Welfare Use by Legal and Illegal Immigrant Households

By Steven A. Camarota September 2015

Download a PDF of this Backgrounder


Steven A. Camarota is the director of research at the Center for Immigration Studies.


This report is a companion to a recent report published by the Center for Immigration Studies looking at welfare use by all immigrant households, based on Census Bureau data. This report separates legal and illegal immigrant households and estimates welfare use using the same Census Bureau data as that study. This analysis shows that legal immigrant households make extensive use of most welfare programs, while illegal immigrant households primarily benefit from food programs and Medicaid through their U.S.-born children. Low levels of education — not legal status — is the main reason immigrant welfare use is high.

Among the findings:

  • An estimated 49 percent of households headed by legal immigrants used one or more welfare programs in 2012, compared to 30 percent of households headed by natives.

  • Households headed by legal immigrants have higher use rates than native households overall and for cash programs (14 percent vs. 10 percent), food programs (36 percent vs. 22 percent), and Medicaid (39 percent vs. 23 percent). Use of housing programs is similar.

  • Legal immigrant households account for three-quarters of all immigrant households accessing one or more welfare programs.

  • Less-educated legal immigrants make extensive use of every type of welfare program, including cash, food, Medicaid, and housing.

  • The overwhelming majority of illegal immigrants have modest levels of education; therefore, the high use of welfare associated with less-educated legal immigrants indicates that legalization would likely increase welfare costs, particularly for cash and housing programs.

  • Restrictions on new legal immigrants’ access to welfare have not prevented them from accessing programs at high rates because restrictions often apply to only a modest share of immigrants at any one time, some programs are not restricted, there are numerous exceptions and exemptions, and some provisions are entirely unenforced. Equally important, immigrants, including those illegally in the country, can receive welfare on behalf of their U.S.-born children.

Introduction

This report is a companion analysis to "Welfare Use by Immigrant and Native Households: An Analysis of Medicaid, Cash, Food, and Housing Programs", recently published by the Center for Immigration Studies. That report examines welfare use for all immigrant-headed households, regardless of legal status, using the Census Bureau's Survey of Income and Program Participation (SIPP). This analysis uses information in the survey to distinguish legal and illegal immigrants to estimate welfare use by legal status. The findings show that, relative to natives, welfare use is high for both legal and illegal immigrant households.

As discussed at length in our earlier study, there is a significant body of research showing that the SIPP provides the most accurate picture of welfare use of any Census Bureau survey. Using the SIPP data to estimate welfare use based on the legal status of the household head shows that both legal and illegal immigrant households make extensive use of the nation's welfare system. Legal immigrant households have higher welfare use than native households for cash, food, and Medicaid, while illegal immigrant households have higher use than natives for food and Medicaid, but lower use for cash and housing programs.

Methodology

In this report we use the terms immigrant and foreign-born synonymously. The foreign-born include all individuals who were not U.S. citizens at birth. We use the term illegal immigrant to describe those in the country without authorization; some publications refer to these individuals as illegal aliens or as undocumented immigrants.

Data Source. Our prior publication, "Welfare Use by Immigrant and Native Households: An Analysis of Medicaid, Cash, Food, and Housing Programs", provides a detailed discussion of why the SIPP produces more accurate estimates of welfare use than other Census Bureau surveys.1 That report also includes a detailed discussion of the structure of the survey, including how the data is collected and weighted.2

Programs Examined. Like our prior analysis, the major welfare programs examined in this report are Supplemental Security Income (SSI), Temporary Assistance to Needy Families (TANF), the Women, Infants, and Children food program (WIC), free or subsidized school lunch, food stamps (Supplemental Nutrition Assistance Program, SNAP), Medicaid, and public housing and rent subsidies.3

Estimating Illegal Immigrants. There is a significant body of research showing that illegal immigrants respond to Census Bureau surveys. The Department of Homeland Security uses the Census Bureau's American Community Survey (ACS) as the basis for estimating the size of the illegal immigrant population. DHS estimates that 90 percent of the illegal immigrant population is included in the ACS.4 The Pew Hispanic Center uses the Current Population Survey and the ACS to estimate the size of the illegal immigrant population based on the same assumption that about 90 percent of illegal immigrants are included in the survey.5 The Center for Immigration Studies has also used these data sources to estimate the illegal immigrant population.6

Researchers also have estimated the legal status of immigrants in the SIPP and found that it produces estimates of the illegal immigrant population similar to estimates based on other data.7 One advantage of the SIPP over other Census Bureau surveys is that it includes two questions on immigration status. First, non-citizens are asked if they came to the United States as lawful permanent residents (LPRs); and second, it asks those who did not come as LPRs if they have changed their status since they arrived.8 The overwhelming majority of these non-LPRs are likely illegal immigrants.9

There were 2.1 million households headed by non-citizen, non-LPRs in the 2012 SIPP. We estimate there should be a total of 3.4 million households headed by illegal immigrants in the 2012 SIPP.10 In a number of cases, respondents did not provide a response to the migration variables and these values were imputed (filled in) by the Census Bureau using a complex algorithm.11 Our analysis focused on these imputed cases when trying to identify the illegal immigrant population that cannot be identified using the migration variables. To identify those illegal immigrants, we use citizenship status, year of arrival in the United States, age, region of origin, educational attainment, sex, receipt of some welfare programs, and marital status. We use these variables to assign probabilities to each respondent. Those individuals who have a cumulative probability that makes them likely to be illegal immigrants are assigned that status, the remaining non-citizen population as well as naturalized citizens are the legal immigrant population.12

By design, this estimate is meant to match the characteristics of the estimates published by DHS by age, duration of stay, and region of origin.13 However, no estimate of the illegal immigrant population is definitive and there is always uncertainty.14 Although we report confidence intervals in this report for households by legal status, these intervals reflect uncertainty in the survey data itself and do not reflect the added uncertainty associated with imputing legal status.

It is also worth noting that households headed by the foreign-born can be only headed by a legal or illegal immigrant. This means that if our estimate of the size of the illegal immigrant population is too high, then our estimate of the legal immigrant is correspondingly too low. By the same logic, if our estimate of the illegal immigrant population is too low, then the size of the legal immigrant population is too high. The same logic applies to welfare use rates. For example, 51 percent (8.3 million) of legal and illegal immigrants in the SIPP reported using one or more welfare programs in 2012. All of the results reported in this analysis total back to these numbers. There is no imputation of welfare use in this report. The analysis relies entirely on what each household reported in the SIPP.

Immigrant Eligibility for Welfare. The appendix to the companion report includes a long discussion of the regulations and laws covering immigrant welfare eligibility. As it makes clear, the 1996 welfare reform changes and other laws are designed to limit immigrant reliance on welfare programs — particularly legal immigrants, since illegal immigrants have been barred from welfare for a long time. These policies include the long-standing "public charge" doctrine that bars entry of immigrants likely to need welfare and the deportation of those who become dependent on it, a five-year bar on most new legal immigrants accessing welfare, and a "deeming" requirement that sponsors' income be considered before welfare can be received.

However, as the appendix in our prior analysis makes clear, these restrictions do not prevent immigrant households from making extensive use of welfare programs because restrictions often apply to only a modest share of legal immigrants at any one time, some programs are not restricted, there are numerous exceptions and exemptions, and some provisions are entirely unenforced. Equally important, immigrants, including those illegally in the country, can receive welfare on behalf of their U.S.-born children. As a result, both our prior report and this one show that immigrant households make extensive use of the nation's welfare system, often at significantly higher rates than native households.

Findings

The Two-Variables Method. The far right column in Table 1 reports welfare use for illegal immigrant households using only the two migration variables discussed in the methodology section of this report. Although there is an undercount of illegal immigrants using this method, the table shows that there is simply no question that households headed by illegal immigrants access a good deal of welfare. In fact, illegal immigrants' use of some programs is quite high. For example, using just the two migration variables, 30 percent of households headed by illegal immigrants are on food stamps and 56 percent have at least one person on Medicaid. Any suggestion that there are no welfare costs associated with illegal immigrants is incorrect. The SIPP shows that households headed by individuals with a very high probability of being in the country illegally make significant use of food programs and Medicaid.

Results from the Full Model. Figure 1 and the second column from the right in Table 1 show welfare use based on a fully developed model, which includes those identified as illegal immigrants using the migration variables plus illegal immigrants identified based on other information in the SIPP. Illegal immigrant-headed households' use of welfare is high at 62 percent, higher than the 30 percent for households headed by native-born Americans. Thus, a much larger share of illegal immigrant-headed households use one or more welfare programs than households headed by natives. Table 1 and Figure 1 also indicate that, while households headed by illegal immigrants make some use of housing and cash programs, their use is lower than that of households headed by the native-born for these programs.

Turning to households headed by legal immigrants shows that, unlike illegal immigrant households, legal immigrant households use every type of program at relatively high rates. Legal immigrant households have higher welfare use than native households for cash, food, and Medicaid. Looking at the individual programs shown in Table 1 indicates that legal immigrant households have statistically significant higher use than natives for SSI, school lunch, WIC, food stamps, and Medicaid. Table 1 makes clear that the inclusion of illegal immigrants in the SIPP does not explain the relatively high use of welfare by immigrant households overall. Illegal immigrants tend to lower the rates for immigrants for some programs, while raising it for others. But both legal and illegal immigrant households make extensive use of the welfare system.

In terms of the share of immigrant households using welfare, households headed by legal immigrants account for most program use. Of all immigrant households (legal and illegal) using one or more welfare programs, legal immigrants account for 75 percent of the total. Of immigrant households using cash, 92 percent are headed by legal immigrants, as are 71 percent of households using food assistance, 74 percent using Medicaid, and 87 percent in public or subsidized housing. This should not be surprising as households headed by legal immigrants account for 79 percent of all immigrant households in the SIPP.

It is fair to say that efforts to curtail immigrant welfare use have not been particularly effective given the high welfare use of legal and illegal immigrants despite all of the restrictions on immigrant welfare use that are in place. Any suggestion that such costs can be avoided in the future by new legislation should be met with significant skepticism. Politically and practically it is difficult to prevent low-income people from accessing these programs once they have been allowed into the country. Further, the results indicate that knowledge of how to navigate the welfare system is extensive in immigrant communities and no change in the law would change this reality.

Refugees and Asylees. Those admitted to the country for humanitarian reasons are allowed immediate access to the welfare system. However, as we discussed at length in the companion study to this report, humanitarian immigration does not explain the high welfare use associated with immigrants. In summary, that report showed: First, of all permanent legal immigrants admitted in the last three decades, only 13 percent were admitted as refugees or asylees. Second, immigrants from Latin America have the highest welfare use, yet few were admitted as refugees or asylees. Third, while the SIPP does not identify countries, the Census Bureau’s Current Population Survey does. Our prior analysis of that survey by sending country showed that households headed by refugee-sending countries had almost the same overall welfare use rates as non-refugee sending countries. In short, refugee use rates are not that different from immigrants overall, and they are not a large enough share of the total legal immigrant population to account for the high welfare use of legal immigrant households.

Education and Legal Status. Figure 3 reports welfare use for legal immigrants and natives based on the educational attainment of the household head. (Table A1 in the appendix at the end of this report provides the same information with more detail.) The key finding in Figure 3 and Table A1 is that less-educated legal immigrants have extremely high use of welfare. The term "less-educated" for the duration of this report refers to those households headed by someone who has not graduated from high school or has only a high school education. For legal immigrant households headed by high school dropouts, 75 percent used at least one program, as did 64 percent of legal immigrant households headed by someone with only a high school education. Less-educated illegal immigrants also tend to use welfare programs across the board.

None of the findings in Figure 3 and Table A1 are surprising. Individuals with low levels of education often earn modest incomes and, as a result, make extensive use of the welfare system. This is the case for households headed by less-educated natives as well. But it does mean that legalizing illegal immigrants will create very significant new welfare costs, especially for cash and housing programs. The SIPP indicates that about three-fourths of adult illegal immigrants have no education beyond high school. Once amnestied, they will become less-educated legal immigrants, the group that uses all types of welfare at high rates. Any discussion of the issue must include an understanding that amnesty would increase welfare costs.

Another important finding in Figure 3 is that households headed by legal immigrants tend to have higher use rates than native households with the same level of education. In 15 of the 20 comparisons by education in Figure 3, legal immigrant households have statistically significant higher use rates than native households with the same education. Legal immigrant welfare use is also higher in four other cases, but the difference is not statistically significant. In one case, native welfare use is higher than immigrant use. Figure 3 shows that households headed by immigrants legally in the country tend to use welfare at higher rates than native households, even natives with the same education. This is an important finding because it indicates that the higher use of welfare associated with legal immigrants for most programs is not simply due to the large share with modest levels of education. Even households headed by well-educated legal immigrants use a good deal more welfare than their native-born counterparts.

The situation is somewhat different for illegal immigrants. Figure 4 contains the same 20 comparisons of welfare use by education as Figure 3, but the comparison is illegal immigrants vs. natives. (Table A1 shows more detailed information by education for illegal immigrant households.) In 10 of the 20 comparisons, illegal immigrant households have higher welfare use than native households with the same education; and in five of the 20 comparisons native use is higher than immigrant use. Illegal immigrant use of welfare is very high for food programs and Medicaid, while use of other programs is much more modest.

Working Households. Table 2 reports welfare use for households that had at least one worker in 2012. It shows very high use of welfare for working immigrant households — both legal and illegal. Almost half (48 percent) of working households headed by legal immigrants use the welfare system. Rates are particularly high for food programs (36 percent) and Medicaid (38 percent). With almost one in nine working legal immigrant households using cash assistance, use of such programs is also high.

It may seem strange that households with workers have such high welfare use rates, but the nation's welfare system is explicitly designed to assist low-income workers with children, especially after the 1996 welfare reform. It is very common for people who work full- or part-time for at least some portion of the year to use welfare programs. For example, a family of two children and a single parent who works full-time earning $20,000 a year would meet the income requirement for just about every welfare program in almost every state.15 So even though the vast majority of households, immigrant or native, had least one worker in 2012, these working households still made extensive use of the welfare system.16

Appendix Table A2 reports welfare use for households that work based on the education and legal status of the household head. Table A2 shows that less-educated working households make extensive use of the welfare system. Former President George W. Bush once observed that he wanted to increase legal immigration so as to "match willing foreign workers with willing American employers."17 But Table A2 makes clear this is a simplistic and one-dimensional way of thinking about immigration. A person may work, but also create very significant costs for the welfare system. However, those costs are diffuse, borne by all taxpayers, while employers get the workers they want and the immigrants improve their lives by coming to the United States. Focusing only on the desire of employers to bring in additional workers or the desire of immigrants to come to America misses the potentially enormous impact immigrant workers can have on American taxpayers.

When welfare is taken into consideration, allowing immigrant workers into the country to perform low-wage jobs is clearly problematic. It would, at least from the point of view of avoiding welfare expenditures, make more sense to hire from the enormous pool of less-educated natives not working rather than adding less-educated individuals through our immigration system.18 To be clear, Table A2 indicates that, in some cases, less-educated native workers will use welfare at rates nearly as high as less-educated immigrants, but the natives are already in the country. Bringing in immigrants in the future, on the other hand, creates new costs for the welfare system.

Households with Children. Table 3 reports welfare use for households with children. Not surprisingly, welfare use is much higher for households with children than for all households. The nation's welfare system is specifically designed to help households with children. Programs like WIC, free/subsidized lunch, and Medicaid for children (referred to as the Children's Health Insurance Program, CHIP) were explicitly created for those under age 18. TANF, formerly called Aid to Families with Dependent Children, is also designed to help low-income children and their families. Other welfare programs benefit children as well.

Even though many welfare programs are targeted at households with children, it is still remarkable that, in 2012, 72 percent of households headed by legal immigrants with children used at least one welfare program. The rate was 87 percent for illegal immigrant households with children and 52 percent for native households. These figures may seem surprisingly high, but it must be remembered that these figures are based only on self-reported use of welfare programs in the public-use file of the SIPP. Moreover, a Census Bureau report published in December 2014 using 2011 SIPP data showed that 65 percent of the nation's children live in a household using one or more welfare programs. So the results in Table 3 are entirely consistent with what the Census Bureau itself has found.19

Legal immigrant households with children have statistically higher welfare use than their native counterparts for food and Medicaid. For cash programs, the difference is not statistically significant and the native rate is higher for housing programs. The same basic pattern exists for households with children headed by illegal immigrants, with the exception that the higher native use of cash is statistically significant. Table 3 indicates that a very large share of legal immigrants come to America with children or have children after they arrive and struggle to support them. As a result, they end up using the welfare system. To be sure, the rates for native households with children are also high. But at a time when many Americans struggle to support their children, we are adding millions of immigrants who are even less likely to be able to do so.

Households without Children. Of all immigrant households, 47 percent have children and 92 percent of these households have at least one U.S.-born child. But receiving welfare on behalf of their children is not the only way legal immigrants access welfare. The right side of Table 3 shows that 31 percent of legal immigrant households with no children used one or more welfare programs in 2012, compared to 20 percent of native households with no children.20 Legal immigrant households without children have statistically significant higher rates of overall welfare use than native households without children; and for every type of welfare program — cash, food, Medicaid, and housing. Thus, it is not the case that the presence of U.S.-born children by itself explains the relatively high welfare use associated with legal immigrants. Even when no children are present, U.S.-born or foreign-born, legal immigrant-headed households report high rates of welfare relative to native-headed households with no children.

Turning to illegal immigrant households without children, Table 3 shows that the difference with native households without children is not statistically significant over¬all, but the native rate is higher for cash. This is an indication that, unlike legal immigrants, welfare use among illegal immigrants is much more associated with children. Table 3 shows that childless illegal households still use some welfare programs at surprisingly high rates, such as 8.7 percent for food stamps and 16.9 percent for Medicaid. Several things should be taken into account when considering these figures. First, fraud and administrative errors allow some number of illegal immigrants to collect benefits.21 Second, illegal immigrants can live with adult green card holders or U.S. citizens who can use these programs. Third, in the case of Medicaid, pregnant women illegally in the country can sometimes be enrolled in the program. Fourth, there is the Emergency Medicaid program that covers predominately illegal immigrants.22 Finally, it must be remembered that illegal immigrant households without children accessing these programs comprise only 14 percent of all illegal households receiving welfare benefits.

Conclusion

If one assumes that legal immigration is supposed to benefit the country, then immigrants allowed into the country should have much lower rates of welfare use than natives. However, our best estimate is that nearly half (49 percent) of households headed by legal immigrants used at least one welfare program in 2012, compared to less than one-third (30 percent) of native households. Households headed by legal immigrants have statistically significant higher use of welfare than native households overall, and for cash and food programs as well as Medicaid. This is the case a decade and a half after the 1996 welfare reform law, which was supposed to greatly limit immigrant access to the welfare system. Among households headed by an illegal immigrant, we estimate that 62 percent use one or more welfare programs. While the overall rate is very high, however, households headed by immigrants illegally in the country primarily use food programs and Medicaid, and their use rates of cash and housing programs are lower than native households' use rates.

The high rate of welfare use by immigrant households may be surprising given the restrictions on welfare enacted in 1996. But those restrictions cover only a modest fraction of legal immigrants at any one time, many programs are not barred, and numerous exceptions exist. Further, benefits are often received on behalf of U.S.-born children who have the same welfare eligibility as any other citizen.

One main reason immigrant welfare use tends to be higher than natives' is the larger share of immigrants who have low levels of education. Households headed by less-educated immigrants have very high rates of welfare use. Those with modest levels of education tend to earn modest wages and, as a result, qualify for welfare programs. Of households headed by legal immigrants without a high school diploma, 75 percent use one or more welfare programs, as do 64 percent of households headed by legal immigrants with only a high school education. Less-educated legal immigrants make extensive use of every type of welfare program, including cash, food, Medicaid, and housing.

The high use of welfare associated with less-educated legal immigrants indicates that legalizing illegal immigrants would likely increase welfare costs because the overwhelming majority of illegal immigrants have modest levels of education. Once legalized, most illegal immigrants would become less-educated legal immigrants, the group with the highest welfare use rates. By itself this does not mean that amnesty is a bad idea. But it does mean that any discussion of legalization must recognize the likely implications for the welfare system. Of course, any amnesty would be accompanied by restrictions on welfare use for the newly legalized, but the current restrictions have not prevented legal immigrants or illegal immigrant households from using the welfare system, so it seems likely that new restrictions would be equally ineffective.

The high rate of welfare use by legal immigrant households does not mean that they came to America to get welfare. Most immigrants come to America to work, and the vast majority of immigrant households, legal or illegal, have one or more workers. But the relatively low education level of a large share of immigrants partly explains why roughly half of legal immigrant households with a worker present accessed one or more welfare programs in 2012. Education is not the only factor that explains heavy use of welfare by legal immigrants, however. Among the most educated legal households, those headed by a person with a bachelor's degree or more, legal immigrant households are still much more likely to use all forms of welfare than native households with the same level of education. So other factors, such as culture and the exchange of information provided by immigrant social networks, also likely play a significant role in explaining why immigrants access welfare programs at high rates.23

A large share of the welfare used by immigrant households is received on behalf of their U.S.-born children. This is especially true of households headed by illegal immigrants. Illegal immigrant-headed households without children make much more modest use of welfare programs. One way to describe what happens in regard to welfare is to recognize that most immigrants come to America to work, and most find jobs. However, many earn low wages because of their education levels, language skills, or other factors. As a result, a large share of immigrants, especially those with children, are eligible for welfare programs, primarily food assistance and Medicaid, and to a lesser extent cash assistance. In addition, immigrant social networks and social service organizations are likely effective in helping immigrants navigate the welfare system.

If we want to avoid heavy use of welfare programs by immigrant households in the future, then selecting only the most skilled immigrants who will tend to use welfare less is probably the best course of action. This analysis makes clear that trying to bar immigrants from welfare usage after they have been allowed into the country almost certainly will not work. It has been tried before with the 1996 welfare reform law. As for illegal immigrants, enforcing immigration laws and encouraging them to return to their home countries is the only way to avoid costs. Legalizing illegal immigrants would likely increase their use of welfare programs over time, as most illegal immigrants have modest levels of education. If we are not prepared to enforce immigration laws and implement a highly selective legal immigration system, then heavy use of welfare by immigrants seems unavoidable.


Appendix


End Notes

1 A detailed discussion and summary of the research showing that the SIPP is the most accurate survey of welfare use can be found in the Methodology section under subsections "Why Use the SIPP" and "The Superiority of SIPP Data" in the companion report to this analysis, Steven A. Camarota, "Welfare Use by Immigrant and Native Households: An Analysis of Medicaid, Cash, Food, and Housing Programs", Center for Immigration Studies, 2015.

2 See the Methodology section under the subsections "Data Source", "Unit of Analysis", and "Weighting" in "Welfare use by Immigrant and Native Households: An Analysis of Medicaid, Cash, Food, and Housing Programs".

3 Although we report figures for SSI and TANF separately, the SIPP variable "rhcbrf" is used to measure all cash welfare and the variable includes several small federal programs, such as those for low-income veterans as well as state assistance programs similar to SSI and TANF. Our reporting of all cash programs includes these programs and it is for this reason that in a few cases cash use is higher than the sum of SSI and TANF. Unlike individual food programs, subsidized housing and public housing cannot be used at the same time. However, some households moved from one type of supported housing unit to another during 2012. For this reason, the total use of housing reported in the tables is often less than the sum of the use of subsided and public housing. Housing programs include several different programs and arrangements, most of which are funded by the federal government. State and local governments also have their own programs.

4 Bryan Baker and Nancy Rytina, "Estimates of the Unauthorized Immigrant Population Residing in the United States: January 2012", Department of Homeland Security, March 2013.

5 Jeffrey S. Passel and D'Vera Cohn, "Unauthorized Immigrant Totals Rise in 7 States, Fall in 14: Decline in Those From Mexico Fuels Most State Decreases", Pew Research Center, November 2014.

6 This report uses the Current Population Survey to estimate the illegal immigrant population. See Steven A. Camarota, "Immigrants in the United States, 2010: A Profile of America's Foreign-Born Population", Center for Immigration Studies, August 2012. In this report we use the ACS to estimate illegal immigrant in the labor force. See Steven A. Camarota and Karen Zeigler, "Are There Really Jobs Americans Won't Do? A detailed look at immigrant and native employment across occupations", Center for Immigration Studies, May 2013.

7 See James D. Bachmeier, Jennifer Van Hook, and Frank D. Bean, "Can We Measure Immigrants' Legal Status? Lessons from Two U.S. Surveys", International Migration Review, Summer 2014; Jeanne Batalova, Sarah Hooker, and Randy Capps, "DACA at the Two-Year Mark: A National and State Profile of Youth Eligible and Applying for Deferred Action", Migration Policy Institute, August 2014.

8 These two migration questions were part of wave 2 in the topical module collected in 2008 and 2009. Immigrants who indicated they were not citizens were asked about their immigration status when they arrived. In the public-use data, non-citizens are recorded as either having entered as a lawful permanent residents (LPR) or not. (The full SIPP data provides more detailed information, but this is not available to the public.) Those non-citizens who indicated they did not enter as LPRs are then asked if their status has changed to LPR. These non-LPRs will also include a small number of temporary visitors (e.g. guestworkers and students) and an even smaller number of refugee and asylees not yet adjusted to LPR status. (Refugees and asylees adjust to LPR status after one year.) Looking at the household heads that are the focus of this analysis indicates that just 5 percent of these non-LPRs have been in the country for less than five years, a clear indication that few are unadjusted refugees and asylees or long-term temporary visitors. As for temporary visitors, virtually none should be receiving welfare as they are not eligible for these programs, so their inclusion in the data should lower the use rates of illegal immigrants.

9 James D. Bachmeier, Jennifer Van Hook, and Frank D. Bean, "Can We Measure Immigrants' Legal Status? Lessons from Two U.S. Surveys", International Migration Review, Summer 2014. This is the case because there is no other immigration status that could account for such a large population.

10 The Heritage Foundation has estimated 3.4 million illegal immigrant households in the 2011 Current Population Survey Annual Social and Economic Supplement. See Robert Rector and Jason Richwine, "The Fiscal Cost of Unlawful Immigrants and Amnesty to the U.S. Taxpayer", Heritage Foundation, 2013. Our analysis of the 2011 and 2012 Current Population surveys confirms this estimate. The SIPP is controlled to similar population totals. This estimate does not include an undercount of illegal immigrants in the SIPP.

11 In a small number of cases, immigrants did not answer the migration questions because they joined the survey after 2009, when the questions were asked.

12 The probabilities are assigned so that both the total number of illegal aliens and the characteristics of the illegal population closely match other research in the field, particularly the estimates developed by the Department of Homeland Security. This method is based on some very well established facts about the characteristics of the illegal population. For example, it is well known that illegal aliens are disproportionately young, male, under age 45, have few years of schooling, etc. Thus, we assign probabilities using these and other factors in order to select the likely illegal population. In some cases, we assume that there is no probability that an individual is an illegal alien, such as having been in the country for 30 or more years.

13 Bryan Baker and Nancy Rytina, "Estimates of the Unauthorized Immigrant Population Residing in the United States: January 2012", Department of Homeland Security, March 2013.

14 When determining the likely illegal immigrant population in the SIPP, the lack of a country variable in the public-use files of the SIPP makes the estimates more difficult than estimating illegal immigrants in the ACS or CPS, both of which provide a country variable. However, Mexican and Central American immigrants have very similar probabilities of being in the United States and together they account for about three-fourths of the illegal immigrant population. Thus, the Census Bureau's decision to group the SIPP data by region rather than country does not prevent reasonable estimates of the illegal immigrant population.

15 The formula for calculating eligibly for welfare programs varies significantly by state. Some income, such as disability payments, is excluded when calculating eligibility in states. Also, some states allow expenditures on things like child care to be deducted from income when calculating the threshold for accessing a program. The benefits from programs like SNAP and TANF vary in size based on income. In this way, the income threshold to use a program can be much higher than one might imagine. For example, if an elderly person is present a family of three can still receive some benefits from SNAP in New York state even if their income is as much as $39,576 a year. The threshold is nearly $47,700 a year in New York if the family has four members. See the New York state income threshold table on the state's SNAP website. In Mississippi, a low-cost-of-living state with a less-generous welfare system, a family of four can earn up to $31,525 and receive SNAP. See the Mississippi income threshold table on the state's SNAP website.

16 Of legal immigrant households, 85 percent had one or more workers, as did 95 percent of illegal immigrant households and 76 percent of native households.

17 White House speech by George W. Bush, January 7, 2004.

18 In 2014, there were more than 25 million native-born Americans ages 18 to 65 with no education beyond high school not working. See Table 6 in Steven A. Camarota and Karen Zeigler, "All Employment Growth Since 2000 Went to Immigrants", Center for Immigration Studies, June 2014.

19 While the unit of analysis in the Census report is children, not households as in this analysis, the findings still indicate that it is the norm for children in America today to live in households using welfare. See Lynda Laughlin, "A Child's Day: Living Arrangements, Nativity, and Family Transitions: 2011", U.S. Census Bureau, December 2014. The report was based on SIPP data collected in 2011. The bottom of Table 1 shows 47.9 million children living in households using TANF, WIC, Medicaid, food stamps, or subsidized lunch and 26.4 million living in households not using these programs.

20 The tiny number of households without children that report using programs designed for children reflects the following: For households without children, use of WIC is due to pregnant women and children born into the household or who joined it after January; use of the school lunch program reflects school-age children who joined the household after January. Use of TANF is due to children who joined the household after January, children born into the household after January, and women in their final trimester of pregnancy, who can also receive the program.

21 In January 2015, Pennsylvania was fined $48.8 million for paying Medicaid, TANF, and SNAP to illegal immigrants. See Lydia Wheeler, "Pa. to pay $48.8M over benefits to illegal immigrants", The Hill, January 16, 2015. See also "Case Worker: Illegal Aliens Got Food Stamps by the 'Vanload'", Judicial Watch blog, April 30, 2013. See also "Former Kansas welfare worker: Illegal immigrants abused the system (Overwhelming fraud)" , Free Republic blog, March 14, 2011.

22 The funds from the program go to the hospital to offset the treatment of parents who cannot pay for their care. As PBS reported in 2013, the program costs about $2 billion a year, and the patients covered "mostly comprise illegal immigrants". The program goes beyond covering treatment in emergency rooms: "in New York, Emergency Medicaid may be used to provide chemotherapy and radiation therapy to illegal immigrants. In New York, California, and North Carolina, it may be used to provide outpatient dialysis to undocumented patients." Some illegal immigrants are certainly aware that Medicaid is paying for their treatment and report that on the SIPP. In this way, illegal immigrants are not formally enrolled in the traditional Medicaid program but report as such when asked by the Census Bureau.

23 Marianne Bertrand, Erzo Luttmer, and Sendhil Mullainathan, "Network Effects and Welfare Cultures", The Quarterly Journal of Economics, August 2000.