Welfare Use by "Immigrant" and "Native" Households

The Center for Immigration Studies (CIS) put out a report in the last ten days titled "Welfare Use by Immigrant and Native Households." The report managed to attract a great deal of media attention (USA Today, National Review, Washington Examiner, etc.). In its report, CIS finds that 51% of immigrant households received welfare benefits in 2012 compared to 30% of native households. My analysis of the same SIPP data finds the CIS conclusions to be misleading.

Using the same identification strategy as CIS, I found that 31.3% of native households and 51.3% of immigrant households received means-tested welfare benefits in 2012. These percentages essentially match the CIS figures.

The problem with these figures is that they exclude a lot of welfare benefits. Most notably, the nation's biggest welfare program -- Social Security -- is not means-tested and, therefore, does not show up in the CIS figures. If you take this same calculation and throw in welfare benefits from Social Security and Unemployment Insurance, the picture looks a lot different:

Under this more inclusive list of welfare benefits, 58.7% of native households and 65.2% of immigrant households received benefits at some point during the year. Immigrants are still ahead by this metric but by a much slimmer margin. Additionally, the headlines that fixated on the majority of immigrants receiving welfare benefits should be rewritten to indicate that the majority of all people, including natives, receive welfare benefits.

Here, I only added Social Security and Unemployment Insurance to the mix because the SIPP microdata makes it easy to do so. But if you were also to include other common welfare benefits like housing and child tax credits, the gap between natives and immigrants would probably close even further, and may disappear altogether.

Of course, measuring what percentage of households received welfare benefits at some point during the year is a meaningless enterprise if your goal is to figure out how much each group "burdens" the public. What matters more is how much money each group received from welfare benefits.

Although I didn't produce a welfare-dollars-per-capita estimate (which I am not sure can be done reliably with SIPP), there is good reason to think that natives receive the most overall welfare money. While 37.4% of native households received Social Security at some point during the year, only 24.9% of immigrant households did. Social Security pays out much higher benefits than other welfare programs and is much more likely to be in use the entire year while other programs are often only used for a few months. Additionally, Social Security often coincides with Medicare use, another massive welfare program.

Given that natives use welfare benefits that tend to have much higher benefit levels, it seems likely that, in dollar terms, natives more than make up for the 6.5 percentage point welfare gap in the graph above.

I know in America we like to arbitrarily distinguish between Good Handouts (Social Security, Medicare, Tax Credits) and Bad Handouts (SNAP, WIC, TANF, SSI, Medicaid). So some will be prone to dismiss my point here as not "really" capturing welfare use. In addition to this point simply being wrong ideologically, it is also practically wrong when it comes to the question of immigration. The public fisc does not care whether a dollar is going to a Good Handout or a Bad Handout. The fact that immigrants are less likely to be using our major old-age welfare programs is clearly a (fiscal) point in their favor.

Indeed, elsewhere in the world, this is one of the main reasons people give for immigration. By adding younger people to your country via immigration, you expand the size of the workforce and make it easier to support our aging population.

Technical Details. I derived these figures using the Survey of Income and Program Participation. I followed the household identification strategy described in the CIS report. Households are defined by the heads of households in January of 2012. From there, I follow those heads through each month of 2012 in SIPP. Natives are those who were born in the US and immigrants are those who were foreign born. For weighting, I used the SIPP longitudinal weight for 2012. To determine whether a household received means-tested benefits in a given month, I used the SIPP variable RHMTRF. To determine whether a household received Social Security or Unemployment Insurance in a given month, I used the variables THSOCSEC and THUNEMP.