The Dual Role of Race and Immigration Among Ascending Neighborhoods in U.S. Metropolitan Areas

Abstract

The racial/ethnic and socioeconomic diversity of immigrants to the United States has grown since the 1990s, along with growing neighborhood socioeconomic inequality. Few studies explain how race/ethnicity and immigration interact to influence neighborhoods’ socioeconomic ascent (increases in residents’ household income, rents, property values, educational and occupational attainment). We draw on two decades of Census microdata, 1990–2010, to create synthetic cohorts that allowed us to measure the interactions between race/ethnicity and immigration on neighborhood ascent. We find that Black immigrants contributed the most to socioeconomic ascent in predominantly Black neighborhoods, while non-Hispanic newcomers—both immigrant and U.S.-born—contribute the most to predominantly Hispanic ascendant neighborhoods. All combinations of White/nonwhite and immigrant/U.S.-born contribute to the ascent of White neighborhoods with White immigrant newcomers to the neighborhood having the highest SES. The presence of White residents increases in nonwhite ascendant neighborhoods, especially neighborhoods in new immigrant destinations. The foreign-born presence grew more in all non-ascendant neighborhoods relative to ascendant neighborhoods of all types. Our analysis shows that immigration influences neighborhood socioeconomic ascent differently across racial/ethnic groups and in neighborhoods with differing initial racial composition. Our findings underscore the need to account for both race/ethnicity and immigration in order to explain the population processes underlying neighborhood change.

Introduction

Two population processes have reconfigured neighborhoods of urban areas in the United States since the 1990s. Immigration changed the racial and ethnic composition of the American population, particularly in large, urban areas and (Hall, 2013; Logan & Zhang, 2010; Massey, 2008; Singer, 2013), changing patterns of neighborhoods racial integration as a result (Bader & Warkentien, 2016; Logan & Zhang, 2010; Pais, 2020). Simultaneously, neighborhood economic inequality has increased during the period as well (Hwang, 2016; Reardon & Bischoff, 2011). Economic segregation, both within and across metropolitan areas, increased as American society has become more economically unequal.

Initial evidence shows that the two trends, immigration and growing economic inequality, may relate to one another. The influx of human, economic, and social capital that immigrants bring to neighborhoods may have spurred revitalization of neighborhoods that helped ignite gentrification. Changes among immigrants entering the U.S. may have contributed to this trend, having once been primarily Mexican and Central American immigrants with few socioeconomic resources to becoming more socioeconomically diverse and hailing from all over the globe (Hall et al., 2011; Hamilton, 2019; Massey, 2008; Tesfai et al., 2020; Waters & Pineau, 2015), including immigrants with higher socioeconomic status (Fry, 2015; Krogstad & Radford, 2018). The pervasiveness of American culture abroad may also have exported the “American Dream” of homeownership and a suburban lifestyle that led immigrants to take part in movement out to new socioeconomically ascendant suburban developments in American suburbs.

The growing economic and racial diversity of immigrants entering the U.S. calls into question whether the association between immigration and the socioeconomic ascent of neighborhoods differs by racial/ethnic groups, and whether the relationship plays out differently across different communities of color. Given longstanding racial inequality in American society and racist perceptions of places, neighborhoods receiving White immigrants may have been more likely to ascend than those with nonwhite immigrants. Similarly, the economic standing of immigrants and their ability to participate in the general ascent of neighborhoods may have also differed by race.

This study investigates the population processes underlying neighborhood socioeconomic ascent—the upwards trajectory of a neighborhood based on economic increases in its housing market and proportional increases in their resident population’s socioeconomic status (SES) (Owens, 2012; Owens & Candipan, 2019). Neighborhoods can experience an increase in their SES when higher-SES populations move into the neighborhood, when the existing resident population experiences an increase in SES, or both (Van Criekingen & Decroly, 2003). Our study focuses on whether, and under what circumstances, immigrant groups contribute socioeconomically to neighborhood ascent, and whether this differs by race. We further argue that the demographic processes involved in ascent is not a uniform process and instead differs in historically Black, Hispanic, Asian, and White neighborhoods.

We take advantage of confidential individual-level data from the Census over two decades, from 1990 to 2010, to estimate the interaction of race and immigration on patterns of neighborhood ascent that have been impossible using publicly available tabulations. We show that although immigration matters during the ascent process, it was not always associated with neighborhood economic improvement. Across ascendant neighborhoods of all types—Black, Hispanic, Asian, and White—the share of immigrants tended to grow slower than in neighborhoods that did not experience socioeconomic ascent. Group contributions to neighborhood ascent varied for Hispanic and Black immigrants relative to their U.S.-born counterparts, with Black immigrants generally having higher SES, and Hispanic immigrants having lower SES, than their U.S.-born counterparts. Nonetheless, the SES of Black and Hispanic residents (immigrant and native households alike) generally lagged the SES of their non-Black and non-Hispanic neighbors (in Black and Hispanic ascendant neighborhoods, respectively). Overall, newcomers to ascendant neighborhoods tended to have higher SES than those that have resided in the neighborhood longer, regardless of race and immigrant status, suggesting that neighborhood ascent was driven by the entry of higher-SES in-movers more often than via incumbent upgrading by long-time residents.

Background

Racial/Ethnic Composition, Immigration, and Neighborhood Socioeconomic Ascent

Immigration since the 1990s has reshaped the racial/ethnic and foreign-born composition of neighborhoods across the U.S. New immigrants entered a society deeply divided by race. Both systemic and individual racism have created, and continue to reinforce, social structures that lead poverty to concentrate in neighborhoods with fewer White residents. These structures led to the devaluation of properties in neighborhoods of color and growing shares of nonwhite residents tend to correlate with increasing rates of poverty, declining incomes, and lower wealth.

Gentrification has been an exception to the general pattern of disinvestment in nonwhite neighborhoods (Brown-Saracino, 2017). In the U.S., racial differences often, but not always, underlie the process. The typical process described often involves new White residents moving into neighborhoods with few existing Whites, displacing people of color. The relative real value that new entrants attain by moving into gentrifying neighborhoods often comes about because of the historical disinvestment in communities of color (Smith, 1996).

Gentrification does not, however, only involve White newcomers moving into neighborhoods with predominantly nonwhite residents. Past work in Chicago, New York, Washington, and elsewhere describe Black- and minority-led gentrification (Boyd, 2005; Davila, 2004; Hyra, 2008; Lacy, 2004; Pattillo, 2005; Taylor, 2002). Bader (2011) finds that high-SES Black households in Chicago, long excluded from White and lower-poverty neighborhoods due to discrimination, are more likely than White households to welcome economic reinvestment in predominantly nonwhite neighborhoods as a way to access improved municipal services, housing, and amenities (see also Freeman, 2006).

Nor is gentrification the only manner in which a neighborhood could ascend socioeconomically. Immigration may be another force to explain the economic ascendance of some neighborhoods. Cities with higher concentrations of immigrants are associated with population growth, increased demographic diversity, regional economic stability, and lower crime (Ruther et al., 2018; Sampson, 2008). Immigrant neighborhoods in these cities could also be sites of economic ascent among existing residents. Results from recent national studies find that predominantly Hispanic and Asian neighborhoods have experienced rising rates of socioeconomic ascent since 1970 and attained similar rates to those of predominantly White neighborhoods since 1990 (Owens, 2012; Owens & Candipan, 2019).

The spatial assimilation theory has long posited that immigrant enclaves provide opportunities for new immigrants to assimilate into the U.S. society and economy (Charles, 2003). Economic ascendance was often marked by moves to whiter and wealthier neighborhoods (Iceland & Scopolliti, 2008). But many upwardly mobile immigrants choose to stay in ethnic enclaves, potentially leading to the economic ascent of those enclaves (Logan et al., 2002), both through their own economic ascent and through the increased social capital to help newer arrivals improve their economic lot as well.

Compositional effects might also explain why immigration spurs neighborhood economic ascendance. The socioeconomic composition of immigrants has changed in recent decades (Tesfai et al., 2020). Early wave immigrants tended to have less SES and fewer skills, while those arriving in the 2000s tended to have higher levels of educational attainment (Krogstad & Radford, 2018). Since higher-SES immigrants tend to migrate to “new destinations”, and they tend to move into neighborhoods with both an existing immigrant presence and a disproportionate number of U.S.-born residents of their same race, the flow of immigrants with more skills and educational attainment may be contributing to neighborhood socioeconomic ascent (Lichter & Johnson, 2009; Lichter et al., 2010; Ruther et al., 2018). Federal immigration policies that facilitated entry to high-skilled professionals and investors, particularly since the 1990s, also meant that increases in capital flowed to many of these cities alongside higher-SES immigrants. Immigration may spark initial economic ascendance that then ends up being followed by gentrification.

Some evidence suggests that neighborhood ascent is more likely in ethno-racially diverse neighborhoods, particularly new immigrant destinations that attract higher-SES Asian and Hispanic populations (Hwang, 2015; Lichter & Johnson, 2009). However, the type of diversity matters. Whether and how neighborhoods experience socioeconomic ascent is complex and depends on the initial racial composition of neighborhoods (Owens & Candipan, 2019; Rucks-Ahidiana, 2020). Therefore, the interaction between race and immigration during ascent processes may play out differently across of different types of ascendant neighborhoods.

Race of Immigrants

Whether immigration leads to neighborhood ascent may depend not only on the racial composition of the existing neighborhood, but also on the racial composition of immigrants as well. Spatial assimilation theories are typically associated with the residential attainment of Hispanic and Asian households. The focus was understandable given that sending countries have, for most of the post-reform era, been in Latin America and Asia.

Black households have traditionally not experienced spatial assimilation. Instead, discrimination and historic inequality has kept Black households from achieving upward mobility and entry into whiter neighborhoods (Logan & Alba, 1993; Charles, 2003). But this work has largely failed to disaggregate housing and socioeconomic attainment by nativity (Cort, 2011; South & Crowder, 1998). Previous research suggests that not disaggregating immigrant and native-born experiences can mask the underlying heterogeneity between the two groups (Crowell & Fossett, 2019; Tesfai, 2019). In particular, the experience of Black immigrants has been largely overlooked (Hamilton, 2014, 2019; Hamilton et al., 2018; Tesfai, 2019). As a consequence, we have little insight regarding whether this new group of Black immigrants experiences trajectories more like the spatial assimilation model of their Hispanic and Asian immigrant counterparts or the place stratification model of their native-born Black counterparts.

While some of the research above suggests that Black immigrants may experience outcomes more like other immigrant groups, other research suggests that race may still play a large role on the experience of Black immigrants. Hwang (2015) found that the entry of Asian and Hispanic residents into predominantly Black neighborhoods increased the chances that the neighborhood would gentrify, results that seem to indicate that wealthy whites prefer neighborhoods where Asian and Hispanic residents may “buffer” them from contact with Black residents (Logan & Zhang, 2010; Parisi et al., 2015). But large shares of the Asian and Hispanic residents that shared neighborhoods with their Black neighbors were themselves immigrants, making it difficult to disentangle whether it was racial buffering or nativity that promoted higher-SES White moves into the neighborhood. Hwang (2020) found that, since the 1990s, Black neighborhoods may have been more likely to gentrify than in previous periods, suggesting that link between race and neighborhood ascent has changed over time as the sociodemographic and immigrant composition of the U.S. has diversified in recent decades (Owens & Candipan, 2019). Thus, disentangling how race and immigration intersect in predominantly Black neighborhoods may unmask some of the complex demographic processes underlying neighborhood ascent.

The Relationship Between Race and Nativity on Neighborhood Socioeconomic Ascent

The research above raises questions about where and how neighborhood ascent occurs in U.S. metropolitan neighborhoods. First, how is socioeconomic ascent related to the racial composition and nativity of residents? While the evidence above suggests that immigration positively correlates with socioeconomic ascent, existing research does not investigate how patterns might differ by the race of immigrants and the racial composition of neighborhoods. Further, existing work leaves unclear how the nativity and race/ethnicity of residents interact during neighborhood ascent. For example, are native-born Black residents contributing to ascent more relative to Black immigrants in majority Black neighborhoods? Are Black immigrants contributing to neighborhood ascent more or less than both native and foreign-born non-Black residents? Our study addresses which groups are contributing more or less to neighborhood ascent.

Second, what are the relative contributions to neighborhood socioeconomic ascent of long-tenured immigrant residents compared to immigrant newcomers to these neighborhoods, and how do these contributions differ from their native-born counterparts? Answering this question will ascertain whether higher-SES immigrants are moving into ascendant neighborhoods, and whether their SES is higher than their U.S.-born counterparts. On the one hand, socioeconomic ascent in immigrant neighborhoods may be driven by immigrants themselves. On the other hand, immigrants may experience socioeconomic stagnation, or even decline, and their native-born neighbors may be the higher-SES group that contributed to increasing socioeconomic status in neighborhoods. And, of course, it may be a mix of both.

Third, how does the racial composition of neighborhoods and the racial compositions of immigrant and U.S.-born residents interact? Given the differences across race noted above, we might expect their relative contributions to neighborhood ascent to depend on the combination of racial composition of the neighborhood and the race of immigrants. Hwang (2015, 2016, 2020) is among the few to identify the role of immigrants as an unexplored mechanism in shaping broader residential migration patterns over time which lead to gentrification in some minority neighborhoods versus others, showing that the relationship between race and gentrification is conditional on immigrant growth, but also that these interactions manifest themselves differently in more diverse areas. However, by focusing on case studies of Chicago and Seattle, this work leaves unresolved a national analysis of these trends. In part this is because a complete analysis cannot be completed with public, aggregated data due to the precision with which we need to answer the question.

Finally, has the socioeconomic status of long-time immigrant residents increased over time in ascendant neighborhoods? Here, we focus on residents that have lived in their neighborhoods a relatively long time, and we observe whether the SES of these native- and foreign-born incumbents has increased over time as their neighborhoods experienced socioeconomic ascent.

Data Sources and Methods

Publicly available Census data make the associations between socioeconomic ascendancy, race, and nativity difficult to disentangle. Revealing cross-tabulations that include the race, socioeconomic status, and nativity of individuals within neighborhoods could lead individuals to be identified. As a result, public-use data only reveal two-way associations. The four-way interaction of race, nativity, residential tenure and socioeconomic status can only be inferred, leading to a large degree of uncertainty when estimating the independent effects and preventing interactive analyses.

We leverage confidential census microdata for census tracts in all metropolitan regions in the U.S. to answer the research questions laid out above. The first step, however, involves classifying neighborhoods as socioeconomically ascendant and then categorizing them based on their initial racial composition.

Identifying Ascendant Neighborhoods and Neighborhood Types Using Public-Use Census Data

Categorizing neighborhoods into different types involves two steps using aggregate public-use census data. First, we measure whether neighborhoods experience socioeconomic ascent from 1990 to 2010. Second, we classify neighborhoods based on their racial composition in 1990.

Socioeconomic ascent encompasses a broad range of upward mobility among neighborhoods, including gentrification as well as wealthy suburbs. Ascent can be the product of the in-movement of higher-SES residents, out-mobility of lower-SES residents, and increasing SES of long-time residents. All three mechanisms would contribute to increases in a neighborhood’s average SES. The first step involves calculating a neighborhood SES score for each tract located in a metropolitan area. We use the public long-form decennial census in 1990 and 2008–2012 American Community Survey (ACS) data for 2010 (the midpoint of the 5-year estimates). We perform factor analysis with principal components extraction to create our composite neighborhood SES score based on data from 1990 and then again with data from 2010. We compute factor scores at the neighborhood level for all tracts located in census-defined metropolitan areas based on five commonly used (and highly cor-related) indicators of neighborhood SES: median rent, median home value, median household income, percent of residents with at least a bachelor’s degree, and percent employed in managerial, professional, and technical occupations (as classified by the Standard Occupation Classification [SOC] system; Owens, 2012; Owens & Candipan, 2019). As expected, each indicator loaded highly onto a single component (pro-max rotation) in both years, 1990 and 2010, with high correlation and communality.

Having SES scores in each year then allows us to observe neighborhood SES change over a two-decade span. Using the 1990 data, we calculated the percentile ranks of all tracts within metropolitan areas (ranging from 1 to 100), and then did the same using the 2010 data. We then calculate the change in each tract’s rank from 1990 to 2010.¹ Following past work (Owens, 2012; Owens & Candipan, 2019), we categorize tracts as socioeconomically ascendant if their percentile rank increased by 10 points from 1990 to 2010. Accordingly, tracts with SES percentile ranks of 91 or greater in 1990 are ineligible to be classified as ascendant since they cannot increase by at least ten points and were excluded from our analyses. All tracts that were neither ascendant nor ineligible are categorized as non-ascendant.² Following these procedures results in two neighborhood SES categories: ascendant (n = 13,158) and non-ascendant (n = 39,320) neighborhoods.³

We hypothesize that population processes underlying neighborhood ascent play out differently in White, Black, Hispanic, and Asian neighborhoods. To that end, we further categorized neighborhoods based on their initial racial composition in 1990. We classified predominantly White neighborhoods as those in which at least 75% of residents are non-Hispanic White in 1990; predominantly Black and Hispanic neighborhoods as those with at least 50% Black and Hispanic, respectively; and disproportionately Asian neighborhoods are those in which Asians composed at least 30% of the population in 1990 (and in which Blacks or Hispanics did not comprise more than 50%). After combining the two SES trajectory classifications and four racial classifications, we arrived at eight (2 × 4) neighborhood categories.

A Synthetic Cohort Approach to Measure SES Using Restricted Census Microdata

While public-use census data can be useful for categorizing tracts and answering questions about aggregate trends in ascendant neighborhoods, answering our research questions about which groups are driving neighborhood SES requires data on individuals’ demographic and socioeconomic characteristics, as well as information on when they moved to their current residence. However, data on socioeconomic status by race/ethnicity by nativity and by year moved to current residence are not publicly available. Therefore, to pursue this analysis, we must draw on the restricted-use microdata for the decennial census and ACS available in the Census Restricted Data Center (CRDC).

We use the long-form restricted microdata to group residents into synthetic cohorts based on race, nativity, and how long they have lived in the neighborhood. First, we construct two residential groups based on tenure in the neighborhood. One group comprises long-time residents who lived in their neighborhoods before 2000, while the other consists of recent movers who entered their neighborhoods between 2000 and 2010. Note that in our analysis, we refer to these two groups as “residential cohorts.”⁴ We then further disaggregate groups by nativity (U.S.- or foreign-born) and race/ethnicity. Foreign-born (immigrant) status is defined based on responses to the item in each survey asking where the householder and/or partner was born and whether their response is a foreign country (with regard to citizenship status). Due to CRDC microdata disclosure restrictions that limit the number of racial/ethnic categories we can use in our synthetic cohort analyses, we construct a binary category for race that denotes co-ethnicity. Co-ethnicity represents individuals who belong to the majority racial/ethnic group in each neighborhood type, for example Black residents in predominantly Black neighborhoods would be categorized as the co-ethnic group for that neighborhood while all non-Black residents would be classified as non-co-ethnic. Thus, for each tract, we aggregate the individual-level data into 2 × 2 × 2 synthetic cohorts for each combination of residential tenure (long-time residents/recent movers), nativity (foreign-born/U.S.-born), and racial/ethnic group membership (co-ethnic/not co-ethnic). Note that census and ACS survey data are not longitudinal. The synthetic cohorts, therefore, will not consist of the same individuals at the beginning and end of our study period (1990 and 2010) due to Census sampling and people leaving the neighborhood—the reason why each cell makes up a “synthetic” cohort.⁵ Table 1 illustrates the residential cohort-race-immigrant groups of interest, using groups in predominantly Black neighborhoods as an example.

Table 1.

Example of synthetic cohort groups in predominantly black neighborhoods

Residential Cohort × Nativity × Race	Predominantly Black Neighborhoods
Majority race
Long-time Resident, U.S. Born, Co-ethnic	Pre-2000 move, Black, U.S. Born
Recent Mover, U.S. Born, Co-ethnic	2000 or later move, Black, U.S. Born
Long-time Resident, Foreign Born, Co-ethnic	Pre-2000 move, Black, Foreign Born
Recent Mover, Foreign Born, Co-ethnic	2000 or later move, Black, Foreign Born
Non-majority race
Long-time Resident, U.S. Born, Not co-ethnic	Pre-2000 move, non-Black, U.S. Born
Recent Mover, U.S. Born, Not co-ethnic	2000 or later move, non-Black, U.S. Born
Long-time Resident, Foreign Born, Not co-ethnic	Pre-2000 move, non-Black, Foreign Born
Recent Mover, Foreign Born, Not co-ethnic	2000 or later move, non-Black, Foreign Born
Residential cohort × nativity	Predominantly black neighborhoods
Majority race
Long-time Resident, U.S. Born	Pre-2000 move, U.S. Born
Recent Mover, Foreign Born	2000 or later move, Foreign Born
Non-majority race
Long-time Resident, U.S. Born	Pre-2000 move, U.S. Born
Recent Mover, Foreign Born	2000 or later move, Foreign Born

This table displays the grouping that we would create for an analysis of residential cohort-race/ethnicity-nativity groups (top panel) and residential cohort-nativity groups (bottom panel), using predominantly Black neighborhoods as an example.

For analyses requiring microdata, we are limited to studying neighborhood SES ascent in predominantly Black, Hispanic, and White neighborhoods. While we are also interested in demographic processes underlying predominantly Asian neighborhoods that ascended, data disclosure limitations do not allow us to present findings using restricted census microdata (i.e., all synthetic cohort analyses).⁶ For predominantly Asian neighborhoods, we present only the descriptive results that draw on the aggregated public-use census data.

Our synthetic cohort analysis involves comparing the mean socioeconomic characteristics of households based on when they moved to the neighborhood, by race and by immigrant status. This approach allows us to examine interactions between residential tenure, household race, immigration, and initial neighborhood context prior to ascent. Our synthetic cohort approach also allows us to analyze whether the mean socioeconomic characteristics of long-time residents increased over time that would provide evidence that improvements to long-time residents’ SES is at least partly driving neighborhood ascent, results that we can further examine by race/ethnicity and nativity.

Calculating Group-Level SES Within Tracts

To calculate the group-level SES in each tract for each of the synthetic cohorts that we identified using census microdata, we follow a process analogous to the method that we used to classify ascendant neighborhoods that drew on aggregate census data except that we use synthetic cohort-tracts as our unit of analysis. That is, for our full model, which analyzes tenure-race/ethnicity-nativity groups (described below), we collapse the millions of observations from raw household-level data into the residential tenure-race/ethnicity-nativity cohorts within all metropolitan tracts (including tracts ineligible to ascend and non-ascendant tracts). The resulting data are arrayed with each row representing a unique tenure- nativity-race cohort for a given tract.

With the data described above we perform a single factor analysis with principal components extraction on the same five SES measures that we used to compute the aggregate neighborhood SES scores: percent of residents ages 25 and older with a college degree; percent of workforce-eligible residents in professional, managerial, or technical jobs classified by the SOC system; median household income; and property value or household rent. This procedure, which again uses all available data from all metropolitan area, produces an SES score for each tenure-race/ethnicity-nativity group in each tract in 2010. For example, our procedure would generate 2010 SES scores in each predominantly Black tract (a) for Black immigrants who moved to the tract prior to 2000, (b) for Black immigrant newcomers that moved into the tract from 2000 to 2010, (c) for non-Black immigrants who moved to the tract prior to 2000, and (d) for non-Black immigrant newcomers that moved into the tract from 2000 to 2010. For the early residential cohort (the long-time residents), we also computed an SES score in 2000. (To be clear, note that we will also perform an analysis at the residential tenure-nativity level. For these analyses, we perform a single factor analysis on a dataset arrayed at the tract-residential cohort-nativity level.)

We then use these group-level SES scores as our dependent variables when we perform our regression analyses (described below) investigating the role of race and immigration in ascendant neighborhoods categorized by racial composition. These analyses, for example, reveal (1) whether immigrants are the high-SES residents of predominantly White, Black or Hispanic tracts experiencing ascent; (2) whether these are longer-tenured or newcomer residents (3) and whether, if immigrants are not the high-SES residents, native-born households of the same race as the neighborhood are the higher-SES residents. Comparing ascent processes by neighborhood racial/ethnic and foreign-born composition reveals whether unique demographic processes occur in different types of socioeconomically ascending neighborhoods.

Analytic Strategy: Demographic Synthetic Cohort Approach Using Restricted-Use Census Microdata

As previously described, our analyses require a modeling strategy that allows us to observe which demographic processes by race/ethnicity and immigration accompany neighborhood ascent. Following past work (McKinnish et al., 2010; Owens & Candipan, 2019), we take a demographic synthetic cohort analysis approach to answer our questions, classifying subgroups by residential tenure, race, and nativity. Since our research questions focus on which groups are contributing socioeconomically to ascent, we restrict our analyses to ascendant tracts only. We perform separate models for predominantly White, Black and Hispanic neighborhoods. Recall, however, that before we apply these filters, we use data from all tracts to compute the synthetic group SES scores in each neighborhood.

We use OLS regression to predict the SES in 2010 for U.S.- and foreign-born residents that moved into the neighborhood before or after 2000. The first set of analyses requires that examine our data analyzed at the tract-residential cohort-nativity (TCN) level where groups are defined based only on tenure and nativity. Our full model is expressed as:

$$\text{2010 SES Score }=\alpha +{\beta }_1\mathit{\text{Tenure }}+{\beta }_2\mathit{\text{Foreign }}+{\beta }_3(\mathit{\text{Tenure}}\times \mathit{\text{Foreign}})+\gamma \mathit{\text{MetroType}}+\epsilon$$ (1)

Tenure denotes our measure for the two residential cohorts, which is an indicator for those that moved into the neighborhood between 2000 and 2010 (newcomers) compared to those already in the neighborhood (long-time residents) that act as the reference; the corresponding β represents the difference in SES position associated with long-time residents compared to newcomers. Foreign denotes whether a resident is foreign-born (reference category is U.S.-born) and the corresponding β represents the SES position of foreign-born residents relative to U.S.-born residents. We also include an interaction between residential cohort and nativity status (Tenure × Foreign) and we interpret the corresponding β as the increase in the effect of newcomers due to being foreign-born. We stratify the models to document the relative contributions of recent immigrant movers vs. long-time immigrant residents to SES ascent in neighborhoods with different initial racial compositions (i.e., predominantly Black, Hispanic, or White). Note that these models are cross-sectional (they capture group-level SES in 2010), though we are able to incorporate a quasi-longitudinal frame by distinguishing between incumbent residents and recent in-movers.

Since SES varies across regions with different immigration histories, we also control for three types of metros (MetroType), drawing on definitions from prior work (Brazil, 2019; Singer, 2015). To do this, we classify neighborhoods into three categories. Established Immigrant Gateways include all metropolitan statistical areas (MSAs) where the proportion of the foreign-born population in 1980 exceeded the average proportion of foreign-born residents across all MSAs in that year. New Destinations are MSAs with a growth rate in the immigrant resident population from 1990 to 2000 that is higher than the growth rate at the national level for that same period. MSAs with both above-average shares and high growth of the immigrant population are classified as established gateways. Finally, Minor destinations consist of all remaining MSAs. We are able to classify 355 MSAs into immigrant metro types after accounting for the 25 MSAs with missing data (2003 OMB definitions).

Next, we predict SES in 2010 by residential tenure, race, and nativity. To do this, we array our data at the tract-residential cohort-race-nativity (TCRN) level where the groups were defined using tenure, nativity, and race/ethnicity. Our full model is formally expressed as:

$$\text{2010SESScore}=\alpha +{\beta }_1\mathit{\text{Tenure}}+{\beta }_2\mathit{\text{Foreign}}+{\beta }_3(\mathit{\text{Tenure}}\times \mathit{\text{Foreign}})+{\beta }_4\mathit{\text{MajRace}}+{\beta }_5(\mathit{\text{Tenure}}\times \mathit{\text{MajRace}})+{\beta }_6(\mathit{\text{Foreign}}\times \mathit{\text{MajRace}})+{\beta }_7(\mathit{\text{Tenure}}\times \mathit{\text{Foreign}}\times \mathit{\text{MajRace}})+\gamma \mathit{\text{MetroType}}+ϵ$$ (2)

This model includes the terms for residential cohort and nativity and adds measures for race. MajRace is a binary measure indicating whether the racial/ethnic background of an individual is the same as the majority racial/ethnic group of the neighborhood. For example, a Hispanic individual in a predominantly Hispanic neighborhood would be considered part of the majority racial/ethnic group of that neighborhood. The first and last lines of Eq. 2 are identical to Eq. 1. The second and third lines add coefficients for MajRace, the two two-way interactions between MajRace and Tenure and Foreign, and the three-way interaction between all three variables. The model includes all interactions between our three main predictors for residential tenure, majority race, and foreign born and, through combinations of coefficients, produce estimates for all eight categories listed in Table 1. Our models allow us to examine the interaction between race, immigration, and residential tenure in population processes of neighborhood ascent, as well as identify which groups contribute most to increasing SES in predominantly Black, Hispanic, and White ascendant neighborhoods. Based on the parameter estimates, we can answer our research questions: (1) how does the SES of U.S.-born residents compare to foreign-born residents, and does this vary in different types of ascendant neighborhoods? and (2) among immigrants, does the relative SES contributions of current residents versus recent in-movers on neighborhood ascent differ by race and neighborhood racial composition?

Finally, we examine whether the SES of long-time immigrant stayers increased over time in ascendant neighborhoods with different initial racial compositions, adding a longitudinal lens. Since our focus for this model is on the trajectories of long-tenured residents, this analysis further restricts the sample to the early residential cohort stayers in ascendant neighborhoods of each type. Having already accounted for residential tenure through our sample filters, these models thus rely on data that is unique at the tract-race-nativity-year (TRNY) level.

$$\text{SES Score}=\alpha +{\beta }_1\mathit{\text{Foreign }}+{\beta }_2\mathit{\text{MajRace}}+{\beta }_3(\mathit{\text{Foreign}}\times \mathit{\text{MajRace}})+{\beta }_4\mathit{\text{Year}}+{\beta }_5(\mathit{\text{Foreign}}\times \mathit{\text{Year}})+{\beta }_6(\mathit{\text{MajRace}}\times \mathit{\text{Year}})+{\beta }_7(\mathit{\text{Foreign}}\times \mathit{\text{MajRace}}\times \mathit{\text{Year}})+\gamma \mathit{\text{MetroType}}+\epsilon$$ (3)

As with previous models, these analyses include terms for race, foreign-born status, and a term for their interaction, again controlling for immigrant metro type. We estimate the model separately for predominantly Black, Hispanic, and White ascendant neighborhoods. We include a measure for Year, denoting the year 2010 (reference year is 2000), which we interact with all other terms. Recall that these models do not track the same individuals over time, but rather capture the mean SES of each of the synthetic groups. These models allow us to answer questions about whether, and to what extent, the average SES of long-time residents increased over time in different types of ascendant neighborhoods. For example, in initially predominantly Black ascendant neighborhoods, did the SES of Black immigrant incumbents increase more from 2000 to 2010 relative to the SES of non-Black native incumbents?

Descriptive Results

Do Changes in the Racial Composition of Residents Occur Alongside Neighborhood Ascent?

Table 2 reports demographic change over time (1990 to 2010) in ascendant and non-ascendant neighborhoods by racial compositions in 1990 and immigrant metro type. The shares of White residents declined on average in predominantly White neighborhoods, but the share of White residents declined far less in ascendant neighborhoods relative to non-ascendant neighborhoods. The differences between non-ascendant and ascendant neighborhoods were highest in established gateways and new destinations. We observed the opposite trend for whites in all non-predominantly White neighborhoods—the share of whites more than doubled in ascendant predominantly Black, Hispanic, and Asian neighborhoods (representing an increase of 5.8, 9.0, and 3.0 percentage points, respectively). Apart from predominantly Hispanic ascendant neighborhoods located in minor destinations, this pattern holds true when also disaggregating by immigrant metro type. Note that the increasing shares of White residents in ascendant minority neighborhoods occurred during a period in which the overall proportion of whites declined in the U.S.

Table 2.

Mean change in racial/ethnic and foreign-born composition, 1990 to 2010, by neighborhood ascent and racial/ethnic category in 1990 and immigrant metro type

			% White		% Black		% Hispanic		% Asian		% Foreign-born		N
			1990	Δ 90–10	1990	Δ 90–10	1990	Δ 90–10	1990	Δ 90–10	1990	Δ 90–10
	Metro type	Neighborhood type
Non-ascent	Established Immigrant Gateways	>75% White	81.1	− 21.7	3.6	3.9	9.6	11.4	5.3	5.8	14.5	8.7	7254
		>50% Black	7.1	− 3.0	78.6	− 6.5	12.3	8.4	1.4	0.8	17.1	7.2	1032
		>50% Hispanic	15.4	− 8.7	7.4	− 1.6	72.8	9.8	3.9	0.5	40.7	3.2	1586
		>30% Asian	21.4	− 8.7	2.2	0.3	8.1	0.3	67.9	7.6	27.6	3.6	167
		No majority	34.4	− 18.5	16.1	− 0.7	31.3	12.6	17.4	6.3	33.6	8.5	1319
	New destinations	>75% White	81.4	− 25.9	6.3	6.2	8.7	15.6	2.9	3.6	7.1	10.2	5952
		>50% Black	14.1	− 6.9	79.2	− 4.8	5.3	11.2	1.1	0.3	5.0	7.3	593
		>50% Hispanic	23.6	− 13.6	7.1	− 1.8	66.9	15.4	1.1	0.0	26.4	9.9	253
		>30% Asian											0
		No majority	39.4	− 23.9	24.1	− 3.1	28.0	26.1	6.8	0.9	18.2	14.9	409
	Minor	>75% White	88.8	− 14.8	6.3	6.5	3.1	5.9	1.3	1.7	3.4	3.7	17,614
		>50% Black	16.4	− 5.9	80.8	2.2	1.9	3.1	0.6	0.3	1.6	2.0	2076
		>50% Hispanic	21.9	− 9.3	5.5	0.1	71.0	9.0	0.6	0.1	10.0	6.9	288
		>30% Asian											2
		No majority	41.7	− 17.8	32.0	4.7	17.2	11.9	3.9	0.4	7.9	5.5	340
	Overall	>75% White	85.5	− 18.6	5.6	5.8	5.7	9.0	2.5	3.0	6.7	6.1	30,820
		>50% Black	13.5	− 5.2	79.9	− 1.3	5.3	5.9	0.9	0.4	6.5	4.3	3701
		>50% Hispanic	17.2	− 9.4	7.1	− 1.4	71.9	10.4	3.1	0.4	34.8	4.5	2127
		>30% Asian	21.5	− 8.6	2.3	0.5	8.0	0.4	67.7	7.2	27.8	3.5	169
		No majority	36.6	− 0.2	20.3	− 0.3	28.4	15.1	13.1	4.3	26.3	0.1	2068
Ascent	Established Immigrant Gateways	>75% White	80.4	− 15.0	3.7	2.4	11.3	6.0	4.1	6.1	15.0	5.7	1628
		>50% Black	6.2	5.5	81.1	− 15.0	10.8	6.7	1.5	2.5	14.8	6.7	312
		>50% Hispanic	17.1	1.4	5.1	− 0.2	72.8	− 3.6	4.5	2.3	38.9	− 2.3	559
		>30% Asian	16.1	1.9	1.7	1.5	9.5	− 1.1	72.2	− 3.0	33.9	− 0.5	38
		No majority	34.9	− 2.5	17.6	− 4.1	32.3	− 0.1	14.1	6.4	30.8	2.7	326
	New Destinations	>75% White	83.2	− 14.0	5.8	2.7	8.8	6.2	1.5	4.8	5.6	5.8	2356
		>50% Black	12.7	13.2	80.9	− 22.5	5.3	6.6	0.8	2.3	5.4	5.5	173
		>50% Hispanic	23.9	0.6	7.1	− 0.6	66.3	− 1.5	1.5	1.3	26.5	1.3	112
		>30% Asian											4
		No majority	39.9	− 6.2	20.0	− 2.3	30.6	6.6	6.5	2.4	18.4	4.9	155
	Minor	>75% White	90.3	− 7.3	5.5	1.9	2.7	3.1	0.8	1.7	2.5	2.3	6550
		>50% Black	19.3	5.3	77.8	− 9.2	2.0	2.5	0.6	1.0	1.6	2.5	559
		>50% Hispanic	27.9	− 0.1	4.1	− 0.2	66.0	− 0.7	0.5	0.7	10.1	3.1	83
		>30% Asian											1
		No majority	40.0	0.9	29.3	− 3.8	18.0	4.7	2.3	1.1	5.6	4.0	97
	Overall	>75% White	87.2	− 10.0	5.2	2.1	5.4	4.3	1.5	3.1	5.1	3.6	10,534
		>50% Black	14.3	6.7	79.3	− 13.2	5.1	4.4	0.9	1.7	6.2	4.3	1044
		>50% Hispanic	19.3	1.1	5.3	− 0.3	71.1	− 2.9	3.6	2.0	33.9	− 1.1	754
		>30% Asian	16.4	3.1	3.3	0.8	8.9	− 0.6	70.8	− 4.0	35.5	− 1.0	43
		No majority	37.1	− 2.9	20.2	− 3.5	29.4	2.5	10.1	4.4	23.3	3.5	578

Racial composition is at the tract level. Black, Asian, and White exclude Hispanic persons. Other race denotes census-defined categories others than Black, Asian, White or Hispanic. Neighborhoods’ 1990 racial/ethnic category is indexed in rows. Neighborhood racial classifications are mutually exclusive, so > 30% Asian neighborhoods are not also > 50% Black or Hispanic. The greyed out the cells for predominantly Asian neighborhoods indicate insufficient data (n < 5). No majority neighborhoods denote those that are not classified as predominantly White, Black, Asian or Hispanic. Immigrant metro types are also mutually exclusive categories and classified based on level and growth in the foreign-born resident population, drawing on definitions from past work (Brazil, 2019)

We largely observe the inverse trends for Hispanic, Black, and Asian residents. The shares of Hispanic and Black residents decrease slightly in ascendant neighborhoods and increased in non-ascendant neighborhoods. Note that these decreases occur during a period of substantial overall growth in the U.S. Hispanic population. On the other hand, the percentage of Hispanic residents increases the most in non-ascendant Hispanic neighborhoods in new destinations (from 66.9% in 1990 to 82.3% in 2010). While the percentage of Black residents decreases in ascendant Black neighborhoods of all immigrant metro types, the disparity between ascendant and non-ascendant is most pronounced in new destinations: from 1990 to 2010, the share of Black residents decreased by 22.5% in ascendant neighborhoods and decreases by less than 5% in non-ascendant neighborhoods. Nearly all predominantly Asian neighborhoods are located in established immigrant gateways where we observe a similar trend as that found in other neighborhood types, albeit to a somewhat lesser degree—the share of Asian residents increases by about 7.6% in non-ascendant neighborhoods but decreases by 3% in ascendant neighborhoods.

Does the Share of Foreign-Born Residents Correlate with Neighborhood Ascent?

The rightmost two columns of Table 2 show initial immigrant composition (in 1990) and growth in non-ascendant and ascendant neighborhoods. Among predominantly White neighborhoods, the overall proportion of foreign-born residents increased over time in both non-ascendant and ascendant neighborhoods, but it increased more in non-ascendant White neighborhoods (6.1 percentage points) rather than ascendant White neighborhoods (3.6 percentage points). In predominantly Black neighborhoods, the foreign-born presence increased during this period in both ascendant and non-ascendant neighborhoods. While non-ascendant Asian and Hispanic neighborhoods experienced growth in the proportion of foreign-born residents (increasing 4.5 and 3.5 percentage points, respectively), the foreign-born presence decreased overall in ascendant Hispanic and Asian neighborhoods. (Note that the overall percentage of foreign-born residents is much lower in predominantly White and Black neighborhoods compared to Asian and Hispanic neighborhoods.)

Synthetic Cohort Analysis Results

Examining the Relative SES Contributions of U.S.- and Foreign-Born Incumbents and Newcomers to Ascendant Neighborhoods

The results in Table 2, which relied on public aggregated census data, showed that changes to both racial/ethnic and foreign-born compositions varied by the initial racial composition of neighborhoods. Table 3 reports the resulting parameter estimates of the synthetic cohort analysis of SES among U.S.- and foreign-born residents that moved into ascending neighborhoods before and after 2000 (Eq. 1). All estimates draw on restricted microdata. Results for each neighborhood racial/ethnic type are displayed in Table 3.

Table 3.

OLS regression models predicting socioeconomic status from cohort and foreign-born status by neighborhood racial composition (in 1990)

	Predominantly Black (> 50% Black)	Predominantly Hispanic (> %50 Hispanic)	Predominantly White (> 75% White)
	Model 1	Model 2	Model 3
U.S. bom (ref)
Foreign born	0.192***	− 0.219***	0.222***
	0.032	0.042	0.026
Cohort 1 (Incumbents): pre-2000 (ref)
Cohort 2: (Newcomers): 2000 or later	0.270***	− 1.180***	0.353***
	0.057	0.047	0.012
Foreign born × Cohort 1
Foreign born × Cohort 2	0.127 +	− 0.05	− 0.080**
	0.07	0.061	0.030
Immigrant metro types
Established Gateway Metro (ref)
New destination metro	− 0.796***	− 0.341***	− 0.842***
	0.038	0.075	0.014
Other destination metro	− 0.783***	− 0.633***	− 0.881***
	0.032	0.057	0.013
Constant	− 0.04	− 0.100**	0.428***
SE	0.029	0.032	0.013
N TCN	2720	2380	20,310

Data unique at the tract-residential cohort-nativity (TCN) level. All models are restricted to ascendant tracts. Cells present regression coefficients with standard errors beneath in italics. The racial categories for Black and White indicate non-Hispanic. All base models are available in online Appendix Table A2.

^*p < 0.05;

^**p < 0.01;

^***p < 0.001 (two-tailed tests)

Figure 1 plots each group’s SES (in 2010) in predominantly White, Black, Hispanic ascendant neighborhoods, predicted from parameter estimates in Table 3. Predicted values show the relative SES contributions for each group—by nativity and residential tenure—to neighborhood ascent within each type of neighborhood.

Fig. 1.

Predicted SES in 2010 by Cohort and Immigration in Predominantly Black, Hispanic, and White Ascendant Neighborhoods. All models are restricted to ascendant neighborhoods only. Analyses are performed separately by neighborhood type, designated by the initial racial/ethnic composition in 1990. Samples are unique at the tract-residential cohort-nativity (TCN) level. Figures display predicted SES in 2010 for each of the four cohort-by-nativity synthetic groups. Cohort 1 refers to those that moved into their current neighborhood prior to 2000 (“stayers”). Cohort 2 denotes those that moved into their current neighborhood in 2000 or later (“movers”). Note that these models do not disaggregate by race/ethnicity. Thus, average group SES encompasses both the SES of the majority racial/ethnic group and all others

In both predominantly White and predominantly Black neighborhoods, foreign-born residents tend to have higher SES than U.S.-born residents with foreign-born newcomers contributing the most to neighborhood ascent.⁷ In predominantly White neighborhoods, the SES of immigrant newcomers is 0.92 unit in 2010, compared to a contribution of 0.65 units via the SES of incumbent immigrant residents. U.S.-born residents who recently moved into predominantly White neighborhoods also have high SES in 2010 (0.78 units), which is nearly double the SES score of incumbent U.S.-born residents living in predominantly White neighborhoods (0.43 units). We find similar patterns with respect to immigrants and incumbency in predominantly Black neighborhoods, where the SES of immigrant recent movers is substantially higher in 2010 compared to incumbent foreign-born residents (0.55 compared to 0.15). In Black ascendant neighborhoods, U.S.-born newcomers’ SES is just under the SES score of immigrant newcomers, while incumbent U.S. residents contribute little to neighborhood ascent.

In predominantly Hispanic ascendant neighborhoods, the story appears to be one of inter-cohort differentiation rather than one between U.S.- and foreign-born residents, and to a lesser extent, intra-cohort similarity. The results for Hispanic ascendant neighborhoods in Fig. 1 may appear slightly odd given that group SES scores are negative. The negative scores arise because even socioeconomically ascendant predominantly Hispanic neighborhoods tended to lose ground to overall economic changes that occurred in all neighborhoods from 1990 to 2010, in part because of the areas of the country where ascendant predominantly Hispanic neighborhoods were located.⁸ Nonetheless, our research objective is to analyze the relative SES contributions of each group within ascendant Hispanic neighborhoods. To that end, we observe the opposite trend as seen in predominantly White and Black neighborhoods—incumbent residents have substantially higher SES than recent in-movers. When examining trends within each residential cohort, we do observe some differences. Among newcomers, SES in 2010 is slightly higher for U.S.-born residents relative to their foreign-born counterparts. On the other hand, SES in 2010 is higher for incumbent immigrant residents compared to U.S.-born incumbents. These intra-cohort differences, however, are substantially smaller relative to the inter-cohort variation that we observe.

Do the Relative Contributions of Foreign- and U.S.-Born Residents Differ by Race of Immigrants and Neighborhood Racial Composition?

Overall, the results above showed that immigrant and U.S.-born contributions to SES ascent varies by neighborhood racial composition. The next set of analyses focuses on the interaction between race and nativity of long-time residents and newcomers in ascendant neighborhoods. Recall that these are cross-sectional regression models predicting tract-level SES in 2010 for each combination of residential cohort, race, and nativity group (TCRN) in ascendant initially predominantly Black, Hispanic, and White neighborhoods. As with the previous analysis, these models show which groups have the highest SES relative to others in ascendant neighborhoods, thus demonstrating group hierarchy in terms of SES contributions to neighborhood ascent.

Table 4 presents coefficients from these regression models. As with the previous analysis, all models control for the three categories of immigrant metro type: established gateways, new destinations, and minor destinations. Across models of all racial/ethnic ascendant neighborhood types, residents in established destinations have substantially and similarly higher SES than new destination and other minor destination metros.

Table 4.

OLS regression models predicting socioeconomic status from cohort, race, and foreign-born status by initial neighborhood racial composition (in 1990)

	Predominantly Black (> 50% Black)	Predominantly Hispanic (> 50% Hispanic)	Predominantly White (> 75% White)
	Model 4a	Model 4b	Model 4c
Description
U.S. born (ref)
Foreign born	− 0.232**	0.023	0.151**
	0.08	0.075	0.050
Cohort 1: pre-2000 (ref)
Cohort 2: 2000 or later	0.197***	0.243***	0.365***
	0.046	0.046	0.025
Non-majority race (ref)
Majority race	− 0.466***	− 0.278***	0.088***
	0.046	0.052	0.024
Cohort 2 × majority race	− 0.167**	− 0.129 +	− 0.015
	0.057	0.067	0.028
Foreign born × majority race	0.708***	− 0.125	0.106 +
	0.108	0.091	0.061
Foreign born × cohort 2	0.345***	0.284**	− 0.123*
	0.09	0.09	0.054
Cohort 2 × Foreign Born × majority race	− 0.443***	− 0.399***	0.150*
	0.129	0.112	0.066
Immigrant metro types
Established gateway (ref)
New destination	− 0.840***	− 0.362***	− 0.822***
	0.032	0.06	0.012
Other destination	− 0.813***	− 0.578***	− 0.861***
	0.027	0.045	0.011
Constant	0.357***	0.03	0.329***
SE	0.042	0.036	0.024
N TCRN	3830	3280	26,810

Data are unique at the tract-residential cohort-race-nativity (TCRN) level. FB refers to foreign born. Race refers to majority race. All models restricted to ascendant neighborhoods only. Cohort is a binary measure denoting residential cohort. Majority Race denotes whether the racial/ethnic background of the group matches the majority or predominant racial/ethnic group of each neighborhood type (e.g., Blacks in predominantly Black neighborhoods are part of the majority race group). Immigrant metro types draw on definitions from past work which categorizes regions based on the growth in their foreign-born population (Brazil, 2019). All base models are available in online Appendix Table A3

^*p < 0.05;

^**p < 0.01;

^***p < 0.001 (two-tailed tests)

To better interpret results from Table 4, we present a series of figures, one for each racial group, that demonstrate how demographic processes vary by race/ethnicity and immigration for early and recent movers to different types of ascendant neighborhoods.⁹ Across all figures, the left four bars report the predicted 2010 SES values for the predominant racial/ethnic group (split by nativity and residential cohort) while the right four bars report the predicted values for the non-predominant racial/ethnic groups.

Figure 2 shows predicted values in predominantly Black ascendant neighborhoods. Overall, Black residents have significantly lower SES than non-Black residents. However, when we disaggregate by race and nativity, we see that Black immigrants contribute to neighborhood SES ascent. The SES of foreign-born Black residents—both long-term residents and newcomers—is higher than the SES of U.S.-born Blacks. The SES of U.S.-born Black long-term residents is the lowest among all groups followed closely by U.S.-born Black newcomers. The substantial difference between foreign- and U.S.-born Black residents underscores how overlooking processes of immigration when examining the racial demography of neighborhood ascent masks meaningful heterogeneity among Black residents.

Fig. 2.

Predicted 2010 SES in Ascendant Initially Predominantly Black Neighborhoods. All models are restricted to ascendant predominantly Black neighborhoods, as designated by initial neighborhood racial/ethnic composition in 1990. Samples are unique at the tract-residential cohort-immigration (TCRN) level. Figures display predicted SES in 2010 for each of the four cohort-by-majority race-by-nativity synthetic groups. Cohort 1 refers to those that moved into their current neighborhood prior to 2000 (“stayers”). Cohort 2 denotes those that moved into their current neighborhood in 2000 or later (“movers”). The left four bars represent stayers; the right four bars represent movers

Among non-Black residents in ascendant Black neighborhoods, the story is more about cohort differentiation between newcomers and longer-tenured residents: recent movers have substantially higher SES than long-time stayers that settled into the neighborhood before 2000. Non-Black immigrant newcomers have the highest SES scores among all groups in 2010 (0.7 units). In contrast, the SES of non-Black immigrant incumbents is among the lowest (roughly a third of the SES score of their Black immigrant incumbent counterparts in 2010).

Figure 3 reports the predicted values using parameter estimates for the predominantly Hispanic ascendant sample. Overall, recent movers have statistically significantly higher predicted SES in 2010 than long-time residents. Non-Hispanic newcomers have the highest SES in 2010—almost 0.6 units among foreign-born non-Hispanics and almost 0.3 units among U.S.-born non-Hispanics. Non-Hispanic incumbents’ SES is substantially lower relative to non-Hispanic newcomers, but also substantially higher relative to Hispanic incumbents. In contrast, Hispanic incumbents and newcomers alike (including both foreign- and U.S.-born Hispanics) tend to have lower SES than non-Hispanics in ascendant Hispanic neighborhoods. These findings do not provide evidence that neighborhood socioeconomic ascent in Hispanic neighborhoods is driven by incumbent upgrading among the existing population. Rather, the results align with the theories of correlation between race and gentrification with newcomers (e.g., higher-SES gentrifiers) changing both the socioeconomic and racial/ethnic composition of Hispanic ascendant neighborhoods.

Fig. 3.

Predicted 2010 SES in Ascendant Initially Predominantly Hispanic Neighborhoods. All models are restricted to ascendant predominantly Hispanic neighborhoods, as designated by initial neighborhood racial/ethnic composition in 1990. Samples are unique at the tract-residential cohort-nativity (TCRN) level. Figures display predicted SES in 2010 for each of the four cohort-by-majority race-by-nativity synthetic groups. Cohort 1 refers to those that moved into their current neighborhood prior to 2000 (“stayers”). Cohort 2 denotes those that moved into their current neighborhood in 2000 or later (“movers”). The left four bars represent stayers; the right four bars represent movers.

Finally, Fig. 4 shows group differences in 2010 SES in ascendant predominantly White neighborhoods. Predominantly White neighborhoods experienced the greatest neighborhood ascent, and all groups seemingly took part in that ascent. Newcomers have higher SES than incumbent residents, with White immigrant newcomers having the highest SES. The SES levels of U.S.-born newcomers are similar to that of nonwhite newcomers, both foreign- and U.S.-born. Among incumbents, the SES of White immigrants is the highest in 2010 (just under 0.7 units), followed by nonwhite immigrants (0.5 units) that, in turn, have slightly higher SES than U.S.-born White incumbents (a little over 0.4 units). The SES of U.S.-born nonwhite incumbents is the lowest among all groups in 2010 (just over 0.3 units). These results regarding foreign-born residents are somewhat surprising and could indicate a case of a highly selective sorting process among a higher-SES foreign-born population that is able to enter ascendant predominantly White neighborhoods. Due to data disclosure limitations, we are unable to show further analyses examining compositional change that would provide more detail about who comprises the nonwhite group.

Fig. 4.

Predicted 2010 SES in Ascendant Initially Predominantly White Neighborhoods. All models are restricted to ascendant predominantly White neighborhoods, as designated by initial neighborhood racial/ ethnic composition in 1990. Samples are unique at the tract-residential cohort-nativity (TCRN) level. Figures display predicted SES in 2010 for each of the four cohort-by-majority race-by-nativity synthetic groups. Cohort 1 refers to those that moved into their current neighborhood prior to 2000 (“stayers”). Cohort 2 denotes those that moved into their current neighborhood in 2000 or later (“movers”). The left four bars represent stayers; the right four bars represent movers

Has the Socioeconomic Status of Long-Time Immigrant Residents Increased Over Time in Ascendant Neighborhoods?

Our final analysis examines change in SES among incumbent groups. Table 5 shows whether the mean SES of U.S.- and foreign-born stayers (i.e., pre-2000 in-movers) in ascendant neighborhoods increased over time, from 2000 to 2010, using data from both years (Eq. 3). Increases in SES for the immigrant subgroups of stayers would provide evidence that these longer-tenured foreign-born residents are at least partly driving neighborhood socioeconomic ascent. Figure 5 displays results from regression models predicting SES among only incumbents (Cohort 1), drawing on Table 5. As with all previous regression models, analyses are restricted to ascendant neighborhoods only. The top panel (5a) provides the baseline SES in 2000 for context. We emphasize the bottom panel (5b), which shows the change in SES from 2000 to 2010, the substantive focus of this analysis.¹⁰ Positive values indicate increases in long-term residents’ SES over time for a given subgroup.

Table 5.

OLS regression models predicting early cohorts’ SES in ascendant neighborhoods from race, foreign born status, and year (by neighborhood racial composition (in 1990))

Description	Predominantly Black (> 50% Black)	Predominantly Hispanic (> 50% Hispanic)	Predominantly White (> 75% White)
	Model 1	Model 2	Model 3
Year (2010)	0.208**	0.260***	− 0.033
	0.065	0.065	0.064
Majority Race	− 0.449***	− 0.351***	0.020
	0.024	0.027	0.015
Foreign Born	0.168***	0.194***	0.289***
	0.045	0.051	0.030
Year (2010) × Foreign Born	− 0.048	− 0.040	0.386*
	0.174	0.171	0.164
Year (2010) × Majority Race	0.095	− 0.042	− 0.034
	0.072	0.079	0.065
Foreign born × Majority Race	0.502***	− 0.191***	0.226***
	0.070	0.056	0.041
Year (2010) × Majority Race × Foreign Born	0.242	0.089	− 0.228
	0.228	0.186	0.205
Constant	− 0.432***	− 0.307***	− 0.272***
	0.021	0.022	0.015
N tract-race-nativity-year	4230	3870	34,350

Data unique at the tract-majority race-nativity-year (TRNY) level. The racial/ethnic categories Black and White indicate non-Hispanic. Cells present regression coefficients with standard errors beneath in italics. All models restricted to pre-2000 cohorts in ascendant tracts. SES is measured from 2000 to 2010. Coefficients for year indicate growth in longer-term residents’ SES from 2000 to 2010. All base models are available in online Appendix Table A4

^*p < 0.05;

^**p < 0.01;

^***p < 0.001 (two-tailed tests)

Fig. 5.

Initial SES (a) and SES change over time (b) among long-time residents in ascendant neighborhoods, 2000–2010. a Initial predicted SES in 2000 Among Stayers. b Change in predicted SES over time (2000–2010) among stayers. Analyses restricted to early cohort (“stayers”) in ascendant neighborhoods. Models performed separately for ascendant initially predominantly Black (a), Hispanic (b), and White neighborhoods (c). All models control for immigrant metro type

In predominantly Black ascendant neighborhoods, Black immigrants from early cohorts experience large increases in SES from 2000 to 2010, suggesting that this particular group is contributing to the increasing socioeconomic fortunes of their neighborhoods (top left of Fig. 5b). Contrast this to U.S.-born Blacks, the group with the lowest SES scores in 2000, who experience a more modest increase in SES over this same period. This result perhaps highlights this group’s vulnerability to displacement in the wake of more progressive neighborhood SES ascent in predominantly Black neighborhoods.

In predominantly Hispanic ascendant neighborhoods, the average SES increases similarly for all groups. The SES of long-time U.S.- and foreign-born Hispanics increases from 2000 to 2010, as does the SES of all non-Hispanic incumbents living in those neighborhoods (top right of Fig. 5b). Both groups of incumbent Hispanic stayers, U.S.- and foreign-born, experience similarly large increases in average SES, though their resultant SES remains substantially lower than the SES for non-Hispanics. The average SES score for non-Hispanics immigrants roughly doubles over time (from − 0.113 to 0.107) and is the highest among the four subgroups in 2010.

In predominantly White ascendant neighborhoods, the SES of immigrant stayers increases over time, suggesting that immigrants residing in these neighborhoods longer contributed partly to socioeconomic ascent (bottom left of Fig. 5b). Interestingly, results here indicate that nonwhite immigrants are driving more of the increases in neighborhood SES, on average, though the differences between White and nonwhite immigrants’ SES is small. Conversely, the SES of longtime native residents, both White and nonwhite, decreases in ascendant predominantly White neighborhoods.

Discussion and Conclusion

Using a synthetic cohort approach on public and restricted census data from 1990 to 2010, this study provides a systematic demographic analysis of race, immigration, and the intersection of both, as drivers of neighborhood socioeconomic ascent, a process of increasing focus in recent empirical work (Hwang, 2015, 2016; Owens & Candipan, 2019; Tesfai et al., 2020; Timberlake & Johns-Wolfe, 2017). Compared to past quantitative research that focuses only on the racial composition of neighborhoods associated with socioeconomic increases in neighborhoods, our analyses reveal key interactions by race and immigration, including how interactions differ in predominantly Black, Hispanic, and White ascendant neighborhoods. Failing to account for both race and nativity masks heterogeneity that explains the complex dynamics of race and immigration that shape neighborhood ascent. Furthermore, by measuring the relative SES contributions of nonwhite residents (by nativity) to neighborhood ascent, we provide new insights in a research area typically dominated by explanations that focus on whites.

For the most part, the population dynamics underlying neighborhood ascent appear to reflect a cohort effect of higher-SES newcomers entering neighborhoods rather than increasing SES among incumbent residents. In predominantly White and predominantly Black ascendant neighborhoods, immigrant newcomers tend to have higher SES than U.S.-born newcomers. Although our data do not allow us to observe the SES composition of the neighborhoods from which immigrants moved, this result perhaps reflects a pattern of spatial assimilation playing out as these households move to neighborhoods on the rise economically. In Black ascendant neighborhoods, disaggregating trends by race and nativity revealed important differences between Black immigrants and U.S.-born Black residents. The average SES of Black immigrants, particularly foreign-born Black newcomers, was substantially higher than that of native Black residents, who had the lowest SES among all groups. Paired with Hwang’s (2020) finding that Black neighborhoods may have a higher propensity to gentrify since the 1990s, our results using national-level data suggest that Black (and non-Black) immigrant in-movers are drivers of that neighborhood change in ways that differed from previous periods (Freeman, 2002).

In contrast to patterns in White and Black ascendant neighborhoods, the overall mean SES of newcomers was lower than the SES of long-time residents in predominantly Hispanic ascendant neighborhoods. At first glance, one might initially conclude that the lower SES of newcomers suggests that neighborhood ascent in Hispanic neighborhoods is driven by incumbent upgrading. However, the intersection of race and nativity explains these population processes in ascendant Hispanic neighborhoods. When we included the interaction between nativity and race, we showed that non-Hispanic newcomers (both foreign- and U.S.-born) indeed had substantially higher SES relative to all Hispanic residents, thus driving the SES ascent of Hispanic neighborhoods. This pattern is consistent with the gentrification narrative that describes higher-SES non-Hispanic newcomers moving into historically low-income Hispanic neighborhoods, thereby accelerating the socioeconomic ascent of those neighborhoods. However, it appears that long-time native Hispanic residents comprise a relatively large share of the population in ascendant Hispanic neighborhoods, which would seem to suggest that a substantial proportion of incumbents are able to remain in the neighborhood as it experiences economic gain (at least temporarily). Nonetheless, the share of White residents has increased, while the share of foreign-born residents has decreased over time in ascendant Hispanic neighborhoods, suggesting some degree of demographic transition that may threaten the residential stability of long-tenured Hispanic immigrants.

Population processes in ascendant neighborhoods are complex. Our findings reveal that the importance of immigration in ascendant neighborhoods depends on the interaction of neighborhood racial composition and race/ethnicity of residents. In ascendant Black neighborhoods, Black immigrants (both incumbents and newcomers) seem to contribute socioeconomically to neighborhood ascent, while U.S.-born Blacks do not. However, non-Black immigrant newcomers contribute the most to neighborhood ascent followed by U.S.-born non-Black residents, suggesting that durable racial hierarchies still remain among Black and non-Black residents. In ascendant Hispanic neighborhoods, all Hispanic residents have lower average SES, though Hispanic immigrants have lower SES than U.S.-born Hispanics. The SES contributions to neighborhood ascent stem from movers, especially non-Hispanic immigrant movers, into Hispanic neighborhoods. In predominantly White ascendant neighborhoods, all groups lift the SES of the neighborhood, though the high-SES movers lift more than longer-term residents. Interestingly, foreign-born residents have higher SES than U.S.-born residents in ascendant White neighborhoods, suggesting a selective residential sorting process in which only high-SES immigrants are able to enter these neighborhoods.

Our findings generally show that in-moving newcomers tend to have higher SES than long-time residents in ascendant neighborhoods, thereby contributing most to socioeconomic change in the neighborhood. This finding leaves open the question about whether the economic fortunes of long-time residents that were able to stay in ascendant neighborhoods, particularly immigrant incumbents, improve as their neighborhoods experience socioeconomic ascent. We explored this question when we analyzed whether incumbents’ SES changed as their neighborhoods ascended (Table 5 and Fig. 5). While not the main drivers of ascent, we do find that the SES of long-term residents does generally improve in ascendant neighborhoods. The SES of Black residents, particularly Black immigrants, increased from 2000 to 2010 in Black ascendant neighborhoods, as did the SES of non-Black residents (albeit to a lesser degree). The same pattern holds in predominantly Hispanic ascendant neighborhoods—the SES of Hispanics and non-Hispanics alike living in predominantly Hispanic neighborhoods increased similarly over time, although these increases did not make up for the absolute SES disparities between incumbent Hispanics (immigrant and native-born alike) and their higher-SES non-Hispanic neighbors. There are, however, clear differences in SES growth between long-tenured native and immigrant residents in ascendant White neighborhoods, where the SES of incumbent U.S.-born residents decreased, both White and nonwhite. In ascendant predominantly White neighborhoods, immigrant incumbents are generally driving more of the increases in neighborhood SES, though the relative of size of the immigrant population in these neighborhoods is much smaller compared to all but predominantly Black ascendant neighborhoods.

Any increases in the SES of immigrants and racial/ethnic minorities that have been able to stay in ascendant neighborhoods should be viewed within the context of the demographic changes accompanying their neighborhoods. When we examined neighborhood compositional changes over time, we found stark differences in racial/ethnic and foreign-born population trends between ascendant and non-ascendant neighborhoods. The share of White residents declines in all predominantly White neighborhoods, but it declines more in non-ascendant, as compared to ascendant, neighborhoods. The opposite trend is largely observed, however, for the co-ethnic population in predominantly Black, Hispanic, and Asian neighborhoods and across all immigrant metro types. Moreover, while the presence of White residents decreases in non-ascendant nonwhite neighborhoods (Black, Hispanic, and Asian), it increases in nonwhite ascendant neighborhoods of all immigrant metro types, especially new destinations. This is consistent with prior work (Owens & Candipan, 2019), but also suggests a more pronounced process of racial transition accompanying SES ascent in metropolitan areas with rapid and recent immigrant growth relative to other areas. Given that this is occurring under the backdrop of great demographic change, in which the overall proportion of whites in the U.S. was declining, these divergences in the proportional presence of whites in ascendant minority neighborhoods are particularly noteworthy. Trends in immigrant growth at the neighborhood level largely mirror the patterns of changing racial composition—increasing shares of immigrants are more likely in non-ascendant neighborhoods compared to ascendant ones. While the foreign-born presence grew in all neighborhoods, it grew the most in non-ascendant neighborhoods. Taken together, this suggests that there may be limits to the economic gains that flow through ascendant neighborhoods experienced by immigrants and racial/ethnic minorities. Despite increasing demographic diversity at a national level, the unequal sorting into ascendant neighborhoods reproduces durable group hierarchies, by race and nativity, in White and nonwhite neighborhoods alike.

That both racial/ethnic and foreign-born change from 1990 to 2010 is most pronounced in new destination metros suggests potentially fundamental and complex differences in the sociodemographic composition of their ascendant neighborhoods relative to those located in established gateways or minor destinations. However, restricted census disclosure requirements limited our ability to show whether immigrant co-ethnics of either residential cohort are contributing to neighborhood SES ascent. Future analyses should further investigate whether sociodemographic processes underlying ascent at the neighborhood level varies across different types of metropolitan areas with different immigration histories.

Note that our study has descriptive aims—to document the demography of neighborhood ascent on a broad geographic scale, and across various neighborhood types. Future work can build on our findings. For example, our results suggest that the contributions of immigrants and co-ethnic residents may depend on the timing of immigration and neighborhood change. Future research should continue to examine whether the various pathways to ascent have changed over time, and since 2010, focusing on the interaction between race, immigration, and neighborhood sociodemographic composition, as some work has already begun doing (Hwang, 2020). Other research could trace the intergenerational residential pathways of immigrants into ascendant neighborhoods of all types. Although our data do not allow us to follow the same immigrant households over time, another issue concerns immigrants’ return to their home countries (e.g., Mexican immigrants in the wake of the 2008 recession), which may influence the SES levels of immigrant groups if those that are leaving have relatively lower SES than those that remain. Finally, future analyses could explore sociodemographic dynamics in neighborhoods experiencing other types of SES change trajectories, including socioeconomic descent.

Results from this study provide new knowledge about neighborhood change and, specifically, how the combination of race and immigration status help to explain neighborhood ascent. Given variation in group SES across different types of neighborhoods, the consequences of neighborhood ascent will depend on the interplay between household race/ethnicity, immigration, and initial neighborhood sociodemographic composition. Our findings underscore the need to account for each to understand population processes underlying neighborhood change. Past work finds that there is a racial/ethnic hierarchy upheld during the neighborhood ascent process. The present study shows that a hierarchy also exists within racial/ethnic groups and between immigrants and U.S.-born residents in these same neighborhoods. This implies the need to disentangle complex connections between race and immigration in urban areas across the U.S., particularly during a period of continuing demographic change on a national level, to fully understand heterogeneity in population processes accompanying ascent.