Immigrant consumption of welfare benefits is one of the most serious economic objections to liberalized immigration. Milton Friedman famously said, “you cannot simultaneously have free immigration and a welfare state.” Elsewhere, he said, “The real hitch is in denying social benefits to the immigrants who are here.” Friedman went further when he argued that illegal immigrants are preferable to legal immigrants because the former don’t have access to welfare:

Mexican immigration, over the border, is a good thing. It’s a good thing for the illegal immigrants. It’s a good thing for the United States. It’s a good thing for the citizens of the country. But, it’s only good so long as it’s illegal.

At a basic level, there is a lot of wisdom in these points. The straight economic case for liberalized immigration is overwhelming because the median worker in the third world would earn wages about 4.1 times higher in the United States than in their home country. Immigrants have a small effect on the wages of native-born American workers mainly because immigrants don’t compete much with American workers, capital markets adjust well to increases in the number of workers, labor demand is relatively elastic, and immigrants increase demand – to say nothing about immigrants increasing GDP in the United States.

Wages are the most internationally distorted price because immigration restrictions prevent people from moving to locations with higher wages, which results in many trillions of dollars in lost production on the table annually. In a free market, goods or factors of production in a location with a lower price move or are moved to locations with a higher price, resulting in price convergence. This won’t work as well with workers as there are many other reasons why people live where they choose to, but immigration restrictions distort the labor market to an incredible degree by locking workers in countries where their wages are low. Immigrants earn higher wages in the United States because they are more productive here than abroad. Still, the United States also has a large welfare state that imposes extra costs on Americans – perhaps to such an extent that the economic benefits of immigration could be much less or even turn negative.

Michael Howard and I have a new briefing paper investigating immigrant consumption of means-tested welfare and entitlement programs and found those worries to be exaggerated.

As background, one must understand the truly massive scale of the U.S. welfare state. Total welfare state spending in the United States amounted to $2.6 trillion in 2020. In that year, the federal government spent roughly $2.4 trillion on welfare and entitlement programs, an amount equal to approximately 37 percent of all federal outlays, while the states spent an additional $229 billion. About $1.8 trillion of federal spending on the welfare state went to Social Security and Medicare, whose intended beneficiaries are the elderly, while the other roughly $600 billion went to means-tested welfare benefits, whose intended beneficiaries are the poor.

Immigrants consumed about 27 percent less welfare and entitlement benefits in 2020 than native-born Americans. Immigrants were 14.6 percent of the U.S. population and consumed just 11.1 percent of all means-tested welfare and entitlement benefits in 2020. Native-born Americans consumed $2.3 trillion of benefits in 2020, compared to $290.4 billion consumed by immigrants. All spending on means-tested welfare and entitlement programs would be about $635.5 billion less if native-born Americans consumed the same per capita dollar amount as immigrants.

2020 was a weird year because of the extra $2.1 trillion in COVID-19 spending that increased the denominator of total federal spending much more than the numerator of welfare spending. In 2019, the share of federal spending on the welfare state accounted for 51 percent of all federal spending. We expect that 2021 spending on the welfare state will be more similar to the 2019 share of federal expenditure than the 2020 share.

Our data primarily come from the Census Bureau’s 2021 Survey of Income and Program Participation, which covers the vast majority of U.S. welfare state spending. You can read our paper’s detailed methodology.

Detailed Results

The average value of welfare benefits per immigrant was $6,063 in 2020, or 27.3 percent less than the $8,335 average for native-born Americans. Figure 1 breaks down the numbers by type of welfare program. Immigrants consumed 36.9 percent less Social Security, 26 percent less Medicare, 10.7 percent less Medicaid, 11.5 percent less SNAP benefits, and 87.6 percent less TANF benefits than native-born Americans on a per capita basis. However, immigrants consumed 11.4 percent more in SSI benefits than natives, which translates to $19 more than natives on a per capita basis. Immigrants individually also consumed 42.9 percent more WIC benefits than native-born Americans, which translates to $7 more than natives per capita.

Immigrants still consumed less than natives when the entitlement programs of Social Security and Medicare are excluded from the analysis. The average immigrant consumed $2,273 in means-tested welfare benefits in 2020, about $310 less than the average American who consumed $2,583. That’s a difference of 12 percent.

Separating native-born and immigrant welfare consumption by age shows more detail. Figure 2 shows the per capita welfare cost for native-born American and immigrant children aged 0–17. Total welfare costs for immigrant children are 45 percent lower than for native-born American children, and the former consume less per capita than the latter for all programs.

Figure 3 shows the per capita welfare cost for working-age immigrants and native-born Americans aged 18–64. The value of welfare benefits consumed by working-age immigrants is about 38 percent less than that consumed by working-age native-born Americans.

Figure 4 displays the per capita welfare cost for native-born Americans and immigrants aged 65 or older. Elderly immigrants consume, on average, about 19 percent less in welfare benefits than elderly native-born Americans. For Social Security benefits and Medicare, the two largest programs, elderly immigrants consume less than elderly natives; for most smaller programs and Medicaid, the opposite is true. This suggests that immigrants ineligible for Social Security and Medicare use cheaper means-tested programs as substitutes. Reducing means-tested benefits for elderly immigrants should be an easy lift for policymakers.

The Big Picture

Elderly immigrants consume more Medicaid benefits than elderly native-born Americans, but natives are more expensive than immigrants in the same age groups on a per capita basis for all other large programs and most smaller ones. Immigrants are already legally restricted from accessing most welfare programs for some years after their arrival, but minor legal changes along the lines we suggest here would significantly reduce immigrant access to all these programs. Rep. Grothman (R‑WI) has introduced multiple bills to end welfare access to non-citizens, which is second best to massively scaling back welfare for all.

In theory, immigrants who have immediate access to welfare benefits upon arrival can impose costs on native-born American taxpayers that could swamp their economic contributions. However, the lower immigrant consumption of welfare benefits is evidence that immigrants don’t impose so large of a cost – although it could and should be lower. Evaluating the net-fiscal effects of immigration – whether the taxes paid due to them being here is greater than their consumption of benefits – requires more complicated calculations and estimates. Stay tuned for those.

The Dynamic Political Effects of Immigrants on Welfare

The results above are only present in 2020, a single point in time, and some other years that other authors and I have studied. Although immigrants consistently consume less welfare and entitlement benefits than native-born Americans, the total size of the welfare state could be larger because of immigrant voting or their indirect effects on political outcomes. Thus, total spending on welfare for natives and immigrants could be much larger because immigrants are present in the United States. In a paper I wrote with Benjamin Powell and J.R. Clark on how the mass immigration of Soviet Jews to Israel affected economic freedom there, we found that immigrants caused a huge improvement in four out of five of the major measures of capitalism – except the size of government variable that worsened initially. We wrote:

The dip in the score for the size of government is unsurprising since these economic migrants were poorer than most of the native born and immediately qualified for welfare state benefits since they became full citizens upon arrival. Also, immigrants to Israel are less likely to generate “blow back” against the welfare state from veteran citizens because of the Zionist notion that the state is obligated to take care of all world-wide Jews when they return to Israel. The decline in economic freedom in this area was largely driven by the fact that transfers and subsidies rose from 16.7 to 22.8 percent of GDP over the decade, while marginal tax rates were also increased.

The size of government variable evened out in the long-run after the mass immigration of Soviet Jews ended in about 1995, falling “23 percent in the decade before recovering to 97 percent of its pre-migration wave score in 2005 and subsequently improving past its original level by 2010.” Even in Israel, the long-run effect was positive from an economic freedom perspective even though the short-run effect was negative.

There are a few reasons to doubt that immigrants would interact with the political system to produce a larger welfare state. The first is that increased immigrant-driven diversity could undermine broad political support for welfare or at least reduce support for expanding it – the “blow back” mechanism we mentioned above in the Israeli context.

A marvelous paper intended to see whether social scientists converge on the same results by using the same data asked 73 research teams to analyze the same cross-country data set to see whether immigrants reduced support for welfare. The paper provided a dismal view of social science as there was little agreement on the total effect immigrants had on support for welfare and wildly different methods that were often opaque. Still, hold your nose and look at the distribution of findings: 25.4 percent of papers found that the average marginal effect of immigrants reduced public support for welfare, 57.7 percent found no effect, and only 16.9 percent found a positive effect. The research that found a positive effect could still be the correct finding though. Although I’m less confident in the overall “blow back” literature because of this paper, the hypothesis is worth considering and is a possible reason to expect that more immigrants will not increase support for welfare.

The second reason to be skeptical that immigrants would increase the size of the welfare state is that immigrants undermine labor unions. In a recent working paper, my co-authors and I found that immigrants reduced “union density by 5.7 percentage points between 1980 and 2020, which accounts for 29.7 percent of the overall decline in union density during that period.” Immigration undermining labor unions is important here because unions have been an important force for progressive policy changes in the United States, including building support for a large welfare state. Workers who join a union become more progressive, are more skeptical of the benefits of markets, and are more anti-trade – and it’s due mostly to treatment effects and not selection effects. This is likely true in other countries, but I haven’t tested it and the literature is thin on this topic.

During the 1922 to 1968 period when immigration was heavily restricted in the United States, real per capita federal expenditures grew by 1,335 percent – 75 times more than the 45-year period after immigration was slightly liberalized in 1968. The New Deal and the Great Society, which created the federal welfare state, were passed when immigration was restricted. As Paul Krugman theorized, immigration restrictions “help[ed] to create the political conditions for a stronger social safety net.” He’s probably correct and free marketeers who support immigration restrictions because they fear that immigrants will undermine economic freedom and grow the welfare state should note that the correlations run in the opposite direction and at least deal with that rather than waste time on analyzing vague cultural variables. In American states from 1970–2010, we find that “larger immigrant shares of the population produce large reductions in the growth of real per capita tax revenue and outlays in the short run that moderate to smaller longer-term growth declines in both.”

The third reason to doubt that immigrants could increase welfare is the “keep ‘em happy!” effect. Many policymakers and voters think that mass immigration creates problems. The best way to keep people happy is to create jobs and the easiest way to do that is through liberalization. This is most obviously true in Jordan, where a sudden mass immigration of refugees in 1990–1991 equal to 10 percent of the pre-migration population resulted in the government undertaking concerted and radical economic liberalization, in part to make sure the new immigrants would have economic opportunities and to reduce the chance of social problems. Welfare is much more expensive, difficult to enact in the face of a potential “blow back” effect, and jobs make people happier.

Conclusion

The average immigrant consumed 27.3 percent less welfare than native-born Americans in 2020. Further reform should reduce that substantially by building a higher wall around the welfare state instead of around the country. Still better would be reducing or eliminating the welfare state for all but reducing or eliminating non-citizen access is a politically feasible first step. In the long-run, there are good reasons to think that immigrants won’t increase the size of the welfare state and compelling evidence that they would reduce support for it. Still, the net-fiscal effects are most important – watch this space for further developments on that important topic.