A new study by the Cato Institute attempts to make the case that the use of welfare by immigrants is not really so high or costly. The report’s lead author, Leighton Ku, was for many years a researcher at two liberal D.C. think tanks, the Center on Budget and Policy Priorities and the Urban Institute. That the libertarian Cato Institute would employ Ku is a reminder that immigration is not an issue that neatly conforms to the liberal–conservative divide. Ku’s central argument is that immigrant welfare use is not worrisome because low-income non-citizens use some programs (the ones he chooses to examine) at rates similar to or even lower than those at which citizens use them.
This is an appealing message for those who, on the right as well as the left, seek to legalize the current illegal population and to increase future immigration, but there are several problems with this analysis.
First, even though the data were available, the authors chose to exclude a number of costly programs, including free or reduced school lunch, WIC, and subsidized and public housing. Second, welfare use by immigrants who have become U.S. citizens is not low. By comparing non-citizens with all citizens, naturalized immigrants as well as the native-born, they obscure the issue.
The third problem is the authors’ decision to look at only those with low incomes. Immigrants are 50 percent more likely to belong to that category (as defined in the report) than are natives. What matters to taxpayers is the overall rate of welfare use by immigrants, which is high, not their use of welfare relative to that of natives with the same income or education level. Because immigrants are more likely to be poor, they are significantly more likely to use welfare; if we compare only low-income immigrants and natives, we would miss this key point.
A main reason that welfare use among immigrants is higher than among natives is the low education of the former. Education is the single best predictor of income, welfare use, and socioeconomic status. In 2011, 28 percent of immigrants (ages 25 to 65) had not graduated from high school, compared with 7 percent of natives.
The fourth problem with the report is that, by examining individuals rather than households, the authors obscure the high welfare-use rates associated with the children of immigrants. Because the vast majority of children (defined as those under 18 years of age) in immigrant households are U.S.-born, most researchers examine households, not individuals, to get an accurate picture of the welfare use associated with immigrants. Counting the U.S.-born children with their parents is vital, because it is the low income of immigrant parents that makes the children eligible. It is the parents who signed the children up for the programs, and the parents clearly benefit by having taxpayers provide for their children. Of course, the cost to taxpayers exists only because the parents were allowed into the country.
The importance of the decision about how to classify the children can be seen by examining Medicaid. Of all children in America on Medicaid, only 3 percent are immigrants themselves, but 30 percent of all children on Medicaid have an immigrant parent.
The graph below shows the percentages of households headed by the native-born and by immigrants (both naturalized and non-citizen) that use at least one welfare program. Welfare includes cash (TANF and SSI), food assistance (SNAP, WIC, and free school lunches), public or subsidized housing, and Medicaid. The percentage of immigrant households using welfare is a good deal higher than that of native households — 36 percent versus 24 percent. The use of food-assistance programs and Medicaid is much higher among immigrants than among natives, although the two groups are similar in their use of cash assistance.
Hispanic immigrant households in particular have a very high rate of welfare use: 50 percent. This is relevant to the current debate over amnesty, as the Department of Homeland Security estimates that about 80 percent of illegal immigrants come from Latin America.