Declining Segregation through the Lens of Neighborhood Quality: Does Middle-Class and Affluent Status Bring Equality?

Abstract

Middle- and upper-class status along with suburban residence are together considered symbolic of the American dream. However, the question of whether they mean access to better quality residential environments has gone largely unexplored. This study relies on data from the 2009 panel of the American Housing Survey and focuses on a range of neighborhood conditions, including indicators of physical and social disorder as well as housing value and a neighborhood rating. Contrary to the tenets of the spatial assimilation model, we find that middle-class and affluent status do not consistently lead to superior conditions for all households. Neighborhood circumstances vary considerably based on householder race and ethnicity, with blacks and Hispanics experiencing the greatest disparities from whites. In addition, suburban residence does not attenuate such differences, and in some cases, well-to-do minorities do even worse than whites in neighborhood quality in suburbs.

1. Introduction

Residential segregation, particularly between blacks and whites, is on the decline. Logan (2011, p. 5) finds that the average black-white index of dissimilarity scores in metropolitan America decreased from 73 in 1980 to 59 in 2010. For Hispanics during the same period, average levels of segregation, which are lower than that between blacks and whites, dropped more minimally from 50 in 1980 to 48 in 2010 (Logan, 2011, p. 11). Suburbs, which have long been associated with prosperous white communities, have seen a far greater representation of minorities (Denton and Gibbons, 2013). In 2010, 51 percent of blacks lived in suburbs, up from 37 percent in 1990, and the percentage of Hispanics in suburbs rose to 59 percent from 47 percent (Frey, 2011, p. 10). While declines in segregation have been found to be slower in suburbs than in cities (Denton and Gibbons, 2013; Fischer, 2008), these are still promising signs that greater racial integration is taking place.

Whether concurrent trends of declining segregation and the growing presence of minorities in suburbs have translated into better residential conditions for individual minorities, particularly those of middle-class and affluent backgrounds, relative to whites, is not well known. Some scholars argue that economic differences between whites and minorities are directly linked to residential segregation (Clark, 2007; Patterson, 1997; Thernstrom and Thernstrom, 1997). Therefore, declines in segregation should result in more equality in the neighborhood quality between middle- and upper-class minorities and whites. However, existing research on race and class has found that while segregation for middle class minorities has declined, it remains high for blacks (Fischer, 2003; Iceland et al., 2005; Iceland and Wilkes 2006; Massey and Fischer, 1999). The persistence of prejudices and negative out-group preferences (Charles, 2000; Farley et. al, 1994; Krysan and Farley, 2002) as well as discrimination, particularly in the form of racial and ethnic steering, will make minorities more disadvantaged in their residential outcomes, relative to whites, no matter how much the class-status gap is closed between minorities and whites (Massey and Denton, 1993; Massey and Fischer, 1999; Turner et al., 2002).

The lack of research on this topic is surprising given that the disparities in wealth between whites and minorities persist (Conley, 1999; Oliver and Shapiro, 1995). Housing and neighborhood conditions have been found to be key factors influencing housing values, which in turn help explain the wealth differences between householders (Fesselmeyer et al., 2013; Flippen, 2004). In 2009, the median net worth of households with a non-Hispanic white householder was $113,149, almost 20 times the median net worth of households with a black householder ($5,677) or Hispanic householder ($6,325) (Kochhar et al., 2011, p. 5). Given that black and Hispanic wealth is much more dependent on their residential circumstances, knowing more about the quality of neighborhoods in which middle class or affluent blacks and Hispanics live, especially if we define such class backgrounds as including homeownership, is important as is focusing on those who live in suburbs where homeownership is even greater (Conley, 1999; Kochhar, 2004; Oliver and Shapiro, 1995).

While there is some research that explores the effects of class and socioeconomic status on residential segregation at an aggregate level (Fischer, 2003; Iceland and Wilkes, 2006; Iceland, et al., 2005; Massey and Fischer, 1999), to our knowledge, only three quantitative studies exist in the literature that directly examine the locational attainment of middle-class or affluent blacks (Adelman, 2004, 2005; Alba et al., 2000). They find that race continues to be important in influencing their residential attainment. Although middle-class blacks have greater shares of whites in their neighborhoods, relative to other blacks, the whites are less affluent than those residing in middle-class white neighborhoods. However, these studies are limited in a number of ways. They do not examine the locational attainment of middle- and upper-class minorities on a national level. The data utilized within these studies are based upon data from the 1990 decennial census or the 1992-1994 MultiCity Study of Urban Inequality, which are about 20 years old (Adelman 2004, 2005; Alba et al., 2000). Neighborhoods within these studies are defined at the census-tract level and not as the characteristics near the person's housing unit, which potentially underestimates the true extent to which middle-class and affluent minorities experience the negative effects of residential segregation. Finally, these studies do not examine the locational attainment of such well-to-do households in suburbs, which is surprising given that many of these households reside in suburbs as well as the fact that suburban residence is considered the pinnacle of achievement in locational attainment models (Alba and Logan, 1991; Fischer, 2008; Friedman and Rosenbaum, 2007; Logan and Alba, 1993; Massey and Denton, 1985, 1988).

In addition to the aforementioned limitations of previous research, almost no research has directly examined the residential outcomes of middle-class or affluent Hispanics and Asians.¹ One exception is a study by Logan (2011) that examines the average neighborhood quality of poor, middle income, and affluent whites, blacks, Hispanics, and Asians using 2005-2009 data from the American Community Survey.² Logan (2011) finds that well-to-do blacks and Hispanics live in poorer neighborhoods than their white counterparts. While this study updates the older research on this topic and offers a broader focus on non-black minorities, it is limited because it is a descriptive analysis that does not control for other factors that could impact the gap in neighborhood quality between well-to-do whites and minorities, namely educational attainment, housing tenure, and suburban location.

The question that remains is how important are race and ethnicity in predicting middle-class and affluent household locational attainment in the 21^st century, given the declines that have occurred in segregation during the past few decades. To address this issue, we conduct bivariate and multivariate analyses of data from the 2009 panel of the American Housing Survey (AHS). The distinct advantages of these data are that they are at the national level, and contain respondents' reports of the physical conditions of their neighborhoods in terms of the presence of abandoned buildings, buildings with bars on the windows, trash/litter/junk, open spaces within a half a block of their housing unit, how residents rank their neighborhoods, and residents' perceptions of crime. Such perception-based measures of residential circumstances have been shown to be significantly related to more objective indicators (Elo et al., 2009) and have been linked in explaining variation in individual health outcomes (Chang et al. 2009; Vinikoor-Imlera et al., 2011; Weden et al., 2008). In addition, the AHS includes household reports of their housing values, which is likely to directly reflect the quality of their neighborhoods. Several questions are addressed using these data: 1) Do racial and ethnic differences in neighborhood outcomes exist among middle-class and affluent households?³ 2) To the extent that differences exist, are they smaller in suburbs? and 3) If racial and ethnic differences exist, do they disappear when controlling for relevant demographic and socioeconomic factors?

2. Theoretical Background

Two theoretical models have been used in the literature to characterize variation in the residential location of households. The spatial assimilation model identifies residential attainment as one of the key outcomes of the status attainment process. Variation in residential outcomes of households is a function of differences in their acculturation, socioeconomic status, and life cycle factors. The model suggests that on the whole, as minorities achieve upward economic mobility, they will transfer these upgrades into superior residences (Massey and Denton, 1985). Minority households with fewer socioeconomic resources will tend to live in lower-quality neighborhoods than majority-group households. Therefore, when focusing on middle-class and affluent households, it is expected that there will be little variation in their neighborhood outcomes, particularly after controlling for demographic factors and the variation that could exist in socioeconomic status of these well-to-do households.

In addition to these variables, life-cycle factors will play an important role in explaining the variation in household residential outcomes. Such factors shape household residential needs and preferences and thereby encourage or discourage their residential mobility (Rossi, 1955; Speare et al., 1975). Marital status and the presence of children will shape household residential preferences as both factors increase the need for more space. In addition, families with children are likely to have stronger preferences for neighborhoods with good schools, low crime, and low poverty (Rosenbaum and Friedman, 2001).

Another important, implicit assumption of the spatial assimilation model is the notion that assimilation involves a move to the suburbs. Suburban residence is widely believed to be the apex of the spatial assimilation process because of the superior residential opportunities thought to be found in suburban communities (Alba et al., 1999). As such, minorities and immigrants alike should have greater access to majority-group members, more affluent neighborhoods, and in general better neighborhood conditions in suburbs.

The main tenets of the spatial assimilation model have found support in the literature. Income and education are positively correlated with tract-level median income, percent white in the neighborhood, and school quality, and negatively associated with neighborhood crime, poverty, and teen fertility rates (Alba and Logan 1991, 1993; Alba et al. 1999, 2000; Logan et al., 1996; Rosenbaum and Friedman, 2007). In addition, households with more income, householders with better education, and native-born households are significantly more likely to live in suburbs than those with less income and education and who are comprised of immigrants (Alba et al., 1999). For Asian households, the tenets of the spatial assimilation model work particularly well. In general, Asians have similar neighborhood outcomes and in some cases even better outcomes than do whites, particularly when controlling for socioeconomic, demographic, and acculturation-related variables (Logan et al., 1996; Rosenbaum and Friedman, 2007).

On the other hand, studies have shown that blacks and Hispanics reside in lower-quality neighborhoods than whites, even among those with greater income or that live in suburbs, controlling for relevant socioeconomic and demographic factors, thereby suggesting that other factors are affecting their variation in household locational attainment (Alba et al., 1999; Alba et al., 2000; Friedman and Rosenbaum, 2007; Logan and Alba, 1993; Logan et al., 1996; Rosenbaum and Friedman, 2007; Woldoff and Ovadia, 2009). The majority of the work done on middle-class household neighborhood quality so far has used qualitative methodological approaches, with the most prominent work having been done on African Americans (Lacy, 2007; Pattillo, 2007; Pattillo-McCoy, 1999) and Hispanics (Vallejo, 2010, 2012). Such studies that have focused on the black middle class have drawn similar conclusions, finding that well-to-do blacks are often living in better environments than their poorer counterparts but not relative to whites of the same economic status (Lacy, 2007; Pattillo, 2007; Pattillo-McCoy, 1999). Indeed, Pattillo-McCoy (1999) finds that many middle-class blacks are in close quarters with their disadvantaged peers and often have to deal with problems found in poor neighborhoods including declining physical conditions, increased crime, and downward mobility for future generations. For Hispanics, Vallejo (2012) finds that upward economic mobility is often times delayed for a generation in spite of economic advancement, resulting in reduced residential mobility.⁴

The influence of structural constraints in the housing market on such disparities is captured by a second theoretical model used to explain variation in household neighborhood outcomes, the place stratification model (Alba and Logan, 1991, 1993; Logan and Alba, 1993). The tenets of this model maintain that the residential opportunities of households are hierarchically ordered, particularly owner-occupied housing. Because housing is a commodity, it can be viewed through its use and exchange values. Majority-group members who own their housing view it in terms of the wealth that can be accrued through its exchange value. As such, majority-group members maintain social distance from other households that may jeopardize this wealth, thereby constraining access of minority households from the best residential locations (Logan and Molotch, 1987). Discriminatory actions, whether they manifest themselves in the search for housing or securing monies to obtain housing, are often built upon majority-group member racial and ethnic prejudices and are the most prominent actions used by powerful groups to constrain other households (Farley et al., 1994; Massey and Denton, 1993).

Results from the 2000 Housing Discrimination Study are consistent with the main propositions of the place stratification model (Ross and Turner, 2005; Turner et al., 2002). Paired tests conducted each between whites and blacks and whites and Hispanics reveal that whites are consistently favored over blacks and Hispanics in housing transactions in both the rental and sales markets. Particularly notable is the fact that between 1989 and 2000 there was a significant increase in the steering of black home buyers to predominantly black neighborhoods, the specific mechanism linking discrimination to residential segregation (Ross and Turner, 2005).

The “new” inequality found in lending patterns is also consistent with the tenets of the place stratification model and no doubt will lead to disparities between minorities and whites in neighborhood outcomes, regardless of their affluence or location in suburbs (Williams et al., 2005). That is, although blacks and Hispanics had achieved record levels of homeownership in the early part of the 2000s, they experienced significant levels of discrimination in financing the purchase of their homes, being more likely to have received high-priced, subprime loans (Avery et al., 2006). Such inequalities in financing have led blacks and Hispanics to be much more likely than whites to buy homes in predominantly minority neighborhoods. In 2000, 29 percent of subprime loans were made in minority neighborhoods, compared to 14 percent of conventional home purchase loans made by traditional lenders (Williams et al. 2005)

Studies have shown that the negative out-group residential preferences of whites, upon which discrimination in the housing and financial markets are built, have persisted into the 21^st century. Research has consistently found that whites maintain unfavorable out-group views of African Americans, regardless of their economic background, reinforcing residential segregation (Bobo and Zubrinsky, 1996; Charles, 2000; Farley et al., 1994; Krysan et al., 2009; Krysan and Farley, 2002). Similar negative white out-group views have been found to exist towards Hispanics, but not to the same degree for Asians, who themselves prefer not to live in communities with a large African American presence (Bobo and Zubrinsky, 1996). The fact that blacks and Hispanics often fare worse in their neighborhood outcomes, relative to whites, than Asians is no doubt partly attributable to the individual actions of whites based upon their negative out-group residential preferences.

The implications of racial bias in the housing market are apparent when looking at the difference in housing values. Various inquires have found that there is a housing “value gap” in existence between minority and white households (Collins and Margo, 2003; Fesselmeyer et al., 2013). In particular, black households in highly segregated communities have been found to have notably less value in comparison to those of other minorities (Flippen, 2004; Sykes, 2003). Collins and Margo (2003) have found that racial disparities in housing value have only gotten worse with time. These differences in housing value have been found to influence the differences in wealth across households by their race and ethnicity (Flippen, 2004).

Recent research has raised questions about the primacy of suburbs implicit in the spatial assimilation model. Several studies have found that African Americans and some immigrant ethnic groups living in suburbs can experience poverty and dilapidation comparable to their counterparts in central cities (Hanlon, 2010; Holliday and Dwyer, 2009; Logan et al., 2002; Patillo-McCoy, 1999). The work of Hanlon (2010), in particular, stresses that suburban communities that closely border cities, so called ‘inner ring’ suburbs, are notable for poverty and high minority concentrations, more similar in nature to central cities than outer-ring suburbs. In the past decade, concentrated poverty in suburbs has been steadily on the rise, outpacing its growth in cities (Kneebone et al., 2011). Thus, while suburban residence can lead to improved residential conditions for some, it is not a universal guarantee of better living for all (Alba et al., 1999; Friedman and Rosenbaum, 2007; Singer, 2008).

3. Hypotheses

Our study focuses on answering three major research questions, which were posed in the introduction. The preceding theoretical discussion suggests the following hypotheses in answering each of these questions.

1. Do racial and ethnic differences in neighborhood outcomes exist among middle-class and affluent households? The spatial assimilation model predicts that few racial and ethnic differences will exist in neighborhood outcomes between well-to-do whites and minorities because of their superior socioeconomic status. Those that do exist will be attributable to life cycle stage differences across groups as well as variation in other demographic characteristics. The place stratification model, on the other hand, suggests that more substantial differences in neighborhood outcomes will exist between whites and blacks and whites and Hispanics. However, relatively fewer differences will exist between whites and Asians. According to the theory, middle-class and affluent whites will be more likely to distance themselves from their black and Hispanic counterparts rather than their Asian counterparts because of the potential threat that might be incurred to the wealth generated from their exchange values by residing near black and Hispanic neighbors.

2. To the extent that differences exist, are they smaller in suburbs? The spatial assimilation model maintains that racial and ethnic differences in neighborhood outcomes should be smaller, if present at all, in suburbs. Mobility to the suburbs is seen as the endpoint of the spatial assimilation process and the focus here is on well-to-do households, thereby leaving little room for racial and ethnic disparities. The place stratification model suggests, however, that racial and ethnic disparities will be just as large in suburbs because of the operation of an explicit stratifying system built upon the profits generated from exchange values.

3. If racial and ethnic differences exist, do they disappear when controlling for relevant demographic and socioeconomic factors? According to the spatial assimilation model, the answer would be yes. The place stratification model predicts that such differences would persist, despite such controls. As discussed above, this finding would result from the fact that discrimination in the housing and financial markets exist, the precise mechanisms used by majority-group members to distance themselves from minority-group members. If the neighborhood quality of middle-class blacks and Hispanics is notably inferior to that of similarly-situated whites, but the neighborhood conditions of Asians is relatively similar to whites, it would likely result from the fact that white prejudices and the structural barriers built upon such prejudices could be constraining the housing choices of particular groups of middle-class minorities.

4. Data and Methods

Our analyses are based on data from the 2009 panel of the American Housing Survey (AHS), a multistage probability sample of approximately 50,000 housing units located throughout the United States that is surveyed every other year. We take advantage of data from the 2009 AHS because these data are recent and contain many indicators of neighborhood quality.⁵ Until now, the quantitative research done on middle- or upper-class residents on a national level has relied on data that are about 20 years old (e.g., Adelman 2005; Alba et al., 2000). In our analyses, we use sampling weights (scaled down to maintain un-weighted cell sizes) to correct for sampling design effects and potential under coverage.

We restrict our analyses here to focus on middle-class and affluent households. Following previous research, we define a household as middle class when its total income falls between two and four times the poverty threshold for a family of four⁶, it owns the home, and the householder has at least a college degree (Adelman, 2004; Alba et al., 2000). Affluent households are those whose income falls above four times the poverty level, own their homes, and whose householders have at least a college degree. In addition to restricting our analyses to middle-class and affluent households, we disaggregate our data on the basis of suburban location.⁷ One set of our analyses examines the middle-class and affluent householders in metropolitan areas, overall, while the second focuses on such householders who live in suburbs. As other researchers have done, we define suburbs as areas that are inside metropolitan areas but not within central cities (e.g., Alba and Logan, 1991; Alba et al., 1999).

To measure neighborhood conditions, our central dependent variables, we mainly rely on data from householders' answers to questions about the characteristics of the neighborhood immediately surrounding the housing unit that are indicative of physical quality, social disorder, and undesirable land uses.⁸ Although these variables are subjective in nature, the items ask respondents about the presence of particular physical or tangible conditions rather than respondents' opinions or attitudes. This helps address any potential variation in how different groups may perceive their neighborhood quality (Sampson 2012). The fact that the questions delineate the geographic area comprising the neighborhood, moreover, improves the chance that respondents' characteristics are much less likely to influence their responses to questions about the neighborhood (Lee and Campbell, 1997). These factors heighten the objectivity of respondents' reports.

Specifically, we use responses to questions asking about the presence of the following conditions within a half block of the building: abandoned buildings; buildings with bars on the windows; trash, litter, or junk in the streets, roads, empty lots or on any properties; and lack of nearby open spaces, such as parks, woods, farms, or ranches. We also use data from a question asking householders if crime was present in the neighborhood. However, the question does not restrict householders to considering crime within a half block of the building. In addition to the five indicators being analyzed one-by-one, we examine a summary index of neighborhood problems based upon the sum of these indicators. This index ranges in value from 0 (no negative conditions) to 5 (all negative conditions are present) and measures the extent to which undesirable conditions are concentrated in neighborhoods. Next, we include the respondent's rating of their neighborhood as a place to live, which is based on a scale from 1 to 10 with 10 being best. Last, we include the housing value of the home reported by households.

Our key independent variable is the householder's race/ethnicity. We use four categories of race/ethnicity (non-Hispanic white, non-Hispanic black, Hispanic, and Asian and Pacific Islander).⁹ We include controls for socioeconomic and demographic characteristics. While we have restricted our sample to middle-class and affluent households, variation exists in their household income and the householder's education and therefore, we control for these factors. Education is represented by a dummy variable indicating whether the householder has more than a college degree (graduate education), with having a college degree forming the reference group. Demographic factors are represented by the householder's age and three dummy variables indicating: (1) whether the householder is foreign born; (2) whether the household is headed by a married couple; and (3) whether children under 18 are present. We also control for the region within which the household lives.¹⁰

We conduct bivariate and multivariate analyses of these data. Bivariate analyses are used to identify how race and ethnicity affects household neighborhood conditions, overall, and within suburbs. Multivariate analyses will be used to identify how race and ethnicity affect household neighborhood conditions, after controlling for household socioeconomic status, demographic factors, and region of residence. Through these analyses, hypotheses derived from the spatial assimilation and place stratification models are tested. Because most of our dependent variables are dichotomous, the data will be analyzed using logistic regression models. However, for the index of neighborhood problems, neighborhood rating, and housing value variables described above ordinary least squares regression is used.¹¹

5. Results

Do racial and ethnic differences in neighborhood outcomes exist among middle-class and affluent households? Table 1 addresses this question, presenting the means for our main dependent variables, focusing on middle-class and affluent households in metropolitan areas, overall, and those in suburbs. With respect to the overall metropolitan results in columns 1 through 4, the data indicate that racial and ethnic differences exist in neighborhood outcomes among middle-class and affluent households. Well-to-do blacks and Hispanics live in lower-quality neighborhoods than middle-class and affluent whites on several dimensions, while well-off Asians generally live in neighborhoods of equal or even better quality than their white counterparts. With respect to the findings for blacks and Hispanics, 15.2 percent of middle-class or affluent blacks (column 2) and 12.2 percent of Hispanics (column 3) report the presence of buildings with barred windows within one-half block of their housing units, a rate more than twice as high as that of middle-class or affluent whites (5.24 percent, column 1). The average index of neighborhood problems for blacks and Hispanics is nearly one and a half times as high as that of whites. Middle-class and affluent Hispanics are more than twice as likely as whites to report the presence of abandoned buildings in their neighborhoods, but the difference between middle-class and affluent blacks and whites is not statistically significant. Well-to-do blacks and Hispanics have median housing values that are $80,000 and $40,000 less than well-to-do whites, respectively. Asians, however, are in a far superior position with a median housing value nearly $100,000 more than their middle-class and affluent white counterparts.

Table 1. Neighborhood Characteristics of Middle-Class and Affluent Households, 2009.

	Percent:
	All Households				Suburban Households
	Non-Hispanic Whites	Non-Hispanic Blacks	Hispanics	Asians	Non-Hispanic Whites	Non-Hispanic Blacks	Hispanics	Asians
Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Reference person reports within 1/2 block of housing unit:
Abandoned buildings	2.77	4.57	5.94**	1.66	2.31	2.23	4.70	1.67
Buildings with bars on windows	5.24	15.16***	12.20***	4.74	1.57	6.25***	7.98***	2.53
Trash or junk	4.31	5.94	5.96	3.06	2.96	4.61	4.30	1.99
No open spaces	56.92	63.92*	66.27**	63.46*	53.61	65.17***	65.25**	60.99*
Presence of crime in neighborhood	16.74	21.91*	20.89	12.09*	12.96	15.09	18.13	11.12
Average index score	0.86	1.12***	1.11***	0.85	0.73	0.93***	1.00***	0.78
Suburban	72.19	65.00**	63.66**	67.77	N/A	N/A	N/A	N/A
Neighborhood rating (10=best)	8.37	8.10***	8.25	8.31	8.45	8.28	8.40	8.35
Median housing value (in dollars)	280,000	200,000***	240,000***	375,000***	300,000	209,000***	250,000*	400,000***
N	5815	400	388	480	4350	260	254	341

^***p<.001;

^**p<.01;

^*p<.05 - differences refer to those between the minority group of interest and whites

Less pronounced but significant disadvantages exist for well-to-do blacks and Hispanics, relative to well-off whites, for the variables indicating a lack of open spaces, the presence of crime, and whether households live in suburbs. In addition, the average neighborhood rating of well-off blacks, 8.1, is significantly lower than that of well-to-do whites, 8.4. Middle-class and affluent Asians are more likely than well-off whites to report a lack of open spaces in their neighborhoods, but that is the only disadvantage that they experience. They report significantly less crime in their neighborhoods than their white counterparts. The only variable for which no racial and ethnic differences exist is that gauging the presence of trash, junk, or litter within one block of the housing unit.

With respect to suburbanites, the findings in columns 5 through 8 of Table 1 reveal fewer disadvantages for well-to-do blacks and Hispanics, relative to whites, than are present in the overall metropolitan analysis. No racial and ethnic disadvantages for these groups are present for the variables gauging the presence of abandoned buildings in the neighborhood and crime as well as average neighborhood rating, findings that depart from the overall metropolitan results. In addition, no significant difference is found between whites and minorities on the variable gauging the level of trash and junk in the neighborhood. However, middle-class and affluent blacks and Hispanics are more likely than similarly situated whites to live in neighborhoods with buildings with barred windows, a lack of open spaces, and more neighborhood problems. Thus, suburban residential location attenuates some but not all racial and ethnic differences in neighborhood outcomes. With respect to median housing value, suburban residence actually magnifies the disadvantages experienced by well-off blacks and Hispanics, relative to their white counterparts. Well-to-do whites have a median housing value that is $91,000 greater than that of blacks and $50,000 greater than for Hispanics. The results for Asians largely parallel those found in metropolitan areas, overall. Well-to-do Asians are significantly more likely than their white counterparts to report a lack of open spaces. However, on most other variables, well-off Asians enjoy comparable neighborhood conditions as middle-class and affluent whites.

Table 2 presents group differences in relevant demographic and socioeconomic characteristics for households in metropolitan areas (columns 1-4) and those in suburbs (columns 5-8). As in Table 1, significance tests are presented to evaluate minority differences from non-Hispanic whites, with the latter serving as the reference group. Despite generally living in the highest quality neighborhoods, middle-class and affluent non-Hispanic whites do not always exhibit higher levels of demographic and socioeconomic advantages predicted to result in superior residential circumstances (columns 1-4). For example, no significant differences are found between well-to-do minorities and whites in terms of the percent of householders with a college degree and more than a college education. While Hispanic and Asian householders are, on average, slightly younger than white householders, the differences are negligible, indicating that none of the groups are potentially constrained in their residential choices by their age. While a greater share of blacks reports the presence of children, relative to whites, the difference is not very large. Moreover, the shares of middle-class and affluent Hispanics and Asians reporting the presence of children at home are about equal to whites, thereby indicating that unequal residential circumstances are not likely to be due to differing residential preferences as they relate to children.

Table 2. Demographic and Socioeconomic Characteristics of Middle-Class and Affluent Households, 2009.

	Percent:
	All Households				Suburban Households
	Non-Hispanic Whites	Non-Hispanic Blacks	Hispanics	Asians	Non-Hispanic Whites	Non-Hispanic Blacks	Hispanics	Asians
Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Householder characteristics
Foreign Born	5.27	13.72***	33.86***	76.41***	4.96	14.47***	34.52***	77.21***
Age	48.66	47.58	46.02**	44.96***	48.65	47.13	45.77**	46.26**
Male	63.03	44.17***	63.38	67.01	63.10	43.19***	61.68	70.05
Education
At least college degree	59.65	57.21	62.28	56.70	61.15	52.61*	60.91	56.55
More than college degree	40.35	42.79	37.72	43.30	38.85	47.39*	39.09	43.45
Household/Housing unit characteristics
Duration in Household	11.12	10.89	9.07***	8.33***	11.20	9.79	8.94**	8.86***
Married household	75.26	60.57***	69.61	81.43*	78.32	60.72***	72.38	84.94*
Presence of kids under 18	41.66	49.11*	48.09	47.46	44.44	50.63	51.67	49.15
Total household income	110,000	93,000***	98,000**	120,000	110,000	100,000***	100,600	120,750*
Region
West	22.87	14.04***	32.58***	50.82***	19.74	14.10	31.87***	46.61***
Northeast	22.26	17.24	11.12***	15.17**	25.96	16.02***	11.22***	16.97***
South	31.23	54.21***	45.49***	22.33***	31.20	59.28***	45.78***	23.49*
Midwest	23.64	14.52***	10.81***	11.68***	23.10	10.60***	11.12***	12.93***
N	5815	400	388	480	4350	260	254	341

^***p<.001;

^**p<.01;

^*p<.05 - differences refer to those between the minority group of interest and whites

On the other hand, Table 2 reveals that there are minority-white differences among middle-class and affluent households on a few key social and demographic variables that could explain why minority-white differences exist in neighborhood conditions (columns 1-4). For example, among well-to-do minorities, 13.7 percent of blacks, 33.9 percent of Hispanics, and 76.4 percent of Asians are foreign-born householders, relative to only 5.3 percent of whites. The imbalanced distribution of foreign-born households for Asians and Hispanics could shape their knowledge about the housing market through their limited proficiency in English and how recent their arrival was.

Among middle-class and affluent black householders, less than half, or 44.2 percent, are males relative to 63 percent of white householders. In addition, 60.6 percent of well-to-do black households are married couples, relative to 75.3 percent of middle-class and affluent white households. Taken together, such white-black differences in family structure could potentially minimize the resources that black households have to secure housing in the best quality neighborhoods. Well-to-do blacks and Hispanics have median household incomes of $93,000 and $98,000, which although quite high, are significantly lower than that of well-to-do whites, $110,000.¹² However, no Asian-white difference exists in median household income, which is consistent with the findings in Table 1 of few differences in neighborhood conditions between these two, well-to-do groups. Moreover, well-to-do Asians are significantly more likely than middle-class and affluent whites to live in married couple families.

Among well-off households in suburbs, a similar pattern of findings exists in Table 2 (see columns 5-8). One difference between the overall results and those for suburbs exists for the educational variables. Among middle-class and affluent black householders in suburbs, 52.6 percent have a college degree compared to 61.1 percent of their white counterparts. However, among well-to-do blacks a significantly greater share of them has more than a college degree, relative to their white counterparts (47.4 percent versus 38.9 percent). This might reflect the fact that the upper echelon of middle-class and affluent blacks are attracted to suburbs, which is consistent with the tenets of the spatial assimilation model.

One other notable result in the suburban section of Table 2 is that while still significant, the magnitude of the differences in median household income between well-to-do whites and blacks and well-to-do whites and Hispanics is smaller, relative to the results for metropolitan areas, overall. Thus, as in the case with the result regarding black education above, there may be a positive selection of the most well-off, minority households to locate suburbs, which would be consistent with the findings for suburbs in Table 1 that show an attenuation in the differences in some neighborhood conditions between well-to-do whites and minorities.

Significant regional variation exists in the location of middle-class and affluent whites and minorities, overall and in suburbs, which may also be important in explaining minority-white differences in neighborhoods conditions observed in Table 1. Well-off blacks are significantly more likely to live in the South, relative to their white counterparts, with over half of all middle-class and affluent blacks living in the South. Well-off Hispanics are significantly more likely than their white counterparts to live in the West and the South but are significantly less likely than well off whites to live in the Northeast and Midwest. Well-to-do Asians are significantly more likely to live in the West than well-to-do whites.

Given that racial and ethnic differences exist in neighborhood conditions, even among well-to-do households, the question to which we now turn is whether they disappear when controlling for relevant demographic and socioeconomic factors. Table 3 addresses this question by presenting the results of the logistic and OLS regression models predicting our neighborhood outcomes in metropolitan areas as a whole and Table 4 does the analysis specifically for suburbs. First, we focus on the coefficients for the race and ethnicity variables and then examine the impact of background characteristics on neighborhood conditions.

Table 3. Logistic Regression Models of Neighborhood Conditions of Middle-Class and Affluent Households, 2009 (weighted).

	Abandoned buildings	Bars on Windows	Trash	No Open Spaces	Crime	Index of Problemsa	Neighborhood Ratinga	Housing Valuea	Suburban Location
Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
Race/ethnicity (ref. white)
Black	0.570**	1.080***	0.236	0.211*	0.184	0.195***	-0.189**	-48602**	-0.329**
	(.250)	(.165)	(.2621)	(.106)	(.126)	(.039)	(.068)	(16062)	(.109)
Hispanic	0.854***	0.755***	0.245	0.320**	0.164	0.199***	-0.020	-42725*	-0.282*
	(.246)	(.191)	(.240)	(.117)	(.140)	(.042)	(.075)	(17586)	(.118)
Asian	-0.318	-0.412	-0.415	0.264*	-0.342*	-0.011	-0.037	48840**	-0.031
	(.408)	(.261)	(.314)	(.119)	(.170)	(.044)	(.077)	(18196)	(.125)
Householder attributes
Foreign born	-0.485	0.085	-0.159	0.025	-0.236	-0.040	0.057	24278	-0.081
	(.288)	(.175)	(0.227)	(.092)	(.126)	(.034)	(.060)	(14114)	(.097)
Age	-0.019**	-0.016**	-0.034***	-0.004	-0.015***	-0.006***	0.023***	4451.127***	0.008**
	(.007)	(.005)	(.006)	(.002)	(.003)	(.001)	(.002)	(383.486)	(.003)
Male	0.085	-0.240*	-0.114	0.152**	-0.213**	-0.006	-0.137***	-18661*	-0.086
	(.152)	(.111)	(.123)	(.052)	(.069)	(.020)	(.035)	(8121.506)	(.058)
More than college	-0.074	0.104	-0.213	0.037	-0.110	-0.012	0.003	13198	-0.170**
	(.148)	(.108)	(.125)	(.051)	(.068)	(.019)	(.033)	(7844.468)	(.055)
Married household	0.112	-0.345**	-0.179	-0.250***	0.021	-0.085**	0.176***	10021	0.422***
	(.175)	(.122)	(.139)	(.063)	(.082)	(.023)	(.041)	(9643.645)	(.066)
Presence of kids < 18	-0.307	-0.356**	-0.343*	-0.073	0.141	-0.042	0.102**	56973***	0.319***
	(.165)	(.127)	(.137)	(.057)	(.075)	(.021)	(.038)	(8884.454)	(.063)
Total household income	-0.002	0.000	-0.000	-0.000	-0.000	-0.000	0.001***	1253.432***	-0.000
	(.001)	(.001)	(.001)	(.000)	(.000)	(.000)	(.000)	(40.842)	(.000)
Housing unit location
Located in suburb	-0.578***	-2.154***	-0.948***	-0.394***	-0.829***	-0.414***	0.248***	12061	N/A
	(.146)	(.118)	(.120)	(.056)	(.067)	(.020)	(.036)	(8487.097)	N/A
Duration in Household	0.014	0.021**	0.018*	0.009**	0.010*	0.006***	-0.012***	-873.394	-0.003
	(0.009)	(0.007)	(.008)	(.003)	(.004)	(.001)	(.002)	(492.251)	0.003
Region (ref. West)
Northeast	-0.188	0.268	0.107	-0.021	-0.485***	-0.049	0.175**	-108600***	1.016***
	(.215)	(.145)	0.172	(.074)	(.107)	(.028)	(.049)	(11507)	(.086)
South	-0.362	-0.318*	-0.155	0.148*	0.096	0.016	0.012	-203294***	0.415***
	(.189)	(.133)	(.157)	(.067)	(.084)	(.025)	(.044)	(10271)	(.070)
Midwest	-0.109	-1.245***	-0.302	-0.114	-0.163	-0.113***	0.141**	-230379***	0.289**
	(.200)	(.193)	(.179)	(.073)	(.096)	(.027)	(.048)	(11361)	(.077)
Intercept	-2.026***	-0.624*	-0.569	0.782***	-0.250	1.504***	6.898***	100966***	-0.109
	(0.367)	(.267)	(.296)	(.136)	(.175)	(.050)	(.089)	(20846)	(.140)
Model chi-squareb/R2	55.37***	608.11***	130.50***	140.62***	271.33***	0.080	0.052	0.224	289.47***
N					7083

^***p<=.001;

^**p<=01;

^*p<=.05

^aOLS regression is used here.

^bThe degrees of freedom are 15 in models 1 through 5; in model 9, they are 14.

Table 4. Logistic Regression Models of Neighborhood Conditions of Middle-Class and Affluent Households in Suburbs, 2009 (weighted).

	Abandoned buildings	Bars on Windows	Trash	No Open Spaces	Crime	Index of Problemsa	Neighborhood Ratinga	Housing Valuea
Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Race/ethnicity (ref. white)
Black	0.101	1.322***	0.435	0.411**	-0.003	0.161***	-0.090	-41054*
	(.417)	(.290)	(.304)	(.129)	(.175)	(.042)	(.081)	(19440)
Hispanic	0.941**	1.524***	0.421	0.419**	0.250	0.238***	0.026	-39260
	(.331)	(.289)	(.340)	(.142)	(.181)	(.046)	(.090)	(21417)
Asian	0.239	0.187	-0.115	0.242	-0.214	0.047	-0.109	59808**
	(.499)	(.428)	(.459)	(.140)	(.212)	(.046)	(.091)	(21600)
Householder attributes
Foreign born	-0.993*	0.141	-0.521	0.126	-0.062	-0.008	0.052	13790
	(.443)	(.299)	(.355)	(.109)	(.158)	(.036)	(.070)	(16795)
Age	-0.001	-0.018	-0.020*	-0.003	-0.010*	-0.003**	0.022***	4558.258***
	(.009)	(.010)	(.008)	(.003)	(.004)	(.001)	(.002)	(446.900)
Male	0.072	-0.376	-0.051	0.183**	-0.280**	0.004	-0.145**	-16734
	(.201)	(.204)	(.172)	(.061)	(.087)	(.020)	(.040)	(9431.038)
More than college degree	0.021	-0.242	-0.255	0.034	-0.054	-0.009	-0.008	11166
	(.192)	(.206)	(.176)	(.059)	(.087)	(.020)	(.038)	(9110.303)
Married household	0.454	0.092	-0.055	-0.243**	0.123	-0.038	0.158**	25644*
	(.262)	(.233)	(.205)	(.075)	(.112)	(.025)	(.048)	(11529)
Presence of kids < 18	0.075	-0.645**	-0.120	-0.052	0.318***	0.009	0.108*	46897***
	(.217)	(.233)	(.185)	(.066)	(.096)	(.022)	(.043)	(10252)
Total household income	-0.004**	0.001	-0.002*	-0.000	-0.000	-0.000*	0.001***	1206.386***
	(.001)	(.001)	(.001)	(.000)	(.000)	(.000)	(.000)	(47.492)
Housing unit location
Duration in Household	0.009	0.041**	0.007	0.008*	0.009	0.004**	-0.010***	-993.576
	(.012)	(.011)	(.011)	(.003)	(.006)	(.001)	(.002)	(572.779)
Region (ref. West)
Northeast	-0.340	-1.229***	-0.086	-0.001	-0.657***	-0.105**	0.138*	-148559***
	(.277)	(.322)	(.237)	(.084)	(.133)	(.028)	(.055)	(13174)
South	-0.365	-0.406	-0.165	0.182*	0.061/	0.027	-0.033	-221682***
	(.251)	(.225)	(.221)	(.079)	(.109)	(.026)	(.052)	(12306)
Midwest	-0.073	-1.418**	-0.363	-0.040	-0.344**	-0.088**	0.150**	-249636***
	(.263)	(.377)	(.258)	(.087)	(.128)	(.029)	(.057)	(13606)
Intercept	-3.562***	-2.529***	-1.949***	0.257	-1.336***	0.917***	7.215***	127438***
	(.523)	(.494)	(.425)	(.157)	(.232)	(.052)	(.102)	(24404)
Model chi-square/R2	28.114*	105.211***	25.906*	70.540***	90.222***	0.019	0.039	0.216
N				5205

^***p<=.001;

^**p<=.01;

^*p<=.05

^aOLS regression is used here.

^bThe degrees of freedom are 14 in models 1 through 5.

The answer to our question of whether minority disadvantages disappear in relation to whites after controlling for relevant socioeconomic and demographic characteristics is no. Among middle-class and affluent households, many of the racial and ethnic differences in neighborhood conditions persist, even in the presence of controls for other relevant factors. Table 3 shows that well-to-do blacks are significantly more likely than well-to-do whites to live in neighborhoods with abandoned buildings, buildings with barred windows, a lack of open spaces, and a greater index of neighborhood problems, controlling for relevant demographic and socioeconomic variables. For example, the odds of middle-class and affluent blacks in living in neighborhoods with buildings that have barred windows is 3.03 times (e^1.08 =3.03) that of their white counterparts with controls in the model. In addition, relative to well-off whites, well-to-do blacks rate their neighborhoods lower, have lower housing values, and are less likely to live in suburbs. Controlling for other factors, the neighborhood rating of well-off blacks is .19 units less than that of their white counterparts; their housing values are $48,602 lower than well-to-do whites; and the odds of middle-class and affluent blacks are .72 times the odds of their white counterparts to live in suburbs.

For the most part, the results in Table 3 are consistent with the white-black disparities in neighborhood conditions reported among well-to-do households in Table 1. However, there are two differences. First, a black-white difference emerges in the presence of abandoned buildings in the neighborhood, with the odds of having such buildings for well-off blacks being 1.79 times the odds of well-off whites. Second, the significant difference in the share of well-to-do blacks and whites reporting the presence of crime is not significant, controlling for relevant socioeconomic and demographic characteristics in the model. This could be explained by the differences in income that exist between the two groups as well as the differences in the shares of these groups living in suburbs.

Table 3 shows that the results pertaining to middle-class and affluent Hispanics largely mirror those of well-to-do blacks. Well-off Hispanics are significantly more likely than well-to-do whites to live in neighborhoods with abandoned buildings, buildings with barred windows, a lack of open spaces, and a greater index of neighborhood problems, controlling for relevant demographic and socioeconomic variables. In addition, well-off Hispanics have lower housing values than well-off whites and are significantly less likely to live in suburbs. The main difference between the results for Hispanics, as compared to that for blacks, is that the neighborhood rating of well-to-do Hispanics does not significantly differ from that of well-to-do whites.

Table 3 reveals, however, that the results for middle-class and affluent Asians, relative to whites, diverge greatly from those of their black and Hispanic counterparts. Consistent with the results in Table 1, column 4 shows that well-to-do Asians are significantly more likely to lack open spaces than well-to-do whites, controlling for other factors. However, this is the only neighborhood characteristic on which Asians are disadvantaged, relative to whites. Examining columns 5 and 8 of Table 3 reveals that well-off Asians are significantly advantaged over well-off whites. Middle-class and affluent Asians are .71 times as likely as well-off whites to report the presence of crime in their neighborhoods and have housing values that are $48,840 greater than their white counterparts, controlling for other relevant socioeconomic and demographic factors.

Table 4 presents results from the logistic and OLS models predicting the locational attainment for middle-class and affluent households residing in suburbs. Many of the results found for metropolitan areas continue to be present in the suburban context. Controlling for relevant factors, middle-class and affluent black and Hispanic households are significantly more likely than well-to-do whites to live in neighborhoods with buildings with barred windows, a lack of open spaces, and a greater index of neighborhood problems. In addition, the housing values of well-off blacks living in suburbs is $41,054 less than that of middle-class and affluent whites, controlling for socioeconomic and demographic characteristics. For Hispanics, no difference exists on this variable. However, well-off Hispanics are 2.56 times as likely as middle-class and affluent whites to live in neighborhoods with abandoned buildings, controlling for other relevant factors.

Of particular note is the fact that, with the exception of housing values, the disadvantages faced by well-off Hispanics in their locational attainment became larger rather than smaller, relative to well-to-do whites, in the suburban models. For example, controlling for relevant factors, in suburbs, the odds of middle-class and affluent Hispanics living in neighborhoods with buildings with barred windows were 4.59 times the odds of well-to-do whites. However, in the overall metropolitan model, the odds of well-off Hispanics living in neighborhoods with buildings with barred windows were 2.13 times the odds of well-to-do whites.

As was the case with the results for overall metropolitan areas, well-off Asians in suburbs are much more equal to their white counterparts than middle-class and affluent blacks and Hispanics. Interestingly, in suburbs, middle-class and affluent Asians have housing values that are $59,808 higher than their white counterparts, which is an even greater advantage to Asians than in the overall metropolitan model.

All told, then, the results in Tables 3 and 4 reveal the persistent effects of race and ethnicity in predicting neighborhood conditions for middle-class and affluent households, in metropolitan areas and their suburbs. In general, middle- and upper-class whites occupy a superior position in the housing market, but they are often joined by Asians who on occasion supersede their position. With the exception of the outcome crime in the overall metropolitan model, most of the disadvantages faced by middle- and upper-class blacks and Hispanics, relative to well-off whites, persist after controlling for key demographic and socioeconomic characteristics. The maintenance of these disadvantages in the multivariate analyses in Tables 3 and 4 are, therefore, more consistent with predictions under the place stratification model than those under the spatial assimilation model. In fact, for well-to-do Hispanics, these neighborhood disadvantages appear to worsen in suburbs, relative to well-off whites, with the exception of the results for housing values.

With respect to the effect of demographic and socioeconomic characteristics, we generally find support for the tenets of the spatial assimilation model. Specifically in the overall analysis presented in Table 3, well-off households that are headed by householders who are older, married, and with children under 18 and that live in suburbs are consistently more likely to live in better quality neighborhoods than well-off households headed by householders who are younger, not married, with no children under 18, and that live in central cities, respectively. Households that have householders with more than a college degree and with greater total incomes are also more likely than those with lower incomes and less education to live in better quality neighborhoods, although these predictors are not significantly related to all neighborhood conditions. Middle-class and affluent households that have lived in their housing units for a long period of time consistently live in poorer quality neighborhoods than their counterparts living in their homes for shorter periods of time, suggesting diminishing returns to owning homes for longer periods of time. The results for householder sex are a bit less straightforward than is the case with the other independent variables, with being male sometimes facilitating household access to better quality neighborhoods and other times not.

With respect to the analysis of locational attainment in suburbs featured in Table 4, the effects of the demographic and socioeconomic variables are relatively similar to the findings in Table 3. Controlling for relevant factors, middle-class and affluent households headed by older householders, those who are married, and with more income are consistently more likely to reside in neighborhoods with higher levels of quality than those headed by younger householders, who are not married, and have less income, respectively. Like the results in Table 3, the findings here support hypotheses derived under the spatial assimilation model. A bit less consistent, however, are the effects of sex of the householder and the presence of kids under 18, which sometimes facilitate well-to-do household access to better neighborhoods and sometimes do not. Controlling for other factors, well-off households living in their owned units for longer periods of time generally have poorer neighborhood quality than those living in their units for shorter periods of time, which is consistent with the findings reported in Table 3.

With respect to region, the results in Tables 3 and 4 show that middle-class and affluent households living in the Northeast, South, and Midwest tend to live in neighborhoods of higher quality than those in the West, but there are exceptions. For example, well-off households in the Northeast, South, and Midwest have housing values that are significantly lower than those in the West, which likely relates to the considerable variation that exists by region in terms of the cost of living. Unfortunately, due to data limitations, we cannot control for metropolitan-level characteristics that might affect well-off household access to better residential outcomes within these areas.

6. Discussion and Conclusions

The primary goal of this paper was to determine if middle- and upper-class status translates into equal neighborhood conditions for households by their race and ethnicity in an era of declining racial residential segregation. To fulfill this overarching goal, the analysis focused on answering three main questions. First, did racial and ethnic differences in neighborhood outcomes exist among middle-class and affluent households? Consistent with expectations derived under the place stratification model, our descriptive analysis revealed that well-to-do black and Hispanic households experienced significantly lower levels of neighborhood quality than their white counterparts. However, with respect to well-off Asians, the descriptive analysis revealed virtually no disadvantages in neighborhood outcomes, relative to well-off whites, thereby being more consistent with hypotheses under the spatial assimilation model.

Second, to the extent that differences existed, were they smaller in suburbs? In our descriptive analyses, we found the answer to be yes and no. Consistent with expectations derived under the spatial assimilation model, in the suburban analysis, differences between middle-class and affluent whites and well-off Hispanics were no longer significant in terms of the presence of abandoned buildings in neighborhoods and the presence of crime. Similarly, differences between well-to-do whites and blacks were no longer significant with regard to the presence of crime and neighborhood satisfaction. However, consistent with hypotheses under the place stratification model, significant disadvantages in neighborhood quality existed for well-off blacks and Hispanics, relative to whites, in terms of the presence of buildings with barred windows in the neighborhood, the lack of open spaces, the index of neighborhood problems, and median housing values. Moreover, although the absolute levels of disadvantage decreased in the suburban analysis, the magnitude of the differences with whites remained about the same.

Finally, did the racial and ethnic differences in residential attainment disappear when controlling for relevant demographic and socioeconomic factors? We found the answer to be no in both the overall, metropolitan analysis and the suburban analysis, particularly for well-to-do blacks and Hispanics, thereby being more consistent with the tenets of the place stratification model than the spatial assimilation model. On some of the outcomes, the magnitude of the disadvantages between middle-class and affluent blacks and Hispanics declined, but it remained statistically significant. Interestingly, for well-off blacks in the overall metropolitan analysis, a significant disadvantage emerged in the multivariate analysis that was not found in the descriptive analysis, with such blacks being more likely than well-off whites to live in neighborhoods with abandoned buildings.

Taken together, while the findings here provide support for hypotheses from both the spatial assimilation and place stratification models, the weight of the evidence leans somewhat more toward the latter. By focusing on middle- and upper-class minorities and whites and those that live in suburbs, we removed much of the variation in the factors that tend to explain why minorities are disadvantaged in their residential circumstances, relative to whites, under the spatial assimilation model. Consistent with the tenets of the spatial assimilation model, we found that middle-class and affluent Asians experience little differences in their neighborhood conditions relative to their white counterparts, and in some instances, such Asians surpass well-off whites in having better neighborhood conditions.

However, we found that even among well-off households, blacks and Hispanics are significantly more likely than well-off whites to experience disadvantages in their neighborhood conditions, consistent with the findings of previous research that uses older data and were conducted when segregation between blacks and whites was higher than contemporary levels (Adelman 2004, 2005; Alba et al. 2000). For metropolitan areas, overall, we found that middle-class and affluent blacks rate their neighborhoods as places to live significantly lower than their white counterparts. Although not large in magnitude, this difference suggests that middle-class and affluent blacks are not as satisfied with their residential circumstances relative to their white counterparts, and perhaps their location in relatively disadvantaged neighborhoods may be due to some constraints on their housing choices. In addition, the fact that blacks in suburbs are found to be more highly educated than whites suggests that they need to achieve more in comparison to whites to attain similar residential conditions. This finding echoes recent research, which shows that white residents will only accept blacks into suburban neighborhoods if they have “middle class values” (Nagel 2013, p. 635). Another very important result shown in our analysis is that the access to resources was not “equalized” in suburban locations. This finding argues against the notion that suburbs symbolize the land of opportunity and equal access for all households, contrary to the tenets of the spatial assimilation model.

These findings have important theoretical implications for the study of residential segregation. The results of our analyses support existing quantitative and qualitative research on minority middle class householders and show that race and ethnicity continue to be central forces in shaping disparities household residential attainment (e.g., Adelman 2004, 2005; Alba et al. 2000; Pattillo-McCoy, 1999; Vallejo, 2012). While residential segregation has declined, particularly between blacks and whites, middle-class and affluent blacks and Hispanics remain disadvantaged in their locational attainment, relative to whites. Thus, despite the fact that minority households have attained more “even” distributions with whites, disparities remain in their residential attainment, relative to whites, thereby warranting further attention to the study of the quality of neighborhoods of blacks and other minorities.

Although our cross-sectional analyses cannot pinpoint the exact mechanisms underlying these significant racial and ethnic disparities, it is clear that class status only takes minority households so far in terms of their residential attainment. Instead, it is likely that continuing racial discrimination in the housing market and mortgage-lending industry constrains the residential choices of affected households (Turner et al., 2002). In particular, the fact that blacks had been more likely than whites to acquire subprime loans is likely to have contributed to such disparities (Avery et al. 2006; Calem et al. 2004). Housing tenure is an important feature of socioeconomic status that deserves further attention as it shapes the locational attainment of well-off households. The current and future well-being of home owners are tied to the wealth accrued through their housing. If the neighborhood conditions where well-off minorities live are of lower quality than that of well-off whites, then future research needs to be done focusing on identifying the precise mechanisms causing such disparities. In addition, because these data are from 2009, and therefore from the midst of the Great Recession, future research should examine whether these disparities have persisted more recently.

The study of white residential preferences would also be important to link to these findings. The fact that well-to-do Asians generally live in neighborhoods of equal and sometimes better quality than well-off whites but that well-off blacks and Hispanics experience many disadvantages, relative to well-to-do whites, suggests that white negative out-group preferences also likely play a role, consistent with the evidence in previous research (Bobo and Zubrinsky, 1996; Charles 2000; Farley et al., 1994; Krysan et al., 2009; Krysan and Farley, 2002).

Future research should also seek to correct some of the other limitations present in the current analysis. For example, more detailed work needs to be done contrasting the various groups within each race and ethnicity. Hispanic householders constitute a wide range of nationalities with varying levels of segregation, and it can be expected that the various well off residents live in different community conditions based in part on these backgrounds (Iceland, and Nelson, 2008; Logan and Turner, 2013). In addition, future work should seek to better characterize the nature of suburban residential attainment. Our definition of suburbs is broad and as such fails to distinguish the variation in suburban locations – such as inner ring suburbs versus peripheral exurbs.

In spite of these limitations, the analyses here revealed that the neighborhood quality of middle-class and affluent black and Hispanic households remains, for the most part, inferior relative to whites of the same socioeconomic standing, despite the progress made in terms of desegregation. Future research on residential segregation should not just focus on the residential distributions of minorities and whites but should instead focus on the quality of such neighborhoods. Given that minority births are outpacing white births, examining the quality of where middle-class and affluent minority households with children live is particularly important because minorities are collectively becoming the majority group in American society (Lichter, 2013). If such inequalities persist, this will no doubt have implication for how inequality may be transmitted to future generations of racial and ethnic minorities.

Highlights.

• Middle- and upper-class status do not erase disadvantages in neighborhood conditions for black and Hispanic households, relative to whites.

• Suburban residence does not “equalize” access to better neighborhood conditions for middle-class and affluent blacks and Hispanics, relative to whites.

• Asians generally enjoy neighborhood conditions that are relatively equal, and often superior, to those of whites.

• Declines in residential segregation do not necessarily translate into better quality neighborhoods for middle-class and affluent minorities, particularly blacks and Hispanics, relative to whites.