Hardship among Immigrants and the Native-Born in the United States

Abstract

This study examines the prevalence of several types of hardship among immigrants by race and ethnicity in the United States. Using data from the 2008 and 2014 Survey of Income and Program Participation panels and logistic regressions, I find that blacks, and to some extent Hispanics, are more likely to report hardships, such as bill paying and housing hardships, than whites and Asians, who are about equally as likely to report hardships. With regards to nativity and citizenship status, I find that immigrants who have become U.S. citizens are less likely to report some kinds of hardship than the native-born population. Undocumented immigrants, however, are more likely to report some kinds of hardships—particularly in the 2008 panel, which collected data closer to the aftermath of the Great Recession, which hit immigrants especially hard—but this relationship is explained by the lower incomes of these households in that panel. Results within racial and ethnic groups are generally (but not always) in the same direction, but are less frequently statistically significant. Overall, these findings suggest that immigrants are not particularly prone to hardship, especially when controlling for other characteristics. In fact, the lower likelihood of some hardships among foreign-born citizens suggests that they are positively selected, in that they may have unobserved characteristics that are protective, such as better health, stronger social networks, or money management skills. Because the foreign-born are less likely to be disadvantaged vis-à-vis the native born when examining hardship than when using the official income poverty measure, this study highlights the importance of using multiple measures when assessing the well-being of immigrants.

New immigrants to the United States often have relatively low levels of education and income. However, over time and across generations, many experience upward mobility, indicative of successful incorporation (Villarreal and Tamborini 2018). Even so, there is considerable variation across outcomes among immigrants and their descendants, with some groups, such as Hispanics, with levels of poverty that remain substantially higher than that among whites, while others, such as Asians, with poverty rates that are fairly similar to those of whites (Bean and Stevens 2003; Iceland 2017; Kasinitz et al. 2008; Sakamoto, Goyette, and Kim 2009).

While there is a considerable body of literature documenting income and poverty by race and nativity, there is much less research on experiences of hardship across these groups. Income is thought to be an important indicator of well-being because of its instrumental importance; that is, money can be used to purchase goods and services to meet basic needs and improve one’s welfare. In contrast, hardship measures are often thought to be outcomes of intrinsic importance (Beverly 2001). For example, individuals who report having insufficient food to eat, or having their utilities cut due to unpaid bills, or not seeing a doctor because of a lack of funds, are experiencing actual deprivation that is sometimes not captured when just measuring income. There is only a moderate correlation between measures of poverty and hardship, partly due to measurement issues, but also because they tap into different, if related, dimensions of well-being (Mayer and Jencks 1989; Iceland and Bauman 2007).

Even with this moderate correlation, because poverty generally is higher among the foreign-born than the native-born in the United States, we would expect that hardship would be more common among immigrants than the native-born as well. In addition, the fact that immigrants who are not citizens lack access to benefits from many programs might further increase hardship, and this issue would be all the more severe among those who are undocumented. Assimilation theory would predict that with rising incomes and greater access to formal institutions across generations, differences would narrow. Alternatively, one could hypothesize that immigrants will report lower levels of hardship than the native-born, especially if we condition on income, because immigrants are often positively selected. That is, they are healthier and may have other characteristics, such as ambition, that might mitigate hardship (Raleigh and Kao 2010).

Thus, the goal of this study is to examine the association between hardship and nativity to see if patterns that have been observed for poverty hold when looking at these intrinsically important outcomes. To this end, I use data from the 2008 and 2014 Survey of Income and Program Participation (SIPP), a nationally-representative panel survey that contains an extended battery of questions on hardships. The data reflect hardships as reported in 2010 and 2013 of the respective panels. I investigate whether reports of several hardships—including (depending on the panel) health, food, bill-paying, and housing hardships, as well ownership of consumer durables, neighborhood problems, and fear of crime—vary by nativity, citizenship and legal status, and race. In doing so, the broader aim is to shed light on the well-being of immigrants. To the best of my knowledge this is the first study to look at the association between this range of hardships and nativity, citizenship, and race using nationally-representative data in the United States.

Background

While patterns of income and poverty by race and nativity have been well-documented, we know considerably less about the incidence of hardship across these groups. Hardships are consumption-based indicators of well-being that are often thought to be superior to income-based measures (Citro and Michael 1995; Beverly 2001). Income measures do not always capture the resources families have to meet needs, such as some types of government transfers, wealth, and access to credit. In addition, as Pilkauskas, Currie, and Garfinkel (2013:403) argue, “Besides capturing the effects of economic resources that income-based measures may miss, consumption-based alternatives, such as hardship measures, are also heuristically attractive because they assess concrete adversities.”

There are a number of possible measures of hardship. Here I focus on seven types: health, food, bill-paying, housing, ownership of consumer durables, neighborhood problems, and fear of crime. These have been used by previous researchers examining the incidence of hardship (Beverly 2001; Heflin, Sandberg, and Rafail 2009; Heflin 2017; Iceland and Bauman 2007), and these indicators are all present in the Survey of Income and Program Participation. These tap into different dimensions of well-being, and have different associations with income. Health, food, and bill-paying hardship are more sensitive to short-term shortfalls in income, while the other four are more dependent on longer-term income (Iceland and Bauman 2007). For example, a job loss or health crisis might produce a short-term income drop that will result in a family having difficulty paying bills in a given month. This possibility has become more common in recent decades as the precarity of work has increased (Kalleberg 2009). That same family, however, may be living in a good neighborhood and has accrued a number of consumer durables over the years. Each are of interest in their own right.

Mechanisms for the link between nativity and hardship

There are many reasons to believe why levels of hardship vary by nativity. Assimilation theory asserts that new immigrants differ in many respects from the native-born population, including culturally and socioeconomically, but over time and across generations these differences narrow, resulting in successful incorporation (Alba and Nee 2003). Applying this approach to the current study, immigrants might be more likely to experience hardships because they have fewer resources, as exemplified by their lower median incomes and higher levels of poverty, than native-born households. For example, in 2017, the poverty rate among the native-born population was 11.0 percent, compared to 14.5 percent among the foreign-born (U.S. Census Bureau 2018a). This means that the foreign-born have less money to meet basic needs, such as food, clothing, shelter, and health expenses. Household income is important also because poorer people are more likely to live in neighborhoods with more affordable housing but worse conditions that could contribute to hardship, such as areas with more crime and environmental hazards and less social capital (Tiebout 1956; Epple and Platt 1998; Bischoff and Reardon 2014). Consistent with assimilation, poverty rates among the foreign-born vary by citizenship status, with citizens having a lower poverty rate (10.0 in 2017) than noncitizens (18.0 percent) (U.S. Census Bureau 2019). Thus, it will be important to examine the role of citizenship status in the analyses.

The segmented assimilation perspective holds that the extent of assimilation could vary across immigrant groups. Some groups might achieve successful incorporation into the mainstream, others may do well socioeconomically but maintain their ethnic distinctiveness, and yet others well experience downward mobility into the underclass (Portes and Zhou 1993; Zhou 1999). These different trajectories are a result of the existing racial hierarchy, maintained by discrimination, that produces unequal outcomes. Some point to large differentials in poverty by race and ethnicity as evidence of this hierarchy, with the lowest levels among whites (8.7 percent poor in 2017) and Asians (10.0 percent), but much higher among blacks (21.2 percent) and Hispanics (18.3 percent) (U.S. Census Bureau 2018b).

Characteristics of immigrants other than poverty can also affect assimilation trajectories. These could include English language proficiency, which is a common indicator of incorporation (Alba and Nee 2003). Also, a relatively high proportion of immigrants from Asia are admitted into the U.S. on the basis of their occupational skills, while a higher proportion of Hispanics enter because they already have kin living in the United States (Min 2006). Immigrants who enter due to their occupational skills have considerably higher levels of education and income on average than those who enter on the basis of family reunification provisions (Chiswick 1986; Feliciano 2005). This likely will results in lower levels of hardship among Asian immigrants and perhaps African immigrants (who also have relatively high levels of education) than Hispanic immigrants (Radford 2019). Relatedly, less-skilled workers, including many immigrants, are more likely to work in jobs that do not offer benefits that could reduce hardship (Kristal, Cohen and Navot 2018).

Overall, the literature on assimilation indicates that Asian Americans experience economic outcomes that are roughly on par with native-born whites (Kim and Sakamoto 2010; Park and Myers 2010; Kasinitz et al. 2008). The evidence concerning Hispanics is less clear. Hispanics achieve upward mobility across generations—for example the second generation have higher levels of educational attainment and income than the first—but they have not achieved parity with whites (Bean and Stevens 2003; Perlmann 2005). They may not achieve that much mobility beyond the second generation, though research on this issue is mixed (Telles and Ortiz 2008; Telles and Sue 2019; Duncan and Trejo 2011, 2014).

There are reasons less related to assimilation or segmented assimilation per se and more to policy that could affect patterns of hardship by nativity. Among these is that many immigrants may have less access to services that may improve well-being. The 1996 Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) and subsequent laws generally restricted benefits to “qualified” immigrants, or those who have legal permanent residency—and if they have lived in the U.S. for at least five years—or refugee status. Unqualified immigrants are ineligible for most government benefits, including cash welfare, Supplemental Nutrition Assistance Program (SNAP), Medicare, Medicaid, and Supplemental Security Income (Pew Charitable Trusts 2014). Take up rates are also lower for many who qualify for program assistance (Cunnyngham 2004). Participating in programs such as TANF, food assistance, and Medicaid likely reduce hardship, though some of the empirical literature (especially the earlier literature) is mixed because of the challenge in measuring selectivity into these programs (McKernan, Ratcliffe, and Iceland 2018; Pilkauskas, Currie, and Grafinkel 2012; Shaefer and Gutierrez 2013).

Undocumented immigrants may also lack access to bank accounts and drivers’ licenses, and often avoid talking with police to report a crime or speak with school officials if there is a problem with their children. They may be less aware of community programs and health services that have no citizenship requirements and may be less likely to seek help when needed because of their precarious status (Yoshikawa, Godfrey, and Rivera 2008; Bernstein et al. 2019; Yu et. al. 2005; Kalil and Chen 2008; Gelatt, Koball, and Bernstein 2018; Potochnick, Chen, and Perreira 2017). Immigrants who are not citizens also are generally ineligible to be employed by the Federal government.

On the other hand, there are reasons why immigrants might be less likely to experience hardships, especially if one controls for level of income. In particular, immigrants typically are positively selected on a number of traits, some observable and some not, such as health, skills, and ambition. There is a large literature on the “immigrant health paradox” that helps explain why immigrants, often with fewer financial resources, often display better health than the native-born population (Hummer et al. 2007; Jasso and Massey 2004; Martinez, Aguayo-Tellez, and Rangel-Gonzalez 2015). To extend this argument, it could be that these traits may provide immigrants with ways to better cope with potential hardships they may face. In addition, many immigrants may have networks of social support that that might have aided in the immigration process that potentially can mitigate the effects of economic strain.

There might be yet further selection among immigrants between those who have become naturalized citizens versus those who haven’t. Immigrants who become citizens have higher levels of schooling, duration in the U.S., and proficiency in English (Chiswick and Miller 2009). The naturalization process itself requires applicants to acquire basic knowledge of the English language and of U.S. history and government and pay an application fee. Moreover, applicants may be denied citizenship on the basis of certain criminal offenses and/or the failure to show that they are of “good moral character” (Bloemraad 2002). This provides another reason to distinguish between the outcomes of immigrants by citizenship status.

Empirical literature on the association between nativity and hardship

There have been relatively few studies that have examined the link between nativity and multiple dimensions of hardship, especially using nationally representative survey data. One study by Capps (2001) that focused on children of immigrants found that such children were more likely than native-born children to live in a household that experienced a few specific hardships, including a food-related hardship, bill-paying hardship, crowded housing, and a lack of health insurance. This study focused on bivariate relationships and did not condition on income. Another study that focused on food insecurity among children found that those with foreign-born noncitizen mothers experienced more persistent and higher levels of food insecurity than the native-born. However, low-income families in which foreign-born mothers are citizens have about the same risk of food insecurity compared to families with native-born mothers (Kalil and Chen 2008); this study’s findings are consistent with another that found that children of non-citizens in particular experience more persistence and higher levels of food insecurity than children of citizens (Van Hook and Balistreri 2006).

Gelatt, Koball, and Bernstein (2018) examine hardship among immigrants, though with a focus on how state enforcement policies affect patterns of hardship. They find no difference by nativity in a few different kinds of hardship, such as meeting expenses, food insecurity, and housing hardship, though immigrants are more likely to live in overcrowded households. In models with a variety of control variables, immigrants—both legal and unauthorized—may be slightly less likely to experience a variety of hardships, though, due to the structure of their models with interaction terms, it is not clear whether these net differences are statistically significant. The paper does not examine differences in the likelihood of hardship among immigrants of different origins.

In a study on the relationship between family structure and hardships that used immigrants status as a control variable, Lerman (2002) finds that recent immigrants were modestly more likely to report two kinds of hardships (inability to pay rent and missing meals for economic reasons) than the native born; the association persisted in models that controlled for income as well. Immigrants who had lived in the U.S. for longer periods of time were not significantly different than the native-born in their experiences of hardship. In contrast, another study that examined the effect of the Great Recession on hardship found that, when controlling for income (this control was in all models), immigrants were less likely to report hardship, measured with an index that included a number of different hardship indicators (Pilkauskas, Currie, and Garfinkel 2012). This paper’s focus was not on immigrants, and did not differentiate among immigrants by citizenship status. Other studies that have used nativity as a variable that did not differentiate by citizenship have found either mixed or null findings (Hernández et al. 2016).

These studies collectively suggest that immigrants who are non-citizens likely fare worse than those who are citizens and the native born. However, this literature is limited in a few ways. First, most studies looked at one, or perhaps two, hardships, such as food insecurity. The couple that looked at more hardships were descriptive (e.g., Capps 2001), did not condition on income or bundled hardships into a single index (Pilkauskas, Currie, and Garfinkel 2012), or did not examine the net effect of nativity or how effects might vary by race/ethnicity (Gelatt, Koball, and Bernstein 2018). The studies sometimes included controls for race, but did not have race-nativity interactions, which is important because of the disparities in well-being by race in the United States.

This paper builds on this literature by: examining seven types of hardship, including some analyses by race, differentiating between citizens, non-citizens, and by legal status, and conditioning some of the models on income to see if differences in the prevalence of hardship is being driven simply by differences in income or by other factors. Using multiple hardship measures is important because they tap into different dimensions of well-being. As noted earlier, some hardships are more sensitive to short-term income shortfalls (e.g., food insecurity), while others are more affected by longer-term income, such as housing problems and neighborhood conditions (Iceland and Bauman 2007). I also have data on program receipt (such as Supplemental Security Income) to see if receipt helps mediate the nativity-hardship relationship. Finally, I analyze data from one year, 2010, where the economy was still reeling from the Great Recession, and another from three years later, 2013, when the recovery was underway. Immigrants were especially hard hit during the recession, so examining immigrant well-being at these two time points is an advantage of this study (Bitler, Hoynes, and Kuka 2017). In summary, the goal of this analysis is to reach a better understanding of the extent to which immigrants experience various kinds of hardship, and whether income alone helps explain observed differences.

Data and Methods

I use data from the 2008 and 2014 panels of the Survey of Income and Program Participation (SIPP), a nationally-representative household survey conducted in the United States (U.S. Census Bureau 2001). The SIPP is longitudinal survey, where panels last from 3 to 5 years. It is a rich source of data on income, program participation, labor force activity, and is one of the relatively few surveys that collects information on experiences with various kinds of hardship. The data on hardships from pre-2014 panels come from the topical module on Adult Well Being, which was typically administered once per panel. Each wave in the SIPP covers a four month period. Specifically, I use data from the wave 6 topic module of the 2008 panel that collected information on hardship in 2010, and data on 2013 from the 2014 panel (this panel asked about hardships in the previous year). Each of these two panels has advantages and disadvantages. The 2008 panel has information on a wide range of hardships. After the 2008 SIPP panel, the SIPP was redesigned and shortened, and most of the topical modules were eliminated. As a result, the 2014 panel contains a much smaller set of hardship measures. However, the 2014 panel has the advantage of having a variable on immigrant year of entry, which could be of substantive importance, and being more current. It is also useful to have hardship measures from two time points with different economic conditions. Thus, I use data from both panels and draw upon their respective strengths.

The sample includes respondents who were in the SIPP survey during the wave that the topical module was administered and who provided valid answers to the hardship questions. Households are the unit of analysis, as hardships are reported for the household as a whole. The sample size is 34,850 in the 2008 panel data and 29,685 in the 2014 panel data. I use household weights provided by the SIPP for a given wave, as these are meant to ensure that the data are representative of all U.S. households in the given time period.

Measures of hardship

When using the 2008 SIPP, I analyze seven types of hardship. For each type, there are a series of questions, and I categorize a household as experiencing a hardship or not as a dichotomous outcome if they answer affirmatively to a certain number of questions, typically based on how previous studies have measured such hardships (Gelatt, Koball, and Bernstein 2018; Heflin 2016, 2017; Iceland and Bauman 2007; Short 2005) and yielding percentages that somewhat approximate poverty rates. The hardships are defined as follows in the 2008 panel:

1. Health hardship (one or more of the following): did not see or go to a doctor/hospital when needed care, did not see a dentist when needed care

2. Food hardship (two or more): food did not last (and didn’t have money for more), could not afford balanced meals, cut or skipped meals, ate less than should, did not eat for a full day

3. Bill-paying hardship (one or more): did not pay utility bill, phone disconnected, did not pay rent/mortgage

4. Housing hardship (one or more): pests, leaks, broken windows, plumbing problems, cracks in walls, holes in floor

5. Consumer durables (lacks five or more): computer, dishwasher, air conditioner, dryer, washer, microwave, cell phone, telephone, refrigerator, color television, VCR/DVD, stove, food freezer

6. Neighborhood problems (two or more): noise, street repair problems, trash/litter, abandoned buildings, would like to move, smoke/odors

7. Fear of crime (two or more): afraid to walk alone at night, stay at home for fear, goes out with others to stay safe, neighborhood is unsafe, carries something for protection, unsatisfied with crime, home is unsafe

Though the main analyses use these dichotomous variables, I also conducted a sensitivity analysis by creating counts of hardships for each dimension and running OLS models. These yielded similar findings, which are shown in appendix tables and discussed briefly at the end of the results section.

The 2014 panel has fewer measures of hardship. Specifically, it has no items at all for health hardships and consumer durables. It also has a fewer number of items for: food hardship (4 in 2014 vs. 5 in 2008), bill paying (1 vs. 3), housing (4 vs. 6), neighborhood problems (2 vs. 6), and fear of crime (2 vs.7). The summary indicators of hardship therefore incorporate different thresholds, including neighborhood problems and fear of crime (1 or more for each of these in 2014 vs. 2 or more in 2010). The wording on the some of the questions also differs slightly, including the accounting period for a few of them (e.g., previous year versus previous month). Thus, it is important to emphasize that the summary measures of the prevalence of hardships are not directly comparable across these two panels. Rather than focusing on levels of hardship in the two different periods, I examine differences in hardship across groups (nativity and race) for the measures available in the two SIPP panels.

Main independent variable: nativity, citizenship status, and documentation status by race

I examine differences in hardship by nativity, citizenship status, documentation status, and race. Nativity is measured by the place of birth, citizenship, and legal permanent status questions, and someone is categorized as native born if they were born in the United States, or if they were born abroad to American parents. Among those who are not native born, we can further distinguish between citizens, noncitizens with legal permanent status, and those without permanent status, who are considered undocumented. To be more precise, the “undocumented” group includes those who are undocumented, nonimmigrants and other non-green card people who are lawfully present, and some people who adjusted from undocumented or nonmigrant status to legal permanent status. So the group more accurately represents those who entered the United States without a green card.

I examine the role of nativity by race/ethnicity of householder, defined as non-Hispanic white, non-Hispanic black, non-Hispanic Asian, or Hispanic. While it would be optimal to have data on specific ethnic groups by nativity (such as Mexicans or Chinese), variables with this level of specificity are not available in the SIPP. Nevertheless, controlling for broad racial categories permits a more fine-grained analysis of hardship by nativity than not having such a variable at all because of at least some broad commonalities (in treatment and outcomes) among pan-ethnic groups (Iceland 2017).

Control variables

The analyses include a number of control variables in the models, including: among immigrants, year of entry (available in the 2014 panel but not the 2008 panel); household income-to-poverty ratio; age of the householder; education of the householder, defined as less than high school, high school diploma, 1–3 years of college, B.A. degree or more; household type, defined as married couple (with and without children), single female parent with children, other household type; employment status of householder, defined as employed full time, employed part time, unemployed, and out of the labor force; lives in a metropolitan area dummy variable; region, with the categories of Northeast, Midwest, South, and West; number of people in household; number of children under 18 present; the household has a person 65 years or older present; the household has a disabled individual present; and English language proficiency (well/very) or not. Since the literature suggests that access to and receipt of benefits might help explain potential differences in hardship by nativity, I also control for: receipt of either supplemental nutrition assistance (SNAP), TANF, General Assistance or Supplemental Security Income; Social Security; and whether has public health insurance or private health insurance.

Analytical Strategy

I begin by presenting descriptive statistics of hardships and summary hardship measures by nativity and race/ethnicity. The subsequent multivariate analysis consists of running a series of logistic regression models with each hardship as a separate dependent variable as specified by:

$$\text{Logit}\left(\text{P}\left(\text{Y}=1\right)\right)={\text{B}}_0+{\beta }_1{X}_1+{\beta }_2{X}_2+\cdots +{\beta }_k{X}_k$$ (1)

Specifically, the probability that a household experiences a hardship (Y) is modeled as a function of a series of covariates, including nativity (X₁), race (X₂), and the series of control variables described above. I also run hardship models by race to see if the role of nativity varies by race, as well as to allow the effects of other variables to vary by race. I run one set of models with nativity and race only, a second that adds household income-to-poverty ratio to see if income mediates the role of nativity, and finally one with the full set of controls to see if these other household characteristics help mediate the nativity-hardship relationship.

For brevity, the multivariate results focus on one short-term hardship, bill paying, and one longer-term one, housing hardship, as the result for others are generally consistent (exceptions are discussed in the text), and each outcome requires its own table. I include results for all of the other hardships in the appendix.

Results

Table 1 shows the percentage of respondents reporting specific hardships, and includes summary hardship indicators as well. As noted in the data and methods, there are fewer hardship measures in the 2014 SIPP panel than in the 2008 panel. Also as noted above, there are some differences in the wording and time frame for some of the hardship measures in the two panels, so most are not directly comparable over time. The main goal in this analysis is not to look at trends, but rather whether hardships vary by nativity in each of these two panels. The table indicates that many households experience hardships one type or another in the given year.

Table 1.

Percentage Reporting Material Hardships, 2010 and 2013

	2010	2013
Bill-paying hardship (one or more)	14.6	12.4
Did not pay utility bill	10.4	10.7
Phone disconnected	3.6
Did not pay mortgage/rent	7.9	7.3
Health hardship (one or more)	12.3
Did not see a dentist	9.6
Did not see a doctor	7.9
Food hardship (two or more)	10.9	12.9
Food did not last	13.5	14.8
Did not eat balanced meals	12.1	13.1
Skipped meals	5.1	8.2
Ate less than should	5.4	8.1
Did not eat whole day	1.4
Housing hardship (one or more)	14.1	16.7
Insect, pest problems	7.5	9.5
Roof leak	4.9
Broken windows	2.8
Plumbing problems	1.9	6.1
Cracks in wall	2.6	7.1
Holes in floor	0.7	1.4
Lack of consumer durables (five or more)	13.0
Computer	24.8
Dishwasher	30.6
Air conditioner	11.5
Dryer	16.8
Washer	14.7
Microwave	2.9
Cell phone	12.8
Telephone	25.0
Refrigerator	0.7
color tv	1.5
VCR/DVD	7.9
Stove	1.4
Food Freezer	62.1
Neighborhood problems (2+ in 2010, 1+ in 2016)	10.9	16.9
Noise problems	13.4	13.6
Street repair problems	12.0
Trash, litter	5.9	7.5
Abandoned buildings	7.1
Would like to move	4.7
Smoke, odors	2.9
Fear of crime (2+ in 2010, 1+ in 2016)	14.5	8.7
Afraid to walk alone at night	20.6
Stay at home for fear	10.5	5.5
Goes out with others	8.6
Neighborhood is unsafe	7.1	5.6
Carries something for protection	6.3
Would like to move due to crime	4.7
Home is unsafe	3.0
N	34,850	29,662

Sources: 2008 and 2014 SIPP panels

Table 2 shows how the summary hardship measures vary by race/ethnicity, nativity, citizenship status, and legal resident status. Among the total population, unauthorized immigrants were the most likely to report hardships, followed by those with legal permanent status. Foreign-born citizens were the least likely to report many kinds of hardship, though for some it was the native-born population, or little difference between the two. Foreign-born citizens, for example, were the least likely to report a bill-paying hardship in 2010 (11.9 percent), followed by the native-born population (14.3 percent), immigrants with legal permanent status (20.0 percent), and undocumented immigrants (23.7 percent). Lack of consumer durables stand out as a hardship that is much more common among the foreign-born of various statuses (16.0, 24.9, and 33.5 percent of citizens, those with legal permanent status, and unauthorized immigrants, respectively, report this hardship) than the native-born (11.7 percent). Thus, it appears that the foreign-born are less likely to prioritize the ownership of such consumer items, which arguably are less essential for well-being, and more likely (relative to the native-born) to address other kinds of needs. Much like with the official poverty rate, hardships are less common among immigrant citizens in all cases than those with legal permanent status or undocumented status; the multivariate analyses will test if these relationships hold after controls.

Table 2.

Hardships by Race, Ethnicity, and Nativity

	Bill-paying hardship		Health hardship		Food hardship		Housing hardship		Consumer durables		Neighborhood problems		Fear of crime
	2010	2013	2010	2013	2010	2013	2010	2013	2010	2013	2010	2013	2010	2013
Total population	14.6	12.4	12.3	NA	10.9	12.9	14.0	16.7	13.0	NA	10.9	16.9	14.5	8.7
Native-born	14.3	12.2	12.1	NA	10.4	12.7	14.0	16.7	11.7	NA	11.0	16.8	14.2	8.4
Foreign-born
Citizens	11.9	12.5	10.6	NA	11.2	11.4	12.1	14.8	16.0	NA	9.2	15.4	15.1	9.8
Legal permanent status	20.0	13.8	15.1	NA	16.8	16.9	15.6	18.5	24.9	NA	13.1	19.3	15.6	10.1
Unauthorized status	23.7	15.1	19.1	NA	19.8	18.6	19.3	21.3	33.5	NA	10.2	12.3	19.0	14.7
Whites	11.3	9.5	11.0	NA	8.5	10.4	12.5	14.5	9.3	NA	9.4	14.8	11.6	6.5
Native-born	11.3	9.6	11.1	NA	8.4	10.5	12.7	14.8	9.0	NA	9.4	14.9	11.5	6.5
Foreign-born
Citizens	10.2	7.7	9.2	NA	8.8	8.2	9.4	9.3	12.2	NA	7.0	10.0	13.0	4.4
Legal permanent status	14.9	4.9	11.9	NA	8.7	7.2	10.8	10.0	20.0	NA	10.2	15.7	10.6	4.9
Unauthorized status	9.7	5.5	13.5	NA	10.8	11.7	14.1	7.0	15.1	NA	7.4	19.5	12.2	11.0
Blacks	26.2	22.3	14.7	NA	17.9	20.9	17.7	21.7	21.5	NA	17.2	22.9	25.4	15.8
Native-born	27.0	22.5	15.1	NA	17.8	20.9	18.2	22.2	21.4	NA	17.9	23.5	26.5	15.9
Foreign-born
Citizens	15.6	22.9	9.1	NA	12.9	20.7	11.3	17.6	18.0	NA	10.1	19.6	12.7	14.2
Legal permanent status	26.4	13.2	13.2	NA	32.3	17.9	20.4	21.8	32.0	NA	17.0	15.1	22.0	17.6
Unauthorized status	38.0	20.1	22.3	NA	19.7	26.4	10.7	15.8	24.1	NA	8.4	15.4	26.5	14.5
Hispanics	21.8	18.6	16.9	NA	18.1	18.7	17.8	22.0	24.7	NA	13.0	22.4	20.3	13.3
Native-born	21.4	18.2	16.3	NA	17.2	17.7	17.4	21.8	21.0	NA	13.5	22.2	21.0	12.6
Foreign-born
Citizens	15.6	18.2	14.3	NA	15.0	15.0	14.7	19.6	20.6	NA	11.4	20.4	19.6	13.3
Legal permanent status	23,3	19.5	18.0	NA	19.3	23.2	17.2	22.7	26.6	NA	14.3	23.1	16.9	12.1
Unauthorized status	29.8	19.8	21.3	NA	24.1	23.4	23.2	26.1	41.5	NA	12.2	25.1	21.5	17.2
Asians	8.5	5.5	8.7	NA	8.2	5.4	13.1	13.5	16.7	NA	8.8	12.6	13.0	8.0
Native-born	7.7	5.6	5.6	NA	4.8	2.0	13.6	16.5	14.2	NA	9.0	11.0	12.2	7.3
Foreign-born
Citizens	8.0	6.0	8.8	NA	9.3	6.3	13.0	12.5	14.8	NA	9.0	13.2	13.4	8.9
Legal permanent status	11.9	5.9	10.6	NA	8.9	7.3	13.6	12.3	21.6	NA	9.7	12.6	13.7	5.4
Unauthorized status	8.2	2.9	12.4	NA	9.0	3.6	12.0	15.6	24.1	NA	6.1	11.9	12.1	8.3

Sources: 2008 and 2014 SIPP panels

With regards to race/ethnicity, the patterns generally confirm expectations, as blacks and Hispanics are more likely to report hardships than whites and Asians. Patterns by nativity, citizenship status, and race, however, are less easy to generalize. Among whites, blacks, and Hispanics, foreign-born citizens tend to be less likely to report hardships than the native born, with more mixed results among Asians. The unauthorized often are the most likely to report hardships, though this doesn’t hold as much for Asians, and some mixed patterns for whites and blacks. Overall, these findings by race and ethnicity do not provide as clear support for any single perspective, though for whites, blacks, and Hispanics in particular immigrants who are citizens on the whole appear to be selective, and this could explain low rates of hardship. For Asians, there may be a mix of factors at work, including differential selectivity and income/education differences by nativity. Citizens likewise might have lower levels of hardship than noncitizens because they also might have higher incomes more generally, or perhaps due to other characteristics. The following multivariate analysis sheds greater light on these issues.

Multivariate Analyses

Table 3 provides descriptive statistics for all of the independent variables in the analyses. About 87.2 percent of householders were native-born in 2010, another 7.2 percent were foreign-born citizens, 2.9 percent were legal permanent residents, and the remaining 2.8 percent were undocumented. The foreign-born groups as a share of the population were a little higher in 2013. In 2010, just over 70 percent of the householder were white (this share was lower in 2013) and the mean household income-poverty-ratio was 3.8, rising to 4.3 in 2013 during the recovery period after the deep recession in the late 2000s. Just under half of households were married-couple ones. A significant proportion of households reported receiving some kind of benefit, such as Social Security or public health insurance, nearly four fifths lived in a metropolitan area, and a plurality lived in the South.

Table 3.

Descriptive Statistics

	2010	2013
Nativity
Native-born	87.2	85.0
Foreign-born, citizen	7.2	8.6
Legal permanent status	2.9	3.4
Unauthorized status	2.8	3.0
Race/Ethnicity
Non-Hispanic white	70.6	67.4
Non-Hispanic black	12.0	12.8
Non-Hispanic Asian	3.2	4.7
Hispanic	11.7	12.9
Other race	2.5	2.2
Household income-to-poverty ratio	3.8	4.3
English language (well/very well)	96.0	98.1
Age	50.6	51.0
Education
Less than high school	10.8	11.1
High school	23.9	27.4
Some college	35.0	29.1
BA+	30.3	32.4
Family structure
Married couple family	49.4	47.7
Female-headed family	12.9	13.0
Other family type	37.8	39.3
Labor force status
Full-time employed	49.5	48.0
Part-time employed	13.2	14.0
Unemployed	4.5	3.5
Out of labor force	32.7	34.5
Household size	2.5	2.3
Children under 18 present	30.4	31.5
Person over 65 present	28.1	28.6
Disabled person present	19.8	23.5
Benefits (% receiving)
TANF/general assistance/SSI/housing assistance	3.9	6.4
Social Security	25.6	25.9
Private health insurance	70.0	64.9
Medicare/Medicaid	30.8	32.7
In metro area	78.9	79.6
Region
Northeast	18.3	18.2
Midwest	22.3	21.9
South	37.3	37.4
West	22.2	22.4

Sources: 2008 and 2014 SIPP panels. Note: individual attributes in the table refer to the those of the householder.

Table 4 shows results for logistic regressions where bill-paying hardship is the dependent variable, using data from the 2008 SIPP panel (reflecting hardship in 2010). The first set of models have results for the full sample, while the next four sets are results by race and ethnicity (whites, blacks, Asians, and Hispanics). Results from the full model indicate that foreign-born householders who have attained citizenship are less likely to report hardships than the native-born. Adding controls slightly weakens the effect, but it remains statistically significant in Model 3. The results in that model indicate that the odds that foreign-born citizens report a bill-paying hardship is 0.79 that of native-born households. Those with legal permanent status do not significantly differ from the native born in any of the models. Those with unauthorized status are more likely to experience a bill-paying hardship according to results in model 1 without controls (1.30 times more likely), but this relationship becomes smaller and not significant in Model 2 when controlling for household income-to-poverty ratio.

Table 4.

Logistic Regressions. Bill Paying Hardship (Odds Ratios) for 2010

	Full Sample						Whites						Blacks						Asians						Hispanics
	Model 1		Model 2		Model 3		Model 1		Model 2		Model 3		Model 1		Model 2		Model 3		Model 1		Model 2		Model 3		Model 1		Model 2		Model 3
Nativity
Native-born (omitted)
Foreign-born citizen	0.70	***	0.71	***	0.79	**	0.89		0.81		0.89		0.50	***	0.56	**	0.68	*	1.04		1.03		0.82		0.68	**	0.66	**	0.78
Legal permanent status	1.15		0.98		0.87		1.37		1.21		0.90		0.97		0.96		1.04		1.62		1.38		1.17		1.12		0.91		0.86
Unauthorized status	1.30	**	1.10		0.91		0.78		0.71		0.46	**	1.03		0.99		1.09		1.03		0.92		1.07		1.52	**	1.23		1.08
Race/Ethnicity
Non-Hispanic white (omitted)
Non-Hispanic black	2.80	***	2.11	***	1.66	***
Non-Hispanic Asian	0.80		0.82		0.78
Non-Hispanic other race	2.67	***	2.20	***	1.72	***
Hispanic	2.16	***	1.62	***	1.12
Household income-to-poverty ratio			0.73	***	0.83	***			0.72	***	0.82	***			0.78	***	0.87	***			0.74	***	0.79	***			0.77	***	0.88	**
English language (well/very well)					1.01						1.25						1.65						0.70						0.91
Age					0.99	***					0.99	***					1.00						1.00						0.99	**
Education
Less than high school (omitted)
High school					1.10						0.95						1.20						1.79						1.20
Some college					1.15	*					0.97						1.43	**					2.86	*					1.18
BA+					0.71	***					0.57	***					1.25						1.29						0.74
Family structure
Married couple family (omitted)
Other family type					1.10						1.07						1.18						1.48						1.03
Female-headed family					1.48	***					1.58	***					1.42	**					1.03						1.44	**
Labor force status
Full-time employed (omitted)
Unemployed					1.46	***					1.38	**					1.60	**					2.36	*					1.52	*
Part-time employment					1.17	**					1.07						1.37	*					0.87						1.36	*
Out of labor force					0.74	***					0.67	***					1.01						0.31	**					0.90
Household size					1.09	***					1.10	**					1.09	*					1.32	**					1.06
Children under 18 present					1.04						1.06						1.00						0.94						1.08
Person over 65 present					0.68	***					0.63	***					0.81						1.25						0.67	**
Disabled person present					2.17	***					2.56	***					1.46	***					1.33						1.82	***
Benefits
TANF/general assistance/SSI/housing assistance					1.05						1.05						1.09						1.60						1.01
Social Security					0.65	***					0.61	***					0.58	***					0.96						0.91
Private health insurance					0.50	***					0.42	***					0.67	***					0.56						0.69	**
Medicare/Medicaid					1.01						1.01						1.14						1.31						1.01
In metro area					1.06						1.08						1.10						0.69						1.04
Region
Northeast (omitted)
West					1.18	**					1.14						1.12						1.11						1.25
Midwest					1.15	*					1.06						1.62	**					1.56						1.01
South					0.99						0.90						1.28	*					1.06						1.03
N	34,850		34,850		34,850		25,059		25,059		25,059		4,325		4,325		4,325		1,177		1,177		1,177		3,245		3,245		3,245

^***p<.001

^**p<.01

^*p<.05

With regards to race and ethnicity, blacks are more likely to report bill-paying hardship than whites in all models. Hispanics are also more likely to report hardships than whites, though this relationship becomes nonsignificant in the final model with controls. There is no significant difference between Asians and whites. With regards to race-specific models, consistent with the descriptive table (Table 2), results in Table 4 vary somewhat. Among whites, there are few differences by nativity and citizenship, except in Model 3 when all controls are included, and this model indicates that undocumented immigrants actually are less likely to experience a hardship, suggesting that there might be some positive selectivity with regard to unobservable characteristics. Among blacks, foreign-born citizens are less likely to report bill hardship, even in models with controls, though the size and significance of the coefficient is moderately reduced in model 3. Both types of noncitizens do not differ from the native-born in any model. Among Asians there are no statistically significant findings. The results for Hispanics are similar to that of the main sample, as undocumented immigrants are more likely to report a bill-paying hardship, but this relationship becomes nonsignificant once we control for household income-to-poverty ratio. Foreign-born citizens are less likely to report hardships than the native-born, and this relationship becomes nonsignificant in the model with all of the controls. Overall, the finding that foreign-born citizens are less likely to report hardship in models without controls as well as in many models with them suggests that immigrant are often positively selected on characteristics that are both observed and unobserved, and this confers advantages that make hardships less likely among this group.

Control variables generally have expected associations with bill-paying hardship, as income is negatively associated with hardship, as is age in many of the models. Female-headed households are generally more likely to report hardships, as are larger households and those with a disabled member present. Householders who are unemployed or employed part-time are generally more likely to report hardships than those employed full-time, with those out of the labor force (perhaps by choice) are less likely. The receipt of welfare is generally not associated with hardship, while those who received Social Security income and those with private health insurance are less likely to report hardships in models where they are significant.

Table 5 shows results where housing hardship in 2010 is the dependent variable. The results for this hardship is similar to that for bill-paying hardship, with a few differences. For the full-sample and among whites and blacks, foreign-born citizens are less likely than the native born to report a housing hardship. Among blacks, this relationship becomes not significant in Model 3 with the full set of controls (though the odds ratio for whites and blacks in their respective Model 3s are nearly identical). The relationship between hardship and foreign-born citizens is in the same direction for Asians and Hispanics, but all of the coefficients are not significant. In addition, for all groups and models, both kinds of noncitizens do not differ from the native-born in their odds of reporting a housing hardship, with one exception: among Hispanics, undocumented immigrants are more likely to report a hardship than the native born, though this relationship becomes nonsignificant once we control for income-to-poverty ratio. The odds ratios for control variables in these models are generally consistent with those in Table 4. Overall, the results from Table 5 indicate a moderate association between nativity and hardship for the full sample, whites, and blacks, suggestive that foreign-born citizens are positively selected on traits that are protective of hardship. Among Hispanics, the finding that the undocumented are more likely to experience housing hardship is mainly due to their low income. More generally, blacks are more likely to report a housing hardship than whites in all models. Hispanics once again are a more likely to report hardships than whites, though this relationship becomes nonsignificant in the final model with controls. There is no significant difference between Asians and whites in Models 1 and 3.

Table 5.

Logistic Regressions. Housing Hardship (Odds Ratios), for 2010

	Full Sample						Whites						Blacks						Asians						Hispanics
	Model 1		Model 2		Model 3		Model 1		Model 2		Model 3		Model 1		Model 2		Model 3		Model 1		Model 2		Model 3		Model 1		Model 2		Model 3
Nativity
Native-born (omitted)
Foreign-born, citizen	0.74	***	0.74	***	0.76	***	0.71	**	0.70	**	0.68	**	0.57	**	0.62	*	0.67		0.95		0.94		0.99		0.82		0 .81		0.84
Legal permanent status	0.94		0.89		0.83		0.84		0.83		0.71		1.15		1.15		1.29		1.00		0.96		1.03		0.99		0.85		0.85
Unauthorized status	1.16		1.09		0.99		0.93		0.93		0.79		0.78		0.76		0.95		0.87		0.85		0.92		1.41	**	1.19		1.17
Race/Ethnicity
Non-Hispanic white (omitted)
Non-Hispanic black	1.52	***	1.34	***	1.17	**
Non-Hispanic Asian	1.19		1.23	*	1.16
Non-Hispanic other race	2.13	***	1.95	***	1.64	***
Hispanic	1.55	***	1.36	***	1.12
Household income-to-poverty ratio			0.91	***	0.96	***			0.92	***	0.96	***			0.85	***	0.92	*			0.96		1.00				0.83	***	0.89	**
English language (well/very well)					0.98						1.06						3.26						0.93						0.97
Age					1.00	*					1.00						1.00						0.99						1.00
Education
Less than high school (omitted)
High school					0.87	*					0.91						0.96						0.69						0.78
Some college					0.89	*					0.94						0.87						0.95						0.82
BA+					0.86	*					0.93						0.81						0.78						0.80
Family structure
Married couple family (omitted)
Other family type					1.38	***					1.38	***					1.17						2.11	**					1.46	**
Female-headed family					1.39	***					1.44	***					1.12						1.55						1.54	**
Labor force status
Full-time employed (omitted)
Unemployed					1.23	**					1.24	*					1.56	*					0.91						1.07
Part-time employment					1.26	***					1.22	**					1.57	**					1.33						1.15
Out of labor force					1.17	**					1.08						1.54	**					0.89						1.20
Household size					1.12	***					1.12	***					1.15	**					1.25	**					1.07
Children under 18 present					0.93	**					0.95						0.82	**					0.91						0.95
Person over 65 present					0.93						0.87						1.03						0.82						1.05
Disabled person present					1.63	***					1.73	***					1.37	**					1.31						1.40	**
Benefits
TANF/general assistance/SSI/housing assistance					1.10						0.94						1.12						1.36						1.36
Social Security					0.85	*					0.85	*					0.94						0.95						0.79
Private health insurance					0.78	***					0.75	***					0.83						0.78						0.96
Medicare/Medicaid					1.00						1.06						0.93						1.20						1.05
In metro area					0.93						0.95						1.05						0.45	*					0.99
Region
Northeast (omitted)
West					1.07						1.12						0.84						1.21						1.14
Midwest					0.86	**					0.86	*					0.69	*					1.19						0.87
South					0.90	*					0.84	**					0.87						1.05						1.26
N	34,850		34,850		34,850		25,059		25,059		25,059		4,325		4,325		4,325		1,177		1,177		1,177		3,245		3,245		3,245

^***p<.001

^**p<.01

^*p<.05

Appendix Tables A1–A5 show results from these models using alternative hardship outcomes. They are generally consistent with Tables 4 and 5, as foreign-born citizens are either less likely to experience hardships or the differences with the native born is not significant. With some outcomes (especially the two more affected by short-term income flows, food and health hardships) undocumented immigrants are more likely to experience hardships, which is attenuated with the addition of controls. The consumer durable outcome is in some respects an outlier, as both types of noncitizens are considerably more likely to experience a dearth of consumer durables when compared to the native-born population, even with controls, perhaps suggesting that these immigrants may have less of a taste for consumer durables or are more likely to save or send remittances than the native-born population. Legal permanent residents are also less likely to report fear of crime, in models with controls only, perhaps suggesting that they often seek out safe neighborhoods (or live in ethnic neighborhoods with relatively high levels of trust). In most of the other models, legal permanent residents don’t differ from the native-born. In models with controls, blacks are more likely to report all hardships than whites, Hispanics are more likely to report some hardships than whites, but for others differences are not significant. In most models whites and Asians do not differ in their likelihood of hardship.

Results using the 2014 SIPP data (with hardships reported in 2013) share some similarities with those from the 2008 panel, though with some differences as well. With regards to bill paying hardship (Table 6), among the population as a whole, there is little difference the likelihood of hardship by nativity and citizenship status in models without controls, though in models with controls both types of non-citizens are less likely to report hardship. This is somewhat different than what we saw in 2010, where unauthorized immigrants were more likely to report a bill-paying hardship in models without controls, but not so with controls. Thus, the 2010 results suggest that in that year undocumented immigrants had low incomes that explained hardship, while the 2013 findings suggests that once we control for income, there may be positive selection based on unobservable characteristics. This could be a function of the Great Recession having a particularly negative impact on immigrants, and perhaps especially undocumented one in industries hard hit by the recession like manufacturing and construction (Bitler, Hoynes, and Kuka 2017; Kochhar 2019). Unlike in 2010, there is no statistically significant difference in bill-paying hardship between foreign-born citizens and the native-born. For most of the specific race groups, the nativity and citizenship indicators are not significant, perhaps in part due to the somewhat smaller sample size in 2014 than 2008, which can be consequential for examining hardship among relatively small racial/nativity groups.

Table 6.

Logistic Regressions. Bill Paying Hardship (Odds Ratios) fur 2013

	Full Sample						Whites						Blacks						Asians						Ilispanics
	Model 1		Model 2		Model 3		Model 1		Model 2		Model 3		Model 1		Model 2		Model 3		Model 1		Model 2		Model 3		Model 1		Model 2		Model 3
Nativity
Native-born (omitted)
Foreign-born, citizen	0.92		0.93		0.98		0.78		0.77		0.85		1.02		1.12		1.00		1.09		1.02		0.98		1.00		1.00		1.02
Legal permanent status	0.88		0.79	*	0.69	**	0.48		0.44		0.46		0.53		0.50		0.53		1.06		1.00		0.92		1.09		0.90		0.82
Unauthorized status	094		0.80	*	0.67	**	0.55		0.48		0.41		0.87		0.85		0.79		0.51		0.46		0.59		1.11		0.89		0.80
Race/Ethnicity
Non-Hispanic white (omitted)
Non-Hispanic black	2.74	***	2.16	***	1.70	***
Non-Hispanic Asian	0.60	**	0.60	**	0.61	**
Non-Hispanic other race	2.60	***	2.11	***	1.62	***
Hispanic	2.26	***	1.81	***	1.32	***
Household income-to-poverty ratio			0.81	***	0.89	***			0.80	***	0.89	***			0.83	***	0.86	***			0.92	*	0.96				0.78	***	0.86	***
English language (well/very well)					1.08						1.99						1.60						2.07						1.01
Age					1.00						0.99	**					1.01	*					1.01						1.00
Education
Less than high school (omitted)
High school					0.82	**					0.83						0.82						0.68						0.89
Some college					0.90						0.91						1.09						0.37						0.90
BA+					0.57	***					0.50	***					0.88						0.60						0.66	*
Family structure
Married couple family (omitted)
Other family type					1.27	***					1.36	**					0.91						2.03						1.23
Female-headed family					1.58	***					1.78	***					1.28						2.60	*					1.42	**
Labor force status
Full-time employed (omitted)
Unemployed					1.71	***					1.97	***					1.33						1.73						1.45
Part-time employment					1.21	**					1.21	*					1.03						1.47						1.40	*
Out of labor force					0.72	***					0.71	***					0.56	***					0.94						0.96
Household size					1.13	***					1.13	***					1.02						1.47	**					1.17	***
Children under 18 present					1.14						1.27	*					1.04						0.37	*					1.14
Person over 65 present					0.60	***					0.56	***					0.68	*					0.81						0.64	**
Disabled person present					2.18	***					2.27	***					1.93	***					1.34						1.98	***
Benefits
TANF/general assistance/SSI/housing assistance					1.11						1.20						1.13						1.40						1.26
Social Security					0.74	***					0.68	***					0.77						0.97						0.86
Private health insurance					0.62	***					0.54	***					0.70	**					0.45	*					0.91
Medicare/Medicaid					1.01						1.18						0.90						0.51						0.98
In metro area					0.99						0.97						1.01						0.69						1.17
Region
Northeast (omitted)
West					0.80	**					0.73	**					0.53	**					0.99						0.98
Midwest					0.77	***					0.67	***					0.91						0.99						0.89
South					0.76	***					074	***					0.75	*					1.26						0.77
N	29,685		29,685		29,685		19,906		19,906		19,906		4,340		4,340		4,340		1,029		1,029		1,029		3,686		3,686		3,686

^***p<.001

^**p<.01

^*p<.05

The patterns for housing hardship are more similar across the two SIPP panels. In 2013, like in 2010, foreign-born citizens were less likely to report a housing hardship in all models (see Table 7). The pattern is apparent for all racial/ethnic groups, though only statistically significant among whites. Findings within other racial/ethnic groups tend to be in the same direction in 2013 as in 2010, but none of them are statistically significant. Results for 2013 with other hardship outcomes are in Appendix Tables A6 to A8. More generally, in 2013 blacks and Hispanics are more likely to report most hardships than whites, while Asians are either as likely or less likely than whites to report such hardships.

Table 7.

Logistic Regressions. Housing Hardship (Odds Ratios), fur 2013

	Full Sample						Whites						Blacks						Asians						Hispanics
	Model 1		Model 2		Model 3		Model 1		Model 2		Model 3		Model 1		Model 2		Model 3		Model 1		Model 2		Model 3		Model 1		Model 2		Model 3
Nativity
Native-born (omitted)
Foreign-born, citizen	0.74	***	0.75	***	0.82	*	0.59	**	0.60	**	0.67	*	0.75		0.80		0.81		0.73		0.70		0.70		0.88		0.87		0.94
Legal permanent status	0.88		0.84		0.84		0.64		0.64		0.63		0.98		0.94		1.09		0.71		0.69		0.75		1.05		0.93		0.99
Unauthorized status	1.02		0.96		0.92		0.43	*	0.41	*	0.40	*	0.66		0.65		0.67		0.94		0.89		1.00		1.27		1.09		1.11
Race/Ethnicity
Non-Hispanic white (omitted)
Non-Hispanic black	1.65	***	1.48	***	1.18	**
Non-Hispanic Asian	1.09		1.12		1.09
Non-Hispanic other race	2.57	***	2.34	***	1.89	***
Hispanic	1.78	***	1.60	***	1.28	***
Household income-to-poverty ratio			0.93	***	0.98	**			0.94	***	0.98	*			0.86	***	0.92	**			0.96	*	1.00				0.86	***	0.94	*
English language (well/very well)					1.06						1.39						0.77						0.97						1.15
Age					0.99	***					0.99	**					1.00						1.00						0.99
Education
Less than high school (omitted)
High school					0.82	***					0.84	*					0.77	*					0.72						0.94
Some college					0.84	**					0.90						0.75	*					0.86						0.94
BA+					0.68	***					0.77	**					0.54	***					0.51						0.64	*
Family structure
Married couple family (omitted)
Other family type					1.62	***					1.67	***					1.25						1.40						1.68	**
Female-headed family					1.54	***					1.55	**					1.26						0.74						1.81	***
Labor force status
Full-time employed (omitted)
Unemployed					1.39	***					1.47	**					1.37						0.70						1.13
Part-time employment					1.16	**					1.04						1.22						2.01	*					1.53	**
Out of labor force					0.97						0.97						0.90						1.30						0.91
Household size					1.10	***					1.11	**					1.05						1.06						1.14	**
Children under 18 present					0.90						0.91						0.86						1.05						0.77
Person over 65 present					0.89						0.91						0.91						0.99						0.91
Disabled person present					2.02	***					2.20	***					1.54	***					1.55						2.06	***
Benefits
TANF/general assistance/SSI/housing assistance					1.09						1.22	*					0.96						0.98						0.98
Social Security					0.78	***					0.85						0.76						0.55						0.71
Private health insurance					0.78	***					0.75	***					0.99						0.77						0.82
Medicare/Medicaid					1.09						0.99						1.35	*					1.03						1.16
In metro area					1.06						1.04						0.99						4.90	*					1.16
Region
Northeast (omitted)
West					1.00						1.21	*					0.75						1.00						0.59	***
Midwest					0.90						1.06						0.79						0.96						0.47	***
South					1.00						1.19	*					0.81						1.28						0.58	***
N	29,685		29,685		29,685		19,906		19,906		19,906		4,340		4,340		4,340		1,029		1,029		1,029		3,686		3,686		3,686

^***p<.001

^**p<.01

^*p<.05

Sensitivity Analyses

I conducted additional analysis in 2013 where I substitute year of entry categories for the various citizenship and documentation statuses (these data are not available in the 2008 SIPP panel). I don’t include both the year of entry and the citizenship variables in the same models because of collinearity between the two sets of variables. The results are very consistent with each other (see Appendix Table A9). With regards to bill paying hardship, results indicate that the most recent immigrants (arrived less than five years and 5–9 years) are less likely to report hardships than the native-born, much as in Table 6, where legal permanent residents and undocumented immigrants were less likely to report bill paying hardship than the native-born. An analogous pattern for housing hardship is observed, where here it is both foreign-born citizens (odds ratio 0.82 in full model in Table 7) and long-term immigrants (0.82 in Table A9) who are significantly less likely than the native-born to report such a hardship. From these results, it is not clear if it is mainly documentation status per se that explains the variation or duration in the United States, or the combination of both since they are correlated with one another.

Finally, I conducted a robustness check by not using dichotomous hardship indicators, but instead counts of hardships for each dimension, and running OLS models with those counts as the dependent variables. Results from these models, shown in Appendix Table A10, are very similar to those from the main logistic regressions. Specifically, in 2010, foreign-born citizens remain less likely to report both bill-paying and housing hardships, even with all of the controls. Undocumented immigrants are more likely to report bill-paying hardship in both, with the association becoming nonsignificant with all controls (though in the logistic regression, the income-to-poverty threshold variable alone explained the association). There is no significant relationship between the noncitizenship categories and housing hardship in both the logistic and OLS models.

Conclusion

The economic well-being of immigrants is an issue of broad concern, and it is indicative of the extent of their incorporation in their new countries and communities. In this study I focus on hardship among immigrants in the United States. Using data from the 2008 and 2014 SIPP panels (reflecting hardship in 2010 and 2013) and logistic regressions, I find that immigrants who have become U.S. citizens tend to be less likely to report some kinds of hardship than the native-born population. I also find that the foreign born who are undocumented are more likely to report some of the hardships (especially in 2010 but not so in 2013) and this is explained mainly by the lower incomes of these households. I also examined whether results vary across racial and ethnic groups. These results often are (though not always) in the same direction, but are less frequently statistically significant. The smaller sample sizes for some specific groups could play a role. More generally with regards to race and ethnicity, blacks are more likely to report various hardship than whites. Hispanics are also more likely to report hardships than whites in all bivariate models, though this relationship becomes nonsignificant for some hardships in models with controls, especially in 2010. More often than not there is no significant difference between Asians and whites.

Previous empirical studies on the association between nativity and hardship are mixed. Some of these looked only at the effect of nativity overall and/or one or two hardships. I find that it is important to differentiate the foreign-born by nativity, citizenship status, and, if possible, documentation status, and the findings are consistent with previous ones that had found that noncitizens, and the undocumented in particular, were in some conditions more like to experience hardship (Kalil and Chen 2008; Van Hook and Balistreri 2006). I find that this mainly occurred in 2010 and among hardships more strongly associated with short-term income flows (bill paying, food, and health hardships), rather than those related to longer-term income flows (Iceland and Bauman 2007), and the findings in fact show that the differences are explained by lower incomes among the undocumented. These immigrants might be incurring greater short-terms hardship in order to live in otherwise more equal neighborhoods. Of note, however, the undocumented in the multivariate analyses were disadvantaged mainly in 2010 and not in 2013. This likely reflects previous findings that immigrants were hit particularly hard by the 2007–2009 Great Recession, though their status improved in the recovery (Bitler, Hoynes, and Kuka 2017; Kochhar 2019). This finding thus indicates that economic shocks often may have a larger effect on immigrants, and the undocumented in particular, due to their more precarious social, economic, and political position.

The finding that foreign-born citizens are less likely to report hardships than the native-born suggests that immigrants frequently are positively selected on a number of traits, some observable and some not, such as health, skills, ambition, and networks of support than can ameliorate the effects of financial strain. Thus, after an initial adjustment period, these immigrants who also are able to meet citizenship requirements are in fact better off than the native born. This notion is also consistent with research which finds an “immigrant health paradox”—where immigrants display better health than the native born (Hummer et al. 2007; Jasso and Massey 2004; Martinez, Aguayo-Tellez, and Rangel-Gonzalez 2014).

The findings are not consistent with perspectives that predict a higher likelihood of hardship among immigrants than the native born due to factors such as less access to government benefits (Pew Charitable Trusts 2014) or work-related benefits (Kristal, Cohen and Navot 2018). Even among noncitizens (who are not eligible for many kinds of government support), income alone explains differences in hardships when they occur. This is not to say that such benefits would have no effect if received (Social Security in general, for example, is associated with lower hardship among respondents). It could very well be that immigrants would be more advantaged than otherwise similar native born households (due to selection) than shown if they had more access to benefits, but I cannot definitively determine this with the data available.

These findings are partially consistent with assimilation theory (Alba and Nee 2003), especially since I find that immigrants who have become citizens—and thus are more acclimated to the United States—report lower levels of some hardship than immigrants with legal permanent status and unauthorized immigrants. Undocumented immigrants exhibit the highest levels of hardship. However, the finding that foreign-born citizens have lower levels of hardship than the native born is not consistent with assimilation, and thus is likely explained by the selection argument described above. To the extent that levels of hardship, much like poverty, vary by race and ethnicity, the findings also provide support to segmented assimilation (Portes and Zhou 1993). Whites and Asians tend to experience similar levels of hardship, consistent with other literature comparing whites and Asians (Kim and Sakamoto 2010; Park and Myers 2010; Kasinitz et al. 2008; Iceland 2019), while blacks, and sometimes Hispanics, are more likely to report hardships, even when controlling for a variety of characteristics (Telles and Ortiz 2008; Telles and Sue 2019). However, we don’t see vast differences in the role that nativity and citizenship variables play across groups; more often than not, group-specific coefficients are not significant, perhaps in part due to sample size.

The findings also indicate that it is important to examine a variety of outcomes when assessing well-being. Poverty, broadly speaking, is by its nature a multidimensional concept; it is not just about the ability to purchase things, but also to meaningfully participate in society and realize one’s capabilities (Sen 1999). An advantage of hardship measures over traditional income poverty measures, such as the official U.S. poverty measure, is that they measure concrete challenges households face, such as not being able to pay bills or having substandard housing (Pilkauskas, Currie, and Garfinkel 2013; Beverly 2001; Heflin 2017). In contrast, income poverty measures, while not uninformative, are proxies for well-being—income is instrumentally important for providing people the ability to meet basic needs, such as housing, food, and paying other bills, with the income they have available.

The results suggest that the well-being of immigrants vis-à-vis the native born appears somewhat worse when using an income poverty measure than hardship measures. For example, while the gap in the official poverty rate was 4.0 percentage points in 2013 (14.3 percent among the native born, compared to 18.3 percent among the foreign born (U.S. Census Bureau 2018a)), the nativity gap for all hardships measured in that year were smaller. As noted above, this could be due to stronger networks, better health, ambition and/or other characteristics that allow them to manage their resources better than the native born. One clue supporting the last of these conjectures is that the one hardship that the foreign-born are much more likely to report is the lack of consumer durables. This suggests that the foreign-born are more likely to forgo consumer items—many which are non-essential—than experience other, what many would consider more serious, kinds of hardship, such as bill paying and housing hardships. In short, these findings support the notion that we need to use a variety of measures to arrive at a more complete and holistic understanding of immigrant well-being.

Data Availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.