to differences in welfare state generosity and labor market structure, but they confirm that immigration motive and education — not nationality — are the primary fiscal determinants across Western economies.

Those who argue for more selective immigration policy, such as CIS's Steven Camarota, point to this data as evidence that the current system admits too many low-skilled immigrants whose fiscal costs exceed their contributions [8]. Those who prioritize humanitarian obligations, such as the American Immigration Council, counter that second-generation outcomes improve dramatically — children of low-education immigrants achieve educational attainment and earnings that converge toward native-born averages, meaning the fiscal investment in the first generation pays off over a 30-to-50-year horizon [12].

Wages and Jobs: Borjas vs. Card

No question in immigration economics has generated more heat — or more contradictory findings — than the effect on native-born workers' wages. The debate is organized around two economists: Harvard's George Borjas and Berkeley's David Card.

Source: Bureau of Labor Statistics

Data as of Mar 22, 2026CSV

Borjas's research, which treats immigrants and native-born workers with similar education and experience as direct substitutes in the labor market, found that immigration between 1980 and 2000 reduced wages for native workers without a high school diploma by 7.4 percent [13]. His framework holds capital stock fixed: if labor supply increases by 10 percent and capital does not adjust, average wages should fall by approximately 3 percent. For the least-skilled native workers — those most directly competing with low-education immigrants — the losses were larger, reaching 3 to 5 percent for high school dropouts [13]. Borjas has argued that the benefits of immigration accrue primarily to employers and consumers in the form of cheaper goods and services, while the costs are borne disproportionately by low-wage native workers who compete directly for the same jobs [14].

Card's research reaches the opposite conclusion. His landmark 1990 study of the Mariel boatlift — when 125,000 Cuban immigrants arrived in Miami over several months — found essentially no wage effect on native workers, including those without high school diplomas [15]. Card and Peri (2016) argue that immigrants and native workers are imperfect substitutes: immigrants tend to fill physically demanding or low-status jobs that native workers avoid, while native workers move into roles requiring English fluency and local knowledge [15]. In this framework, immigration is complementary rather than competitive for most native workers. Card's work shows that in practice, capital does adjust to increased labor supply — businesses invest more when labor is abundant, expanding the economy rather than just dividing a fixed pie.

The empirical reality lies somewhere between these poles. The weight of evidence suggests that immigration has small negative effects on wages for the least-educated native workers — those without high school diplomas, roughly 8 percent of the native workforce — and neutral-to-positive effects on all other groups [13] [15]. The Economic Policy Institute estimates that roughly 16 million native-born workers lack a high school diploma or equivalent and could face direct competition from low-education immigrants [16]. For the 92 percent of native workers with at least a high school diploma, immigration effects on wages are negligible or positive.

Sector-Specific Impacts

The concentration of immigrant labor varies dramatically by industry. Immigrants constitute approximately 73 percent of farmworkers, 24 percent of construction workers, and 25 percent of workers in STEM occupations [17]. These concentrations produce different dynamics:

Agriculture: The workforce is overwhelmingly immigrant, much of it unauthorized. Native-born workers have largely exited farm labor regardless of wage levels. The practical question is not whether immigrants displace native farmworkers — there are very few to displace — but whether the availability of low-wage immigrant labor has prevented the mechanization and wage increases that would otherwise occur. The USDA has acknowledged that labor-intensive crops like strawberries and lettuce would face severe harvest disruptions without immigrant workers [17].

Construction: Borjas and others have documented measurable wage suppression in construction trades with high immigrant concentration, particularly for plasterers, roofers, and painters [13]. The counterargument, advanced by the National Association of Home Builders, is that the construction industry already faces a severe labor shortage — an estimated 500,000 unfilled positions — and that removing immigrant workers would increase housing costs more than it would raise wages for remaining workers [18].

Technology: The H-1B program, which admits approximately 85,000 workers annually, has been simultaneously criticized for displacing American tech workers (particularly older workers at companies like Disney and Southern California Edison that replaced domestic staff with H-1B contractors) and defended as essential for maintaining U.S. competitiveness in AI, semiconductors, and biotech [19]. The Trump administration's 2025 imposition of a $100,000 fee per H-1B petition — up from roughly $2,000-$5,000 — is designed to limit use of the program. Critics argue this will disproportionately harm startups and small firms that cannot absorb the cost, while tech giants like Amazon, Microsoft, and Google can [19].

Innovation and Entrepreneurship

The economic case for immigration rests partly on measurable contributions to innovation. The American Immigration Council's 2025 report found that 231 of the Fortune 500 — 46 percent — were founded by immigrants or their children, the highest share since the organization began tracking in 2011 [12]. These companies generated $8.6 trillion in revenue in fiscal year 2024 and employed over 15.4 million people worldwide [12]. The list includes Apple, Amazon, Google, Tesla, and hundreds of less prominent firms.

Immigrants are also disproportionately represented in patenting. Foreign-born researchers account for more than half of U.S.-based Nobel Prize winners in recent decades, and immigrants file patents at higher rates per capita than native-born Americans [12]. Immigrants and their children founded 80 percent of Fortune 500 companies in professional services, 65.6 percent in manufacturing, and 57.5 percent in information technology [12].

The restrictionist response to these statistics is twofold. First, the Fortune 500 founders are overwhelmingly high-skilled immigrants who entered through employment-based visas or as children of such immigrants — they tell us little about the economic impact of low-skilled or unauthorized immigration, which is the actual policy dispute. Second, the fact that immigrants found companies does not prove that those companies would not exist absent immigration; the underlying demand and capital that created them would likely have found domestic entrepreneurs. George Borjas has argued that "immigration is not manna from heaven" and that counting immigrant-founded firms without accounting for the companies and jobs they may have displaced overstates the net contribution [14].

The pro-immigration counter: the United States' ability to attract global talent is a competitive advantage that countries like Japan — which has resisted large-scale immigration — have failed to replicate. Japan's points-based system, introduced in 2012, admits far fewer immigrants, and its economy has stagnated for decades while its working-age population contracts [20]. Japan now estimates it will need over 6 million foreign workers in the next 20 years to maintain GDP [20].

The Cost of Removal: Mass Deportation Economics

The Trump administration's pursuit of large-scale deportation has generated several economic impact assessments. The Penn Wharton Budget Model estimated that a four-year deportation program would reduce GDP by 1.0 percent by 2034, while a ten-year program would reduce it by 4.9 percent by 2054 [21]. The four-year policy would increase primary deficits by approximately $350 billion with economic feedback effects; the ten-year policy by approximately $987 billion [21].

The Peterson Institute for International Economics projected that deporting 8.3 million undocumented immigrants would lower GDP by 7.4 percent and employment by 7.0 percent by 2028 [22]. The American Immigration Council estimated a GDP contraction of 4.2 to 6.8 percent — comparable to the Great Recession [23].

Source: FRED / Bureau of Economic Analysis

Data as of Mar 22, 2026CSV

Industry-specific workforce losses would be concentrated: up to 1.5 million workers removed from construction, 225,000 from agriculture, one million from hospitality, 870,000 from manufacturing, and 461,000 from transportation [23]. Mass deportation would remove more than 30 percent of workers in major construction trades [23]. For every 500,000 immigrants removed from the labor force, an estimated 44,000 U.S.-born workers would lose their jobs as well, because the businesses employing them would contract or close [23].

Those who support enforcement-first policies make a different economic argument. The Federation for American Immigration Reform contends that unauthorized immigration costs taxpayers approximately $150 billion annually in education, healthcare, and law enforcement expenditures, net of taxes paid [24]. From this perspective, the upfront cost of deportation is an investment that reduces long-term fiscal drag. Restrictionists also argue that the economic models projecting GDP losses from deportation are measuring the wrong thing: GDP per capita, not aggregate GDP, is the relevant metric for living standards, and removing low-productivity workers could raise GDP per capita even as it reduces total output.

Historical precedents are limited and contested. The termination of the Bracero Program in 1964, which had supplied Mexican agricultural workers, did lead to some wage increases for domestic farmworkers — but also accelerated mechanization of crops like tomatoes and cotton rather than producing broad-based employment gains for native workers [25].

Housing, Schools, and Infrastructure

The second-order costs of immigration fall most heavily on high-immigration metropolitan areas. Housing is the most visible pressure point. The Harvard Joint Center for Housing Studies has acknowledged that immigration increases housing demand, but argues that rising costs are primarily caused by insufficient construction rather than excess demand [26]. Between 2022 and 2023, even as immigration increased, housing price growth slowed — suggesting that immigration is one factor among many, including interest rates, zoning restrictions, and construction costs [26].

CIS's Steven Camarota, testifying before Congress in 2024, presented a different analysis: that immigration-driven population growth in already supply-constrained metros like New York, Los Angeles, and Miami acts as a multiplier on housing costs, because new construction cannot keep pace with demand in jurisdictions with restrictive zoning [27]. He estimated that unauthorized immigration cost the education system $68.1 billion annually even before the 2021-2023 surge [27].

Public schools in high-immigration districts face enrollment pressures, language accommodation costs (English Language Learner programs cost 20-100 percent more per student than standard instruction), and staffing challenges [6]. Healthcare systems absorb uncompensated emergency care costs, though immigrants also contribute to the healthcare workforce — foreign-born workers constitute roughly 18 percent of healthcare workers nationally [16].

Immigrant homeownership partially offsets these costs: immigrant households contributed $579.1 billion in taxes in 2022, including property taxes that fund local schools and infrastructure [5]. The fiscal balance at the local level depends heavily on the specific community: a suburb absorbing high-skilled H-1B families with high incomes and few children has a very different fiscal profile than a small city experiencing rapid refugee resettlement.

Community Disruptions and Public Safety

Academic research consistently shows that immigrants — including undocumented immigrants — have lower crime and incarceration rates than native-born Americans. A 2020 PNAS study using Texas Department of Public Safety data found that undocumented immigrants had substantially lower felony arrest rates than either legal immigrants or native-born citizens [28]. Between 1980 and 2020, the foreign-born share of the U.S. population roughly doubled while total crime dropped by approximately 60 percent [29]. Cato Institute research found that immigrants were 60 percent less likely to be incarcerated than the native-born population [29].

Research on sanctuary jurisdictions — cities that limit cooperation between local law enforcement and federal immigration authorities — found that sanctuary policies reduced deportations by about one-third, primarily by reducing deportations of people with no criminal convictions, without any detectable effect on crime rates [30].

These aggregate statistics, however, do not capture the community-level disruptions that fuel political opposition to immigration. Springfield, Ohio — population 58,000 — absorbed roughly 20,000 Haitian immigrants between 2020 and 2025, a 34 percent population increase that strained schools, healthcare facilities, and housing [31]. Most arrived legally under Temporary Protected Status, but the pace of settlement overwhelmed local infrastructure regardless of legal status. Residents reported overcrowded classrooms, longer emergency room waits, and housing competition that drove up rents [31].

In Aurora, Colorado, members of the Venezuelan gang Tren de Aragua were documented operating in several apartment complexes, though the city's mayor acknowledged that claims of a "complete gang takeover" were exaggerated — police arrested eight suspected gang members, and the problem was confined to a few properties [32]. The viral claims that Haitian immigrants in Springfield were eating residents' pets were investigated and found to be baseless by city officials [31].

The pattern is consistent: aggregate crime data favors immigrants, but rapid, concentrated settlement creates real quality-of-life disruptions for receiving communities — disruptions that are distinct from crime rates and should not be dismissed as mere xenophobia.

Benefits Fraud

Documented cases of benefits fraud involving immigrant communities have attracted legislative attention. Investigations into childcare funding fraud in Minnesota and California revealed organized schemes exploiting program oversight gaps [33]. The Deporting Fraudsters Act, introduced in March 2025 and passed by the House in early 2026 on a 231-186 vote, would make public benefits fraud a deportable offense for noncitizens [33]. Democrats opposed the measure, arguing it would erode due process by permitting deportation before criminal conviction [33].

Benefits fraud is a real phenomenon but requires scale context. The Government Accountability Office has documented fraud across federal benefit programs broadly — not exclusively among immigrants. Native-born individuals commit benefits fraud at higher absolute numbers simply because they constitute a larger share of benefit recipients. The question of whether immigrant fraud rates are higher per capita than native-born rates remains empirically unresolved, with CIS and Cato reaching different conclusions based on different measurement approaches.

The Chilling Effect on Public Health

Immigration enforcement has measurable public health consequences. A 2025 KFF/New York Times survey found that 29 percent of immigrant adults skipped or postponed healthcare, with 19 percent of that group citing immigration-related fears [34]. In Dallas, back-to-school vaccinations for Hispanic children dropped from 11,500 in August 2024 to 5,800 in August 2025 — a 50 percent decline [35]. Eighty-four percent of surveyed healthcare workers reported significant or moderate decreases in patient visits since January 2025 [36].

These effects extend to U.S.-citizen children in mixed-status households. When parents fear that seeking medical care, enrolling in nutrition programs, or registering for school could expose them to immigration enforcement, their American-born children miss vaccinations, well-child visits, and SNAP benefits to which they are legally entitled [34]. The downstream costs — treating preventable diseases in emergency rooms, managing chronic conditions that could have been caught early — are borne by the same taxpayers who fund enforcement operations.

Enforcement advocates argue that this dynamic, while regrettable, is a consequence of illegal presence rather than of enforcement itself, and that the solution is compliance with immigration law rather than reduced enforcement.

Source: FRED / Bureau of Labor Statistics

Data as of Mar 22, 2026CSV

International Comparisons

The United States is unusual among wealthy nations in how it structures immigration. Canada and Australia operate points-based systems that weight age, education, work experience, and language skills to select immigrants likely to contribute economically. In Australia, 73 percent of immigration is classified as economic; in Canada, 58 percent [20]. The U.S. system, by contrast, allocates roughly two-thirds of permanent resident visas to family reunification and only about 15 percent to employment-based categories.

Canada's Express Entry and Australia's SkillSelect programs can adjust points thresholds based on labor market needs — a flexibility the U.S. H-1B system, with its fixed annual cap of 85,000 and lottery allocation, does not possess [20]. The Niskanen Center has argued that the U.S. is losing the "global race for talent" because peer nations offer faster processing, clearer pathways to permanent residence, and employer-responsive allocation systems [37].

Germany, facing a projected decline in working-age population, relaxed its Skilled Labour Immigration Act to streamline entry for non-EU workers in healthcare, IT, and engineering [20]. Japan, long resistant to immigration, introduced a points test in 2012 and now estimates it needs 6 million foreign workers over the next two decades to sustain GDP [20].

The restrictionist case draws on these comparisons differently. Canada's points system, despite its design, has faced criticism domestically: the C.D. Howe Institute published a 2024 report titled "Quality Over Quantity" arguing that Canada had admitted too many immigrants too quickly, overwhelming housing and healthcare infrastructure [38]. Canada subsequently cut its immigration targets in late 2024. This suggests that even well-designed selection systems cannot solve the infrastructure constraints that accompany rapid population growth.

The Strongest Case for Reducing Immigration

The economic case for significantly lower immigration levels rests on several empirical pillars that pro-immigration advocates often dismiss too quickly.

Wage competition for low-skill native workers is real. Borjas's estimates of 3-5 percent wage suppression for high school dropouts are at the conservative end of the range; some studies estimate larger effects in specific local labor markets [13]. Roughly 16 million native-born workers without high school diplomas face the most direct competition [16]. These workers are disproportionately Black and Hispanic, meaning immigration's wage effects fall hardest on the most economically vulnerable Americans.

The "lump of labor" fallacy has limits. The standard pro-immigration argument holds that more workers create more demand, expanding the economy rather than dividing a fixed pie [39]. This is broadly true when labor supply increases gradually and capital adjusts. But when labor supply increases faster than capital investment — as may occur during immigration surges — the adjustment period can involve years of depressed wages in affected sectors. The lump of labor fallacy is a fallacy in the long run, but in the short-to-medium run, in specific occupations, the adjustment costs are real and fall on identifiable people [39].

Fiscal costs in high-immigration states are substantial. The CBO confirmed that the 2021-2023 immigration surge imposed net fiscal costs on state and local budgets [6]. States like Texas, California, New York, and Florida bear these costs without proportional federal reimbursement. A FAIR estimate puts the net annual cost of unauthorized immigration at $150 billion [24].

Infrastructure cannot scale instantly. Adding millions of people to housing markets, school systems, and healthcare networks in supply-constrained metros produces measurable harm to existing residents, including previous immigrants. Springfield, Ohio's experience with a 34 percent population increase illustrates that pace matters as much as absolute numbers [31].

The Strongest Case for Current or Higher Immigration Levels

The fiscal math is positive in aggregate. The Cato Institute's 30-year study found immigrants reduced federal deficits by $14.5 trillion since 1994 [10]. The CBO projects $1.2 trillion in additional federal revenue from the immigration surge over the 2024-2034 decade [4]. Even undocumented immigrants contribute $96.7 billion annually in taxes [5].

Removal would be more expensive than accommodation. Every economic model that has assessed mass deportation projects significant GDP losses, ranging from 1 percent to 7.4 percent depending on scope and timeline [21] [22]. The fiscal cost of deportation operations themselves — $350 billion to $987 billion — dwarfs the annual fiscal costs that deportation is meant to eliminate [21].

The labor market needs immigrant workers. With unemployment at 4.4 percent as of February 2026, the U.S. labor market is near full employment [40]. The construction industry has 500,000 unfilled positions. Agriculture cannot find domestic workers at any realistic wage. An aging native-born population means fewer workers entering the labor force; immigration is the primary mechanism keeping the working-age population growing [41].

Innovation depends on immigrant talent. Forty-six percent of Fortune 500 companies were founded by immigrants or their children [12]. Immigrants file patents at disproportionate rates. The $100,000 H-1B fee risks driving talent to Canada, Australia, and the U.K. [19].

Second-generation convergence redeems first-generation costs. Children of low-education immigrants dramatically outperform their parents in educational attainment and earnings, approaching native-born averages by the second generation [2]. The NAS lifetime fiscal model shows this trajectory clearly: the first generation may be a net fiscal cost, but the second and third generations are net contributors [2].

What the Data Cannot Resolve

The immigration debate is ultimately a distributional question that economics alone cannot settle. Immigration produces aggregate gains — virtually all serious economists agree on this — but distributes those gains unevenly. Employers and consumers benefit from lower prices for food, construction, and services. High-skilled immigrants generate innovation and tax revenue. Low-wage native workers, concentrated among Black and Hispanic Americans without college degrees, bear measurable if modest wage costs. Local communities absorbing rapid settlement bear infrastructure and service costs that federal policy creates but does not fund.

The question of how much weight to assign humanitarian obligations versus fiscal optimization is a moral and political question, not an economic one. The NAS model shows that admitting an immigrant with less than a high school education costs $314,000 over a lifetime [2]; it also shows that their children will likely become net contributors. Whether the first-generation cost is an acceptable investment or an unacceptable burden depends on values that no regression analysis can adjudicate.

Similarly, the question of pace — how many immigrants can be absorbed per year without overwhelming local infrastructure — is an empirical question with no consensus answer. Canada's recent decision to cut immigration targets suggests that even nations with sophisticated selection systems can overshoot.

The honest summary: immigration makes America richer in aggregate, makes specific Americans poorer at the margins, and imposes real costs on specific communities. How to weigh those realities against each other is the political question. The economic data can inform that question. It cannot answer it.