Housing and neighborhood quality among undocumented Mexican and Central American immigrants

Abstract

Extensive research has documented the challenges that undocumented immigrants face in navigating U.S. labor markets, but relatively little has explored the impact of legal status on residential outcomes despite their widespread repercussions for social well-being. Using data from the 1996–2008 panels of the Survey of Income and Program Participation to impute documentation status among Mexican and Central American immigrants, we examine group differences in residential outcomes, including homeownership, housing crowding, satisfaction with neighborhood and housing quality, problems with neighborhood crime/safety, governmental services, and environmental issues, and deficiencies with housing units. Results from our analysis indicate that undocumented householders are far less likely to be homeowners than documented migrants, and also live in more crowded homes, report greater structural deficiencies with their dwellings, and express greater concern about the quality of public services and environmental conditions in their neighborhoods. In comparison to native whites, undocumented migrants’ residential circumstances are lacking, but their residential outcomes tend to be superior to those of native-born blacks. Overall, our results highlight the pervasive impact of legal status on stratifying Mexicans’ and Central Americans’ prospects for successful incorporation, but also underscore the rigidity of the black/nonblack divide structuring American residential contexts.

1. Introduction

The demography of the American population has been permanently transformed as a result of large-scale immigration over the past four decades. Since 1970 more than 50 million immigrants have come to America in search of well-paying jobs, nice homes and neighborhoods, and the opportunity to improve their lives and those of their children. America’s newest members include those from all parts of globe, of varying racial/ethnic backgrounds, and with a diverse set of skills and resources, but no group is as substantial in size as those from Mexico and neighboring countries. At least one-third of all arrivals to the U.S. since 1970 have hailed from Mexico or Central America, and their geographic dispersion in recent years to areas with little prior history of immigration means that their presence is altering social relations in a wide mix of communities. While this population has made considerable inroads in gaining access to American economic, educational and other social institutions, its ultimate incorporation into the mainstream is hampered by that fact that half of Mexican and Central American immigrants lack legal authorization to live and work in the U.S. (Hoefer et al., 2011; Passel and Cohn, 2011).

Better understanding the comprehensive impacts of legal status of Mexican and Central American immigrants’ success in America is of considerable importance given not just because their sheer size, but due to their unique contribution to America’s economic, social, and political future. Considerable research has explored how undocumented status impedes access to competitive jobs and safe work environments, obstructs academic achievement and educational completion, and limits the health and well-being of individual migrants (Donato and Massey, 1993; Flippen, 2012; Greenman and Hall, 2013; Hall et al., 2010; Kaushal, 2006; Kossoudji and Cobb-Clark, 2002; Massey, 1987; Rivera-Batiz, 1999; Suárez-Orozco et al., 2011) Comparatively lacking is research evaluating the impacts of legal status on residential attainment and quality. Such research is crucial not only because of the important role that residential location plays in structuring life chances, but because of persistent and deep racial/ethnic inequalities in residential outcomes (Rosenbaum and Friedman, 2006; Sampson, 2012).

For several decades, researchers have analyzed the historical and contemporary barriers making minority families prone to residence in dilapidated homes and in poor neighborhoods with weak institutional supports, heightened crime, and other social ills (Friedman and Rosenbaum, 2004; Krivo and Kaufman, 2004; Rosenbaum, 1996). This work consistently documents a rigid racial/ethnic hierarchy – one in which whites traditionally hold position on top, and blacks on the bottom – yet the influx of millions of low-skilled and unauthorized immigrants into communities raises the question of where Mexican and Central American immigrants fit into this racial/ethnic structure. Answering this question is crucial not only to understand the challenges that unauthorized migrants may face in securing housing and residence in desirable neighborhoods, but because it informs discussions regarding the emerging shape of the American colorline.

Our goal in this study is to provide some evidence on these topics by estimating the impact of legal status on Mexican and Central American immigrants’ residential attainment and by comparing their outcomes to three broader racial/ethnic groups: native Latinos, native whites, and native blacks. In contrast to related work, we draw a large, nationally-representative sample of U.S. householders from the Survey of Income and Program Participation spanning the 1998–2010 period and use detailed information on citizenship, visa status, and public program participation to impute legal status for immigrant householders. Moving beyond the conventional practice of using homeownership to proxy residential quality, we consider various specific measures of neighborhood and housing quality, including both subjective and objective ones, to assess residential life along multiple dimensions.

2. Background

Two general theoretical perspectives have been employed to evaluate racial/group disparities in residential processes and outcomes: assimilation and stratification. The assimilation perspective dates back to the early writings of Robert Park and other Chicago School scholars who argued that minority groups initially congregate in dense inner-city ethnic enclaves out of economic and cultural necessity, but as group differences between ethnic and majority groups shrink, minority group members come to rely less on the security of the enclave and leave it behind for more desirable residential spaces. Massey (1985) expanded these arguments into the spatial assimilation model, arguing that the distribution of ethnic populations across neighborhoods is a function of group’s relative acculturation and socioeconomic standing. From an empirical standpoint, the assimilation model implies that groups with greater economic resources, which have been in the U.S. for longer durations, and which have a stronger command of the English language will be observed in more-desirable residential settings. The assimilation model has been supported by considerable research: immigrants with higher earnings and longer stays in the U.S. are more likely to live in suburban and other more-advantaged neighborhoods (Adelman et al., 2001; Alba et al., 1999, 2000; Rosenbaum and Friedman, 2001; South et al., 2005a, 2005b) and to be homeowners and reside in homes with fewer structural deficiencies (Clark, 2003; Myers and Lee, 1998; Rosenbaum and Friedman, 2006). The spatial assimilation model has also been informative at explaining racial/ethnic differences in residential outcomes, with prior studies finding that assimilation-related factors account for a substantial portion of racial/ethnic inequalities in homeownership (Alba and Logan, 1992; Krivo, 1986, 1995), household crowding (Krivo, 1995), and exposure to neighborhood poverty (Adelman et al., 2001), violent crime (Logan and Stults, 1999; Peterson and Krivo, 2010), and hazardous pollutants (Crowder and Downey, 2010; Oakes et al., 1996).

While the assimilation perspective has received considerable empirical support, it has failed to fully attenuate racial/ethnic disparities in residential outcomes. The stratification perspective was formulated to explain the persistence of race/ethnicity in residential processes and highlights the structural barriers that limit housing and neighborhood access among minorities (see Alba and Logan, 1991). According to the stratification model, discriminatory behaviors and policies of banks, lenders, real estate agents, and landlords create racially-stratified housing markets that restrict housing and neighborhoods opportunities for some minority groups (Massey and Denton, 1993; Ross and Yinger, 2002; Yinger, 1997). The stratification perspective also highlights the racially-disparate residential preferences that act to constrain not only the supply of minorities’ housing possibilities, but also undercut neighborhood and housing quality due to the linkages between segregation, poverty, and neighborhood disadvantage (see Quillian, 2012; Sampson, 2012).

While these two models of residential attainment have been extensively applied to evaluating differences in housing and neighborhood conditions between whites and blacks, there is considerably less scholarship evaluating residential outcomes for Latinos, particularly with respect to factors other than homeownership and neighborhood income level. Moreover, little research has accounted for the sizeable portion of Latino immigrants who lack authorization to live in the U.S. Recent estimates pin the undocumented population at about 11.5 million, at least 70% of whom hail from Mexico or Central America (Hoefer et al., 2011; Passel and Cohn, 2011). Documentation status is widely known to have powerful effects on several dimensions of immigrant well-being and given that a near majority of Mexican and Central American immigrants lack documentation, it also is likely to have important implications for their incorporation. Previous work has found that undocumented workers have lower wages, slower wage growth, weaker returns to human capital, fewer ancillary benefits, and work in jobs with higher fatality rates than documented workers (Donato et al., 2008; Flippen, 2012; Hall et al., 2010; Kaushal, 2006; Kossoudji and Cobb-Clark, 2002; Massey and Bartley, 2005; Rivera-Batiz, 1999). Documentation status also influences educational achievement and attainment (Abrego, 2006; Bean et al., 2013; Gonzales, 2011; Greenman and Hall, 2013) and is related to health care access and health outcomes (Arbona et al., 2010; Cavazos-Rehg et al., 2007). This past research, along with the literature on residential attainment, provides several arguments for why immigrants’ residential outcomes – i.e., the quality of their housing units and the neighborhoods where they live – may vary according to legal status.

One possibility, in line with the assimilation perspective, is that undocumented migrants possess characteristics that limit their preferences for or accessibility to quality housing. Undocumented migrants, for example, tend to be younger and are less likely to be partnered than their legal counterparts (Hoefer et al., 2011; Passel and Cohn, 2011). They also have less schooling and, as implied above, lower incomes and fewer financial holdings than documented immigrants. These compositional differences between documented and undocumented migrants are likely to reduce homeownership for undocumented migrants, and to diminish both the quality of and satisfaction with their housing units and neighborhoods by limiting their ability to attain desirable housing and neighborhood locations.

Alternatively, the impact of legal status may withstand measured controls for demographic and socioeconomic characteristics. For example, lacking papers is likely to severely restrict the ability of migrants to obtain mortgages and guarantee financial security. Even among migrants with falsified social security cards or with access to institutions permitting the use of home-country identification, undocumented migrants are likely to be reluctant to interact with banks, lenders, insurers and other institutional actors in the housing-search process out of fear of detection. Indeed, Suro et al. (2002) report that unauthorized status is the primary reason why many Latino immigrants lack bank accounts and Amuedo-Dorantes and Bansak (2006) find that undocumented Mexicans are 7.3 times less likely to hold bank accounts than documented Mexican migrants (also Bair, 2003). Limited access to credit is not only likely to limit undocumented migrants’ ability to purchase housing but is also likely to make them more susceptible to exploitation on behalf of landlords or at least to severely restrict their opportunities for housing. Aside from limited access to housing, undocumented migrants may concentrate in less-desirable neighborhoods in order to avoid interaction with authorities. Segregation research provides suggestive evidence of this sort of process, finding that segregation between Mexican and whites is heightened in areas with large undocumented populations (Hall and Stringfield, 2013). As a result of a limited set of housing options and possible anxieties over detection, undocumented migrants may be willing to accept living in homes of substandard quality or in problematic neighborhoods. Moreover, challenges in securing housing combined with limited economic reserves may prompt undocumented migrants to combine resources and coreside, increasing the likelihood of overcrowding.

Owing largely to data limitations in identifying undocumented migrants, there are relatively few studies that have explored the connection between legal status and residential attainment. Moreover, the few studies that have considered documentation status are somewhat limited in geographic scope or by measurement issues. Paral (2004) and Toussaint-Comeau and Rhine (2004) suggest that undocumented Latino immigrants are less likely to become homeowners than documented ones, but because these studies use citizenship as a proxy for documented status, they are likely to underestimate the true impact of legal status. Similarly, McConnell and Marcelli (2007) find that among new lawful immigrants, those with prior unauthorized experience had odds of ownership that were about 20% lower than those who were not previously undocumented, but their data lacks information on residential statuses while respondents were actually undocumented and those successful in transitioning to an authorized status are likely to be a select group and not necessarily representative of all undocumented migrants. To alleviate these measurement issues, recent studies use data from the Los Angeles Family and Neighborhood Survey which includes specific questions on current authorization status to evaluate the impact of legal status on housing-induced poverty (McConnell, 2012), expenditures on housing and housing burdens (McConnell, 2012, 2013) and neighborhood advantage (Cort, 2011). These studies find moderate to large effects of legal status on residential outcomes, with Cort (2011) concluding that undocumented Latinos have replaced blacks as the group living in the most disadvantaged neighborhoods (also see McConnell and Marcelli (2007)). Whether these results are limited to Los Angeles County (or during the 2000–2002 period) or are generalizable to the rest of the nation is unknown.¹

Existing residential research has also focused almost exclusively on two types of residential outcomes: homeownership and neighborhood income/poverty level. Both are important domains of residential life, but serve mostly as proxies for specific aspects of neighborhood and housing-quality. Comparatively lacking is information on the detailed features of quality that undocumented migrants may be lacking, including neighborhood amenities, cleanliness, and safety, and the structural soundness of properties. We do so in our analysis, but rely on survey respondent self-reports to assess neighborhood and housing quality and overall satisfaction. While we expand on past work by moving beyond conventional measures of residential attainment, these more-specific measures are somewhat prone to subjectivity. It is widely acknowledged that in assessing their social position, immigrants make comparisons not only to those in their current countries, but also to those in their prior ones (Franzini and Fernandez-Esquer, 2006; Gelatt, 2013). Consequently, it is important to keep in mind in interpreting the more-subjective outcomes (in contrast to our more objective ones [ownership and household crowding]) that group differences can reflect both variation in real living conditions and variation in how comparisons are framed.

3. Data and methods

We rely on the 1996, 2001, 2004, and 2008 panels of the Survey of Income and Program Participation (SIPP) – a panel study focusing on U.S. households’ employment and public program experiences – for this project. The SIPP design draws a large, nationally-representative sample of U.S. households and collects information on each household member every four months for approximately four years. At each interview, respondents are asked a set of core questions and topical questions specific to each wave that cover the reference month and three preceding months. In cases where respondents are non-English speakers, SIPP provides translators.² While we use information on all household members in some measures, our interest in residential attainment necessitates a focus on household heads to avoid duplication of residential statuses within the same household. In addition, because the elderly have distinct processes of residential attainment (Mateyka, 2012), we restrict our sample to householders between 18 and 64. We also restrict our analysis to foreign-born Mexican and Central American (MCA) immigrants, U.S.-born Latinos, non-Latino whites, and non-Latino blacks.³ Finally, we restrict our sample to respondents participating in the second wave of each SIPP panel – which includes questions on place of birth, citizenship, and visa status – and the wave during which the Adult Well-Being Module was administered, which incorporates items on housing and neighborhood quality.⁴ Our final analytic sample includes 3792 MCA immigrants, 3663 native Latinos, 62,621 native whites, and 7896 native blacks.

Our analysis explores residential attainment and satisfaction along several dimensions. Homeownership – the conventional outcome of interest in residential scholarship – is a binary indicator of whether the householder’s housing unit is owner-occupied. Household crowding is the number of occupants of the household divided by the number of rooms in the home.⁵ To evaluate more specific components of householders’ living conditions, we consider four measures of neighborhood quality and two measures of housing quality. First, as a general measure of neighborhood dissatisfaction, we distinguish householders who are “very” or “somewhat’’ dissatisfied with overall conditions in their neighborhood from those who are very or somewhat satisfied.⁶ Next, we use factor analytic techniques to condense specific questions on neighborhood conditions. Specifically, we reduce 13 items to 3 factors of neighborhood problems using principal factor analysis with an orthogonal varimax rotation, which we refer to as: safety problems, public service problems, and environmental problems.⁷ (Each of the items used and their factor loadings are shown in Appendix Table A1) For our measures of housing quality, we create a binary indicator of dissatisfaction with householders’ homes that distinguishes between those who are “very’’ or “somewhat’’ dissatisfied from those who are very or somewhat satisfied.⁸ We also use factor analysis to reduce 7 items on housing unit problems – e.g., leaky roof, broken windows, and holes in walls – into a single factor of physical condition problems (specific items and their loadings are shown in Appendix Table A1).

The key explanatory variable in our analysis is documentation status. To identify legal status, we apply the approach developed by Hall et al. (2010), which uses information on respondents’ citizenship, legal permanent resident status, and participation in federal public assistance programs to impute documentation status for Mexican and Central American immigrants. Specifically, respondents who indicate that they are citizens or legal permanent residents (either currently or at entry) are classified as legal.⁹ We also track respondents’ participation in all federal assistance programs that undocumented immigrants are not eligible for (e.g., Food Stamps, Medicaid, SSI, TANF) throughout the entire SIPP observation period; if an immigrant reports receiving benefits from one of these federal programs in their own name (as opposed to dependently through someone else in the household [e.g., a U.S.-born child]) at any point while in SIPP, they are also classified as legal. The balance is either undocumented or falls into one of the following categories: refugees/asylees, students and exchange visitors, tourist/business travelers, temporary workers, and diplomats and other political representatives (U.S. Department of Homeland, 2012). SIPP does not sample tourists and other short-term visitors. Those admitted as diplomats are accounted for by deeming MCA foreigners who are themselves or are married to a high-ranking public official to be in the country legally. To account for those on student visas, we exclude householders who are themselves, or have a spouse who is, enrolled in college full-time as legal. The residual group that we are unable to account for is temporary workers. Authorized temporary workers, however, form a comparatively small portion of MCA immigrants (U.S. Department of Homeland, 2012). Nevertheless our results should be assessed with some caution as the group we refer to as undocumented workers potentially includes a small proportion of legal temporary workers.¹⁰

We incorporate several human capital characteristics implied by the assimilation model to encourage access to high-quality neighborhoods and housing, including educational attainment (in years of schooling), employment status (currently working vs. not), and monthly household income per person in the household (in thousands). Sociodemographic controls include age, gender, marital status, and the presence of children. For immigrants, we also include arrival recency, defined as having arrived in the U.S. in the last 5 years.¹¹ Dummies for each of the four census-defined regions are included with the Western region – where a majority of MCA immigrants live – serving as the referent.¹² Our models also incorporate year (1998, 2003, 2005, 2010) fixed effects and, to account for the uneven geographic distribution of our five groups, fixed effects for state of residence. All covariates that vary during the SIPP observation period refer to the same month in which residential outcomes were measured. Summary statistics for and correlations among all variables used in the analysis are shown in Appendix Table A2.

To examine residential attainment, we estimate OLS and logit (for the three binary outcomes [ownership, neighborhood dissatisfaction, housing dissatisfaction]) models that regress each of the eight measures of residential attainment/quality on group indicators and other explanatory variables. For each outcome we show results for the full sample of immigrants and natives to evaluate the hierarchy of group differences in residential attainment, and subsequently show results for the MCA immigrant sample that attempt to isolate the impact of legal authorization on residential attainment and quality. Descriptive results are weighted using the wave-2 person weights provided by SIPP.

4. Results

4.1. Group differences in residential attainment and quality

We begin our analysis by comparing residential outcomes for our five racial/legal groups. Table 1 shows mean differences in each of our measures of residential attainment, neighborhood quality, and housing quality. The two most common patterns gleaned from the group differences are consistent with decades of prior research on racial residential stratification: whites enjoy the most favorable residential settings and, with a few important exceptions, blacks experience the worst. More relevant to this study, undocumented MCA immigrant householders have residential statuses that are – on several dimensions – nearly as bad as or worse than native black householders. For example, just one in four undocumented MCA householders are homeowners, while twice as many documented MCA immigrants, native Latinos, and native blacks are, and nearly three times as many native whites are. Undocumented householders also live in homes with significantly more people per room than any other group, particularly native whites and native blacks. They are also more likely than their documented counterparts, as well as native Latinos and whites, to report problems with neighborhood services. Additionally, undocumented householders report relatively high levels of dissatisfaction with neighborhood and housing quality, problems with the neighborhood environment (e.g., litter, noise) and problems with the physical condition of their housing units. Overall, the unadjusted differences in Table 1 are consistent with the stratification of undocumented migrants’ residential settings in which they rank lowest or near-lowest on nearly every dimension of residential attainment and quality. In the models that follow, we explore the extent to which these differences persist after accounting for group variation in sociodemographic characteristics, acculturation, and area of residence.

Table 1.

Residential attainment and quality by legal/racial group for SIPP household heads, 1998–2010.

	Undocumented MCA Immigrants		Documented MCA Immigrants	U.S.-born Latinos	U.S.-born non- Latino whites	U.S.-born non- Latino blacks
Homeowner	.244	d,l,w,b	.513	.517	.719	.485
People per room	.943	d,l,w,b	.902	.648	.447	.485
Neighborhood Quality
Not satisfied with overall quality	.074	w	.067	.067	.038	.086
Safety problems	−.019	l,b	.002	.093	−.050	.265
Services problems	.283	d,l,w	.201	.199	−.074	.293
Environmental problems	.163	w,b	.133	.151	−.057	.247
Housing unit quality
Not satisfied with overall quality	.097	w	.080	.082	.056	.102
Physical condition problems	.098	w	.075	.059	−.024	.066
N of persons (unweighted)	775		3,017	3,663	62,621	7,896

Notes: d, l, w, and b indicate that means for undocumented MCA immigrants differ from documented MCA immigrants, native Latinos, native whites, and native blacks significantly at p < .05 (two-tailed t); weighted by wave-specific person weights.

4.2. Multivariate models of residential attainment and quality

Traditional models of residential outcomes have centered on the extent to which groups differ in their ability to purchase housing and on basic dimensions of the housing environment. In line with this prior work, we first examine differences in homeownership and housing crowing (people per room). As shown in the first column of Table 2, even with controls for demographic, socioeconomic, and area of residence characteristics, undocumented householders are severely limited in their access to ownership, with all other groups having significantly higher log odds of homeownership. In particular, the odds of ownership are between 58.9% (e^(.463) − 1) and 155.0% (e^(.936) − 1) higher for the other minority groups and 272.5% (e^(1.315) − 1) higher for native whites than for undocumented householders. When we restrict the sample to MCA immigrants (second column of Table 2) and incorporate a control for arrival recency, the ownership deficit for undocumented householders remains and is consistent with arguments that the inability to or fear of interacting with institutional actors in the housing market (realtors, lenders, insurers) restricts housing opportunities for undocumented migrants. Other covariates in the model tend to operate in a manner consistent with the attainment model: ownership increases with income, education, and employment, is more likely among married householders and those with children, and increases with age. In line with the assimilation perspective, ownership is also less likely among immigrants who arrived within the last 5 years.

Table 2.

Models of residential attainment for SIPP heads, 1998–2010.

	Homeownera		Household crowding
	(1)	(2)	(3)	(4)
Documented immigrant	.853***	.864***	−.053***	−.044*
	(.097)	(.111)	(.010)	(.022)
Native Latino	.936***	–	−.202***	–
	(.097)	–	(.010)	–
Native white	1.315***	–	−.327***	–
	(.091)	–	(.009)	–
Native black	.463***	–	−.286***	–
	(.094)	–	(.009)	–
Recent arrival	–	−.795***	–	.090***
	–	(.124)	–	(.024)
Age (in years)	.127***	.174***	−.006***	.003
	(.006)	(.031)	(.001)	(.006)
Age squared	−.001***	−.002***	.000***	−.000
	(.000)	(.000)	(.000)	(.000)
Female-headed	.035	.084	−.038***	−.036
	(.025)	(.173)	(.003)	(.033)
Children present	.418***	.227*	.192***	.195***
	(.026)	(.114)	(.002)	(.023)
Married	1.356***	.969***	.075***	.029
	(.024)	(.119)	(.002)	(.024)
Education (in years)	.058***	.023*	−.013***	−.013***
	(.004)	(.010)	(.000)	(.002)
Working	.356***	−.151	−.024***	.003
	(.026)	(.123)	(.003)	(.025)
Household income	.175***	.349***	−.017***	−.107***
	(.006)	(.053)	(.000)	(.008)
Region (West = ref)
Northeast	.298	−.673	.000	.009
	(.308)	(1.579)	(.032)	(.330)
Midwest	−.102	.175	−.082*	.153
	(.308)	(1.336)	(.032)	(.280)
South	.962***	.972	−.044	−.041
	(.269)	(1.466)	(.028)	(.305)
Constant	−.6.828***	−.6.505***	1.142***	1.093
	(.300)	(1.427)	(.030)	(.119)
R-squared	.223	.168	.366	.202

Notes: Model 1 N = 77,972; Model 2 N = 3792.

^*p < .05.

^**p < .01.

^***p < .001.

^aLogit coefficients shown; standard errors in parentheses; includes year and state fixed effects.

The third and fourth columns of Table 2 show corresponding results for household crowding – the number of people per room in the housing unit. Here too we see evidence of stratification by legal status, with undocumented householders having significantly higher levels of housing crowding than any other group. Multiplying the group coefficients by mean number of rooms (5.9) indicates that homes headed by an undocumented migrant contain about 1.93 more people than similarly-sized homes headed by native whites, 1.69 more people than native black householders’ homes, and 1.19 more people than homes headed by native Latinos. When the sample is limited to MCA immigrants, undocumented householders still exhibit higher levels of housing crowding. Normalizing by the average number of rooms for MCA immigrants (4.8) indicates that a home headed by an undocumented migrant contains, on average, about .21 more people than a similarly-sized home headed by a documented migrant. As with ownership, other covariates in the household crowding models work in expected directions with crowding decreasing with age, education, and income, but increasing with events that augment the size of the household (marriage, children).

Table 3 reports parallel coefficients for our various measures of neighborhood quality: dissatisfaction with the overall condition of the neighborhood, and factor-based scales of problems associated with neighborhood safety, public services, and the local environment. Looking first at the binary indicator of neighborhood dissatisfaction, the only significant group coefficient is for native blacks whose odds of being unhappy with the overall condition of their neighborhood are about 72.6% (e^(.546)-1) higher than undocumented householders.¹³ There does not appear to be any meaningful difference by legal status or nativity in householders’ general satisfaction levels. Other variables in the model are consistent with the attainment model: the odds of being dissatisfied are lower among married householders, those with children, and those with higher incomes and education. However, in the model limited to MCA immigrants, acculturation and SES measures are not significantly associated with neighborhood satisfaction, in contrast with the assimilation perspective.

Table 3.

Models of neighborhood quality for SIPP heads, 1998–2010.

	Not satisfied with qualitya		Safety problems		Services problems		Environmental problems
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Documented immigrant	−.070	−.216	.027	.009	−.117***	−.148**	−.038	−.098*
	(.158)	(.181)	(.036)	(.042)	(.035)	(.050)	(.031)	(.042)
Native Latino	.015	–	.200***	–	−.138***	–	−.048	–
	(.016)	–	(.035)	–	(.034)	–	(.031)	–
Native white	−.203	–	.131***	–	−.245***	–	−.094***	–
	(.148)	–	(.033)	–	(.032)	–	(.029)	–
Native black	.546***	–	.390***	–	.078*	–	.181***	–
	(.152)	–	(.034)	–	(.033)	–	(.030)	–
Recent arrival	–	−.043	–	−.059	–	−.018	–	−.077
	–	(.208)	–	(.046)	–	(.055)	–	(.046)
Age (in years)	.039***	−.054	.019***	.007	.002	−.020	.010***	−.018
	(.011)	(.051)	(.002)	(.012)	(.002)	(.014)	(.002)	(.012)
Age squared	−.001***	.001	.000***	.000	.000	.000	.000***	.000
	(.000)	(.001)	(.000)	(.000)	(.000)	(.000)	(.000)	(.000)
Female-headed	.142**	−.089	−.027**	−.078	.253***	.125	.092***	.034
	(.047)	(.299)	(.010)	(.065)	(.010)	(.078)	(.009)	(.065)
Children present	−.164**	−.080	−.041***	−.003	−.041***	−.059	−.060***	−.014
	(.051)	(.203)	(.009)	(.045)	(.009)	(.054)	(.008)	(.045)
Married	−.297***	.174	−.070***	−.062	.002	.153**	−.064***	.058
	(.048)	(.213)	(.009)	(.047)	(.008)	(.056)	(.008)	(.046)
Education (in years)	−.046***	.023	−.025***	.001	−.017***	−.007	−.020***	−.004
	(.007)	(.018)	(.001)	(.004)	(.001)	(.005)	(.001)	(.004)
Working	−.338***	−.065	−.093***	−.106*	−.139***	−.087	−.116***	−.090
	(.045)	(.215)	(.009)	(.049)	(.009)	(.059)	(.008)	(.049)
Household income	−.132***	−.125	−.024***	−.031	−.021***	−.059**	−.023***	−.052**
	(.013)	(.093)	(.002)	(.016)	(.001)	(.019)	(.001)	(.016)
Region (West=ref)
Northeast	−.252	2.084	−.033	.296	−.025	.596	.009	1.077
	(.630)	(1.315)	(.113)	(.643)	(.109)	(.773)	(.098)	(.641)
Midwest	.208	−.082	−.214	.197	.075	.501	.064	.632
	(.602)	(1.242)	(.115)	(.528)	(.111)	(.635)	(.099)	(.527)
South	−.992	.887	−.252*	.197	.095	.681	−.109	.315
	(.543)	(.934)	(.099)	(.574)	(.095)	(.690)	(.086)	(.573)
Constant	−.2.086***	−.2.254	.315**	−.132	.477***	.198	.429***	.113
	(.576)	(1.219)	(.109)	(.556)	(.106)	(.669)	(.095)	(.555)
R-squared	.044	.035	.049	.042	.067	.040	.067	.054

Notes: Model 1 N = 77,972; Model 2 N = 3792.

^*p < .05.

^**p < .01.

^***p < .001.

^aLogit coefficients shown; standard errors in parentheses; includes year and state fixed effects.

Results for the neighborhood safety problems scale are shown in the third and fourth columns of Table 3. The group coefficients indicate that native householders of any race/ethnicity report significantly greater problems with neighborhood crime and safety than do MCA immigrant householders. Native blacks express the greatest concern with safety, being about two-fifths of a standard deviation higher on the safety-problems scale than undocumented MCA householders. Like with neighborhood satisfaction, there are no meaningful differences between documented and undocumented MCA immigrants in their concerns about neighborhood safety. That native Latinos voice greater anxieties about neighborhood safety than MCA immigrants may suggest that across generations, Latinos are concentrating in more dangerous or higher-crime neighborhoods. Conversely, it could simply result from immigrants being comfortable with crime/safety in American neighborhoods relative to former neighborhoods in Mexico, regardless of actual levels of criminal activity in the neighborhoods occupied by immigrant and native householders. Aside from these group differences, reported concerns with neighborhood safety, in the pooled-group model, are lower among married and female householders, those with children, and those with higher education and incomes.

Reported problems with neighborhood public services – e.g., quality of local hospitals, policing – shown in the fifth and sixth columns point to considerable group differences. In particular, documented MCA immigrants, native Latinos, and native whites report significantly lower levels of concern with public services than do undocumented householders. Native blacks, however, score significantly higher on the services problems scale than any group, reinforcing their exceptional position in racial residential stratification. Nevertheless, restricting the sample to MCA immigrants reinforces the disparity in reports of public service problems between documented and undocumented migrants, with the former scoring about one-seventh of a standard deviation lower on the scale than the latter.

Similar results are revealed for local environmental issues – e.g., problems with trash, noise, or odors – in which undocumented householders express greater concern with the physical condition of their neighborhoods than native white and Latino householders, although only the difference with native whites – in the pooled-group model – reaches statistical significance. The heightened perceptions of neighborhood problems among native blacks is also revealed on this measure, who are more likely than any other group to report concerns with the local environment. The MCA immigrant only model, shown in the final column of Table 3, reports a significant and negative coefficient for legal status, suggesting that undocumented migrants are more likely to describe their neighborhoods as suffering from poor environmental conditions.¹⁴

Lastly, the results for housing quality are shown in Table 4. The first set of models (columns 1 and 2) predict the log odds of being dissatisfied with the overall quality of the housing unit in which householders live. There are minimal group differences in overall housing satisfaction, although native blacks odds of reporting residence in undesirable housing units are about 43.5% (e^(.361) − 1) higher than undocumented migrants. Variation in the physical condition of housing units is considered in the final two columns on Table 4. In the combined-group model, native whites are the only group that differ significantly from undocumented migrants in terms of the physical state of respondents’ home, with whites – as expected – having modestly fewer problems with their housing units. When the sample is restricted to MCA immigrants, however, a larger and statistically-significant effect of legal status is revealed. Specifically, the negative coefficient for documented immigrants implies that legal MCA migrants have moderately fewer problems with their homes than undocumented ones.¹⁵ Supplemental analyses of the specific items comprising the physical problems scale indicate that, in comparison to their authorized counterparts, undocumented migrants are especially likely to report problems associated with pests and insects, exposed wires, and holes in walls. Other covariates on both measures of housing quality – in the combined group model – indicate that dissatisfaction and housing problems tends to increase with age, but are lower among married householders and those with greater socioeconomic resources.

Table 4.

Models of housing unit quality for SIPP heads, 1998–2010.

	Not satisfied with qualitya		Physical condition problems
	(1)	(2)	(3)	(4)
Documented immigrant	−.203	−.262	−.045	−.078*
	(.141)	(.161)	(.031)	(.038)
Native Latino	−.050	–	−.022	–
	(.140)	–	(.030)	–
Native white	−.020	–	−.048*	–
	(.130)	–	(.021)	–
Native black	.361**	–	.006	–
	(.134)	–	(.029)	–
Recent arrival	–	.231	–	.045
	–	(.175)	–	(.045)
Age (in years)	.052***	.036	.012***	.021
	(.009)	(.047)	(.002)	(.012)
Age squared	−.001***	−.000	.000***	.000
	(.000)	(.001)	(.000)	(.000)
Female-headed	.048	−.044	−.020*	−.030
	(.041)	(.265)	(.008)	(.064)
Children present	−.057	−.251	.014	−.026
	(.043)	(.184)	(.008)	(.045)
Married	−.411***	.135	−.093***	−.006
	(.040)	(.192)	(.008)	(.046)
Education (in years)	−.062***	−.017	−.013***	−.008*
	(.006)	(.016)	(.001)	(.004)
Working	−.248***	−.030	−.118***	−.065
	(.039)	(.198)	(.008)	(.049)
Household income	−.207***	−.185*	−.017***	−.042***
	(.012)	(.093)	(.001)	(.016)
Region (West = ref)
Northeast	.029	−.2.394	.113	.096
	(.419)	(1.604)	(.098)	(.633)
Midwest	−.891	−.2.403	.024	.160
	(.495)	(1.436)	(.099)	(.536)
South	−.1.061	−.2.557	−.190*	−.070
	(.378)**	(1.430)	(.085)	(.584)
Constant	−.1.497***	−.407	.325***	−.215
	(.422)	(1.534)	(.094)	(.563)
R-squared	.041	.035	.023	.038

Notes: Model 1 N = 77,972; Model 2 N = 3792.

^*p < .05.

^**p < .01.

^***p < .001.

^aLogit coefficients shown; standard errors in parentheses; includes year and state fixed effects.

To summarize our findings, we report predicted values for each of our eight residential outcomes, by group, in Table 5. These predictions are based on the pooled-group models in Tables 2–4 and include all of the measured controls shown there. To highlight the racial ordering to our results, we bold the values for the group with the least-favorable residential ranking. As noted above, undocumented migrants have substantially lower regression-adjusted rates of homeownership (.451) and considerably higher levels of household crowding (.796 people per room) than all other groups. While they are positioned near the bottom of the group hierarchy on the more specific measures, native black householders ranked lowest on all dimensions of neighborhood and housing quality. Specifically, African Americans report considerably greater dissatisfaction with the quality of both their neighborhoods and housing, express much greater concern with neighborhood safety, public amenities, and environmental conditions, and state more problems with the structural soundness of their housing units. While these more detailed measures of neighborhood and housing quality are arguably more sensitive to subjective interpretation, the magnitude of the deviations of blacks from the other groups (including undocumented migrants on all but housing physical condition) underscore the exceptional nature of blacks’ residential experience (Rosenbaum and Friedman, 2006).

Table 5.

Predicted residential attainment and quality by legal/racial group for SIPP heads, 1998–2010.

	Undocumented MCA immigrants	Documented MCA immigrants	U.S.-born Latinos	U.S.-born non-Latino whites	U.S.-born non-Latino blacks
Homeowner	.451	.658	.676	.753	.566
People per room	.796	.743	.595	.469	.511
Neighborhood quality
Not satisfied with overall quality	.044	.041	.044	.036	.073
Safety problems	−.155	−.128	.045	−.024	.235
Services problems	.195	.079	.157	−.050	.273
Environmental problems	.056	.018	.104	−.038	.237
Housing unit quality
Not satisfied with overall quality	.055	.045	.052	.054	.077
Physical condition problems	.041	−.004	.019	−.007	.047

Note: Bolded values refer to group with the poorest residential outcome.

5. Conclusion

Previous research has shown pervasive effects of undocumented status on many aspects of well-being for Mexican and Central American immigrants, including low earnings, weak returns to human capital, poor youth educational attainment, and employment in risky occupations (Donato and Massey, 1993; Flippen, 2012; Greenman and Hall, 2013; Hall et al., 2010; Kaushal, 2006; Kossoudji and Cobb-Clark, 2002; Massey, 1987; Rivera-Batiz, 1999; Suárez-Orozco et al., 2011). Housing and residential context not only are additional dimensions of well-being, they also have profound consequences for social and economic integration (see Sampson, 2012). Like with most questions related to legal status, however, data constraints have prevented researchers from systematically investigating the relationship between legal status and residential attainment. We overcome these limitations by employing an imputation strategy to identify the documentation status of Mexican and Central American (MCA) immigrant respondents in the 1996–2008 panels of the Survey of Income and Program Participation. We then assess the relationship between legal status of MCA immigrants and several key dimensions of residential attainment, including homeownership, crowding, satisfaction with the neighborhoods and housing, and factors assessing the quality of respondents’ neighborhoods or housing units. We compare documented and undocumented MCA immigrants with native whites, blacks, and Latinos in order to gauge where MCA immigrants fit into the racial/ethnic hierarchy that has long defined American residential life.

Our findings clearly show that legal status is associated with poorer residential outcomes for MCA immigrants. Models comparing documented and undocumented immigrants indicate that, net of personal and immigration-related characteristics, lacking legal authorization to live in the U.S. is associated with inferior residential contexts. Specifically, our analysis shows that undocumented immigrants are substantially less likely that their documented counterparts to be homeowners. They also tend to live in more crowded housing units and report greater problems with the physical condition of their homes, such as holes in the walls, pests, and exposed wires. We also find evidence that lacking legal status puts immigrants at risk of living in lower-quality neighborhoods, particularly those with inferior public services and environmental problems (e.g., trash, noise, odors). Certainly part of the explanation for why undocumented migrants live in less-desirable residential environments is that lacking authorization restricts accessibility to quality housing and neighborhoods or induces households to settle in less-advantaged neighborhoods where perceived fears of detection are minimized. Alternatively, undocumented migrants may be more likely to consider their migration to be temporary in nature (Chavez et al., 1997) and thus more willing to accept less-favorable residential conditions over the short run. However, post-IRCA border controls have substantially reduced cross-border movement and likely altered migration intentions of undocumented migrants (Massey et al., 2002), which is suggested in survey data finding undocumented and documented migrants to have similar plans to stay in the U.S. (Massey and Akresh, 2006; Wampler et al., 2009).

Our research also informs discussion of how undocumented MCA immigrants fit into the structure of racial/ethnic stratification in residential outcomes. Homeownership, not surprisingly, is one of two dimensions where undocumented MCA immigrants fare worse than all other groups in our analysis. This is likely because lacking legal status impedes access to credit markets, as well as making migrants reluctant to interact with banks and other institutions involved in the process of buying a home. Household crowding is the second feature on which undocumented immigrants clearly fare worse than all other groups, although their distinctiveness is less striking than is true for homeownership. On most other dimensions of housing and neighborhood quality examined, undocumented immigrants fall somewhere in the middle of the three native comparison groups. They fare worse than native whites but better than native blacks in terms of neighborhood services and environmental problems. The physical conditions of their housing are comparable to those of other minority groups, but worse than those of native whites. Finally, self-reported levels of satisfaction with both neighborhood and housing quality were comparable to those of native whites, and significantly better than those of native blacks. Thus, despite the unique disadvantage posed by lacking legal status, the position of undocumented MCA immigrants in the racial/ethnic hierarchy of residential attainment is not clearly at the bottom but rather depends upon the specific aspect of residential attainment examined.

It is useful to remind readers that several of our measures of neighborhood and housing quality are based on subjective self-reports of neighborhood and housing concerns and, thus, group differences in these outcomes reflect not only real differences between groups, but also differences in the reference frames that group members employ to evaluate problematic conditions. If groups systematically differ in their observation or scrutiny of neighborhood or housing concerns, we are likely underestimating any effects of legal status. Indeed, the fact that MCA immigrants score lower on concerns with local crime and safety could result from immigrants’ evaluation of crime relative to their experiences in home countries suffering from social or political violence (Correia, 2010; Davis and Hendricks, 2007; Menjivar and Bejarano, 2004). These dual frames of reference may, thus, generate positive sentiments toward feelings about crime and safety in American neighborhoods even if their actual experiences are similar to or worse than those experienced by native households. To the extent that subjective evaluations differ between immigrants and natives, we may also be overstating the gap between native blacks and other groups. Despite this important caveat, native blacks’ substantial divergence from other groups along several dimensions is a testament to the enduring power of the black/nonblack divide in structuring residential life in America.

Another potential limitation to this study is that in addition to undercounting undocumented migrants, large surveys like SIPP potentially suffer from non-random coverage bias of households containing unauthorized persons. While it is possible that those participating in surveys such as SIPP are less threatened by involvement and may, thus, represent a slice of the total unauthorized population that has relatively less to lose from survey participation, a recent evaluation of the SIPP-based legal status allocation employed here suggests that that the demographic profile of undocumented persons in SIPP compares favorably to samples of undocumented populations derived from administrative and non-self-reported data (Bachmeier et al., forthcoming). Specifically, the authors conclude that there is “little [evidence] to suggest that misreporting of legal status is so widespread in the SIPP to lead to substantially biased estimates of the unauthorized immigrant population’’. Nevertheless, to the extent that undocumented respondents in SIPP are positively selected, estimated impacts of legal status will be biased toward zero, and our results of the residential consequences of lacking authorization are likely to be somewhat conservative. In addition, if our sample of undocumented householders is selectively advantaged, then estimated gaps between undocumented migrants and native blacks may be artificially large.

Overall this study offers further evidence of the pervasive impact of undocumented status in structuring social inequality. In addition to being segregated in jobs with low pay and few opportunities for advancement, lacking complete access to educational opportunities and public assistance programs, and suffering from mental and physical health impairments, our research demonstrates that immigrants’ residential attainment is strained by lacking legal authorization. From a policy standpoint, inferior residential contexts for undocumented migrants should be of concern given the widespread deleterious effects of living in disadvantaged areas and their negative impacts on social and economic mobility (see Sampson et al., 2002). This is particularly true for children, of whom approximately 4.5 million of whom are the U.S.-born offspring of unauthorized parents (Passel and Cohn, 2011). While ongoing policy discussions give hope that the undocumented population in the U.S. will be brought out of the shadows, theirmiring in poor residential environments is likely to have lasting social costs – especially among the youngest members – given that early residential experiences are associated with weakened outcomes in adulthood (Crowder and South, 2011; Jackson and Mare, 2007; Sharkey, 2008; South and Crowder, 2010; Wodtke et al., 2011; Wodtke, forthcoming).

Appendix A

Table A1.

Loadings for neighborhood and housing factors used in analysis.

	Neighborhood safety problems	Neighborhood services problems	Neighborhood environment problems	Housing unit physical condition problems
Afraid to walk alone at night	.59	−.02	.09	–
Stayed home because felt unsafe	.74	−.02	−.05	–
Taken someone because felt unsafe	.72	−.03	−.06	–
Carried something to protect self	.36	.03	.02	–
Perception of neighborhood crime	.44	.14	.22	–
Street noise or heavy traffic	.04	−.04	.47	–
Streets in need of repair	−.05	.03	.44	–
Trash, litter, garbage in streets or lots	.03	−.02	.52	–
Odors, smoke, or gas fumes in neighborhood	−.02	−.02	.36	–
Satisfaction with hospitals/health clinics	−.02	.64	.00	–
Satisfaction with police services	.01	.76	.03	–
Satisfaction with fire department services	−.02	.75	−.08	–
Overall satisfaction with public services	.01	.65	.04	–
Rats, mice, roaches or other pests	–	–	–	.38
Leaky roof or ceiling	–	–	–	.40
Broken windows	–	–	–	.47
Exposed electrical wires	–	–	–	.39
Toilet, water heater, or other plumbing broken	–	–	–	.39
Holes in walls or ceiling	–	–	–	.54
Holes in floor	–	–	–	.39
Eignevalue	2.31	2.24	1.82	1.27

Table A2.

Summary statistics for and correlations among variables used in analysis.

		Mean	SD	Min	Max	(1)	(2)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)	(14)	(15)	(16)	(17)	(18)	(19)	(20)	(21)	(22)	(23)	(24)	(25)
(1)	Homeowner	.67	.47	.00	1.00	1.00
(2)	People per room	.49	.31	.05	6.00	−.14	1.00
(4)	Safety problems	−.01	.88	−.75	4.95	−.08	.06	1.00
(5)	Services problems	.00	.86	−.60	4.03	−.15	.08	.34	1.00
(6)	Environmental problems	.00	.77	−.73	4.56	−.16	.08	.53	.72	1.00
(7)	Not satisfied with overall quality	.06	.24	.00	1.00	−.08	.06	.17	.17	.22	1.00
(8)	Physical condition problems	.00	.76	−.25	9.25	−.07	.06	.18	.19	.26	.40	1.00
(9)	Undocumented MCA	.01	.11	.00	1.00	−.10	.16	.00	.04	.02	.02	.02	1.00
(10)	Documented MCA	.04	.21	.00	1.00	−.07	.29	.00	.05	.04	.01	.02	−.02	1.00
(11)	Native Latinos	.06	.24	.00	1.00	−.08	.13	.03	.06	.05	.02	.02	−.03	−.05	1.00
(12)	Native whites	.78	.41	.00	1.00	.21	−.26	−.09	−.16	−.13	−.06	−.05	−.21	−.41	−.47	1.00
(13)	Native blacks	.10	.30	.00	1.00	−.13	.00	.10	.12	.11	.05	.03	−.04	−.07	−.08	−.64	1.00
(14)	Recent arrivala	.16	.37	.00	1.00	−.19	.04	−.01	.00	−.03	.04	.01	.18	−.18	–	–	–	1.00
(15)	Age	43.13	11.82	18.00	64.00	.32	−.22	−.04	−.02	−.04	−.01	−.02	−.08	−.05	−.08	.09	.01	−.22	1.00
(16)	Female-headed	.18	.38	.00	1.00	−.17	−.21	.02	.12	.08	.03	.02	−.03	−.06	−.02	.00	.06	.07	−.03	1.00
(17)	Children present	.33	.47	.00	1.00	.15	.45	−.01	−.04	−.05	−.02	.00	.04	.12	.04	−.06	−.05	−.08	−.20	−.33	1.00
(18)	Married	.55	.50	.00	1.00	.35	.27	−.06	−.08	−.10	−.06	−.05	−.01	.06	−.01	.06	−.11	−.14	.15	−.52	.51	1.00
(19)	Education (in years)	13.47	2.87	.00	20.00	.15	−.26	−.11	−.11	−.12	−.08	−.08	−.17	−.30	−.08	.28	−.05	.02	.02	.04	−.01	.04	1.00
(20)	Working	.83	.37	.00	1.00	.10	.02	−.07	−.09	−.09	−.07	−.07	.02	.01	−.01	.06	−.09	.01	−.19	−.11	.17	.13	.16	1.00
(21)	H’hold monthly income per capita (thou)	2.32	2.45	−17.94	70.80	.14	−.29	−.10	−.10	−.11	−.08	−.08	−.06	−.11	−.06	.15	−.06	−.02	.12	.03	−.21	−.07	.33	.20	1.00
(22)	Northeast	.17	.38	.00	1.00	−.01	−.02	−.02	−.03	.01	.00	.00	−.02	−.07	.00	.07	−.04	.05	.03	.01	−.01	−.01	.06	.05	.06	1.00
(23)	Midwest	.25	.43	.00	1.00	.06	−.06	−.01	−.05	−.03	−.01	−.02	−.04	−.07	−.09	.13	−.05	.02	.00	−.01	.00	.02	.01	−.02	−.01	−.26	1.00
(24)	South	.36	.48	.00	1.00	.02	−.02	.01	.02	−.03	.01	.01	.00	−.03	−.01	−.10	.16	.08	.00	.00	−.01	.01	−.04	−.04	−.06	−.35	−.43	1.00
(25)	West	.22	.41	.00	1.00	−.08	.12	.02	.06	.05	.01	.01	.06	.17	.11	−.09	−.10	−.11	−.03	.00	.01	−.02	−.02	.02	.02	−.24	−.30	−.40	1.00

Notes: Weighted by person-weights.

^aRestricted to MCA immigrants.