Gentrification and the Increasing Significance of Racial Transition in New York City 1970

Abstract

Gentrification is characterized as a spatial manifestation of economic inequality. An unsettled debate about gentrification is the extent to which it is also marked by distinct changes in neighborhood racial composition over time. This study uses a balanced panel of census data and retail data for New York City between 1970 and 2010 to extend prior research on the trajectory of gentrification and racial transition. This analysis finds an inverse relationship between Black and Latino residents and the pace of gentrification that increases over time. Consistent with theories of gentrification, it consistently trends with increasing household income. When income growth is disaggregated by race, Blacks and Latinos either have no effect or dampen the pace of gentrification by 2010. These findings support popular claims that even middle-class Blacks and Latinos are increasingly unable to remain in gentrifying neighborhoods as processes of change extend across the city.

Keywords

gentrification New York City racial transition

Introduction

A provocative article in the New York Times proclaims “The End of Black Harlem,” implicating gentrification in remaking Harlem, a worldwide symbol of Black urban culture, “for wealthier White people” (Adams 2016). NBC News announces that “Gentrification in Detroit Leaves Black-Owned Business Behind” (K. H. Taylor 2015). And a WNYC public radio and The Nation magazine podcast uses poignant narratives of residential dispossession linking gentrification to the evocative racial legacy of redlining, blockbusting, and White flight in “There Goes the Neighborhood: Race & Gentrification” (Carroll 2016). A growing body of qualitative research on gentrification supports journalistic accounts with evidence of racial (and class) conflict following the influx of middle-class White residents in historically Black and Latino neighborhoods in New York City, Chicago, Portland, Oregon and Washington, D.C. (Betancur 2002; Hyra 2015, 2017; Sullivan and Shaw 2011). Numerous interpretative accounts of gentrification reinforce popular beliefs that its processes and mechanisms mirror and possibly amplify the durable racial hierarchy in the United States and structure highly unequal and contentious urban landscapes (Betancur and Smith 2016; Kirkland 2008; McGee 1991; Moore 2009). Yet few quantitative studies find compelling evidence that gentrification is associated with racially patterned neighborhood changes (exceptions include Goetz 2011, Hwang and Sampson 2014, Wyly and Hammel 2004). In one national quantitative study, Freeman and Cai (2015) find an influx of White residents into Black neighborhoods during the first decade of the twenty-first century, a sharp reversal from the historical trend of White avoidance of Black neighborhoods, but the study is inconclusive as to whether trends are consistent with gentrification, whereas other national studies explicitly concerned with gentrification and economic growth during the 1990s find limited evidence of neighborhood racial transition (Ellen and O’Regan 2011; McKinnish et al. 2010).

Since British sociologist Ruth Glass first introduced the term gentrification into the lexicon of urban scholarship in 1964, characterizing it as a rapid process in which “all or most of the original working-class occupiers are displaced and the whole social character of the district is changed” (Glass 1964, p. xviii), this social and economic process of neighborhood restructuring has become one of the most controversial challenges facing cities in the twenty-first century. City leaders across the country have attempted to assuage the pace of gentrification and lessen its disruptive effects, particularly for low-income and poor residents, by increasing the supply of affordable housing or supporting living wage ordinances. Although such hard-won redistributive policies are necessary, they are likely inadequate for averting substantial “racial transition”— change in neighborhood racial composition of gentrifying neighborhoods that is statistically different from change observed citywide or in nongentrifying neighborhoods¹—and attendant racial conflict. That is because conventional antigentrification policy treats processes and practices animating change as principally spatial manifestations of economic inequality (Slater 2011), while underappreciating how concomitant social processes, such as racial preferences and the contemporary legacy of spatialized racial inequality, intersect with gentrification (Hwang and Sampson 2014; Kirkland 2008; Lees 2000; Smith 1996).

Recent studies have begun to advance gentrification research theoretically and empirically by systematically examining how the durable racial hierarchy that governs residential selection also influences the pace and the trajectory of gentrification during the 2000s (Hwang and Sampson 2014; Timberlake and Johns-Wolfe 2017). This article extends that research by undertaking a statistical periodization of gentrification across four intercensal periods between 1970 and 2010; by examining how racial transition, not just neighborhood composition, associates with the pace of gentrification over consecutive periods; and by interacting racial transition and household income examining how upwardly mobile or middle-class Blacks and Latinos affect the pace of gentrification. While this cohort has been touted as harbingers of gentrification in predominately Black and Latino neighborhoods in Chicago and New York City (Anderson and Sternberg 2013; Bostic and Martin 2003; Pattillo 2007; M. M. Taylor 2002), over time, the share of middle-class Black and Latino residents declines as the pace of gentrification increases.

This study relies on the Longitudinal Tract Data Base (LTDB), a balanced panel of census tracts for New York City joined with the National Establishment Time Series (NETS) database, to address three empirical questions. First, and most simply, are the socioeconomic, housing, and commercial attributes of gentrifying neighborhoods in New York City meaningfully different from nongentrifying neighborhoods each decade? And, if so, is the magnitude of racial transition over the course of a decade significant? Second, how does racial transition relate to gentrification over time? Finally, what can we say about the mutually constitutive nature of income growth and race on the pace of gentrification. How do Blacks and Latinos with the strongest income growth affect the pace of gentrification? In the following section, I review relevant literature and establish the theoretical framework for analyzing racial transition and gentrification. Then I describe the data and the gentrification construct used in this analysis, and outline the analytic strategy. The “Findings” section is organized around the aforementioned empirical questions. I conclude with a discussion of the planning and policy implications of this study.

Background and Theoretical Framework

Ruth Glass’s (1964) classic study of gentrification describes processes of urban restructuring in a working-class neighborhood in the London Borough of Islington, as affluent newcomers acquired “shabby, modest mews and cottages” and beleaguered “large Victorian houses,” and converted them into elegant and expensive quarters, consequently displacing poor and working-class residents. But Glass’s early articulation of gentrification does not explicitly discuss processes of racial transition, likely because during the 1950s and early 1960s, the ethno-racial composition of Islington was nearly 90% White English/Welsh/Scottish/Northern Irish.² Urban scholars in the United States adopted the term gentrification to depict similar processes of urban restructuring. Representations of American cities during the 1950s, 1960s, and 1970s as deplorable sites of abandonment, physical dilapidation, and economic devalorization inadvertently helped construct the requisite “rent gap” that primed disinvested areas for gentrification (Smith 1996). Yet most classic studies of gentrification in the United States have not thoroughly investigated how gentrification intersects with the legacy of racial segregation and other forms of racial injustice or how trajectories of gentrification may affect social groups differently (Atkinson 2003; Kirkland 2008). Many classic studies theorize gentrification in the United States by drawing evidence from aberrant neighborhoods of the day experiencing economic growth amid citywide decline, such as New York City’s SoHo, Baltimore’s Federal Hill, and Philadelphia’s Society (Smith 1979; Zukin 1982)—overwhelmingly White working-class neighborhoods able to stave off White-flight and widespread disinvestment. Consequently, these neighborhoods were considered more stable, desirable, and safer investments that, according to Schaffer and Smith (1986), White middle-class “gentrifiers” are less “squeamish” about for relocating.

Black and Latino inner-city neighborhoods were stereotyped as ghettos, marked by entrenched patterns of residential segregation (Massey and Denton 1993), White avoidance (Ellen 2000), and the ongoing legacy of urban renewal, redlining, racially restrictive covenants, blockbusting, and other racially discriminatory housing policies, practices, and social norms (Brooks and Rose 2013; Gotham 2000; Satter 2009). Yet, as gentrification advances, both temporally and spatially across cities, neighborhoods that previously seemed impervious to upgrading begin to exhibit signs of revitalization and eventually gentrification (Lees et al. 2010; Slater 2011; Smith 1996; Sutton 2010). Studies of Central Harlem, for example, exemplify a protracted process of racial transition. Schaffer and Smith (1986) respond to impressionistic reports of gentrification in Central Harlem during the 1970s. Drawing on U.S. census and housing data, the authors find nascent signs of gentrification that they attribute to a small number of affluent Black households. Although Central Harlem was 96% Black at the time of the study, Schaffer and Smith (1986) deduce that if gentrification proceeds, it will eventually lead to White influx and Black displacement. They argue that the racial wealth gap would limit the number of affluent Black households who can afford to relocate to or remain in Central Harlem as market pressures advance. Other studies of Central Harlem describe retail renewal and housing renovation as markers of gentrification during the mid- to late-1990s (M. M. Taylor 2002), and racial conflict and transition as symbolic of gentrification during the mid to late-2000s (Adams 2016; Freeman 2006; Hyra 2008). Stages of change in Central Harlem are consistent with research on national trends whereby “White invasion” into Black neighborhoods was an exception and “remained exceedingly rare” until 2000 (Freeman and Cai 2015, p. 306).

From a citywide perspective, Hwang and Sampson (2014) find the pace of gentrification in Chicago between 2007 and 2009 to be inversely related to the share of Black and Latino residents in 1995, and gentrification was further attenuated when the share of Black residents was greater than 40%. Hwang and Sampson (2014) attribute the threshold effect to the stigmatization of Black (and Latino) neighborhoods. Contrary to Freeman and Cai’s (2015) proposal that diminishing racism likely explains the unprecedented White entry into Black neighborhoods, Hwang and Sampson (2014) find that racial bias and perceptions of neighborhood disorder deter processes of gentrification, even more than observed neighborhood disorder, and likely contribute to the reproduction of racial inequality across the city.

Racial Transition, Neighborhood Diversity, Residential Mobility

Researchers use multiple measures and neighborhood attributes to discern how gentrification affects social group spatial relations. Freeman (2009) employs two gentrification constructs to analyze its effects for neighborhood diversity based on the assumption that once gentrification commences, neighborhood diversity declines. Freeman (2009) tests this assumption using an entropy index, a conventional analytic measure of racial diversity, and finds that racial diversity actually increases in gentrifying neighborhoods over the course of the study period. But, relative to other types of neighborhoods, gentrifying neighborhoods are more diverse at the beginning of the study period and typically remain diverse. However, the entropy index makes it difficult to surmise whether gentrification indeed exacerbates racial transition because the index measure how evenly groups are distributed across neighborhoods, but it obscures the direction of change by social group when more than two groups are included in the analysis. Although popular debates correlate gentrification with an influx of middle-class White residents, Freeman’s (2009) neighborhood-level analysis is inconclusive on the question of social mixing or entry or exit by social group.

Ellen and O’Regan (2011) improve upon Freeman’s (2009) analysis with a national study of racial residential mobility, both entry and exit, in economically growing low-income neighborhoods during the 1990s. They find that although neighborhood entrants were more likely to be White, the original minority residents were not more likely to exit gentrifying neighborhoods, dispelling claims of racialized patterns of neighborhood change and displacement. Ellen and O’Regan (2011) argue that perceptions of gentrification-related racial transition may rely on outmoded ideas about neighborhood change or focus solely on White entrants, to the exclusion of the larger story of transition. However, the White–non-White racial binary Ellen and O’Regan (2011) use to analyze resident mobility patterns limits attribution. It erases variability across social groups and implies that Blacks, Latinos, and Asians have comparable patterns of residential mobility and housing access. Yet the preponderance of literature on housing discrimination and residential segregation suggests otherwise (Brooks and Rose 2013; Massey and Denton 1993).

Timberlake and Johns-Wolfe (2017) take a different approach to understanding racial transition and gentrification focusing on Chicago and New York City. They find that for neighborhoods that gentrify by 2010, the percentage of Black or Latino residents in 1980 is positively associated with neighborhoods remaining majority Black or Latino rather than becoming majority White. Timberlake and Johns-Wolfe (2017) assert that past neighborhood ethno-racial composition predicts future gentrification. In other words, middle-class Whites do not gentrify all neighborhoods. In fact, gentrified neighborhoods are more likely to be majority White if they are proximate to the central business district and they have a very low percentage of Black residents in 1980, belying popular fears that gentrification is tantamount to “Negro removal redux.” It is worth noting that the approach Timberlake and Johns-Wolfe (2017) use underestimates transitioning neighborhoods that fail to meet the racial “majority” threshold. For example, a gentrifying neighborhood that was 90% Black and 10% White in 1980 and 52% Black, 40% White, and 8% Asian by 2010 would not be categorized as “gentrified White” in their study since it would not have reached the majority threshold. Moreover, it is questionable whether New York City and Chicago are useful comparison cities for studying racial transition and gentrification, given their distinct patterns of racial segregation during the 2000s. Furthermore, gentrification has progressed at an arrested pace across large swaths of Chicago’s predominately Black and Latino south and west sides, while gentrification in New York City began decades earlier (Zukin 1982, 1987) and has spread geographically to many historically Black and Latino neighborhoods. Nevertheless, both Timberlake and Johns-Wolfe (2017) and Hwang and Sampson (2014) advance our understanding of gentrification and neighborhood racial composition beyond the conventional class-based analysis. Yet both studies estimate gentrification during the last decennial period relative to racial composition during an arbitrarily chosen earlier period. While their temporal logic is sound for predicting gentrification, estimating just two periods obscures the uneven nature of gentrification that unfolds in varying degrees across successive periods.

Outstanding questions remain regarding how change in neighborhood racial composition associates with gentrification, and how the effects vary over time within a city as more neighborhoods experience gentrification. This study contributes to extant debates by first testing the validity of the gentrification construct by examining differences in the composition and character of gentrifying and nongentrifying New York City neighborhoods across five decennial periods. This research tests three primary hypotheses:

Hypothesis 1 (H1): In gentrifying neighborhoods, racial transition observed over the course of a decade becomes more significant by the end of the study period.

Hypothesis 2 (H2): As the pace of gentrification increases, the share of Black and Latino residents declines.

Hypothesis 3 (H3): Blacks and Latinos with the greatest income growth do not significantly advance the pace of gentrification over time.

Data, Measures, and Approach

This analysis utilizes two primary datasets. The first is the LTDB³ developed by researchers at the Spatial Structures in the Social Sciences at Brown University and combines U.S. Census data for 1970 to 2010 and the American Community Survey (ACS) five-year estimates (2006–2010) into a tract-level database that includes socioeconomic, demographic, and housing variables. The benefit of using the LTDB for this analysis is that it applies consistent 2010 tract boundaries back to 1970, correcting for inconsistent tract definitions that arise across decades. The LTDB also applies appropriate population weights in each decade to construct a balanced panel of tract-level data ideal for analyzing neighborhood changes across four intercensal periods (1970–1980, 1980–1990, 1990–2000, and 2000–2010) and attributing differences to the shifting population rather than changes in tract boundaries. The LTDB contains 2,116 census tracts representing New York City’s five boroughs—the Bronx, Brooklyn, Manhattan, Queens, and Staten Island. Consistent with prior studies of neighborhood change, I eliminate census tracts if, in any period, the population is less than 200 (e.g., John F. Kennedy Airport), primarily institutionalized (e.g., Rikers Island Correctional Facility), or the tract corresponds to open spaces (e.g., city parks and cemeteries). Approximately 17 tracts are consistently eliminated each decade. To exploit the longitudinal nature of the LTDB, the data are structured in a tract-year format in which socioeconomic, demographic, housing, and retail characteristics of census tracts represent specific neighborhoods in each decade.

The second dataset is the NETS, a proprietary establishment-level dataset constructed by Walls and Associates from Dunn & Bradstreet archival records that includes business characteristics for all establishments operational between 1990 and 2014. This analysis includes approximately 1,850 establishments per year representing 12 North American Industry Classification System (NAICS) categories (e.g., bars, cafés, convenience stores, hardware stores, pharmacies, restaurants, supermarkets, nail salons, barber shops, beauty supply stores, electronic stores, bookstores, and liquor stores). The NETS data were geocoded, spatially joined to census tracts, and merged with the LTDB to capture the changes in the retail typology and retail density (establishment count per acre) over time previously found to be associated with gentrification (Chapple and Jacobus 2009; Meltzer and Capperis 2017; Zukin 2011; Zukin et al. 2009). To provide a more reliable estimate of neighborhood retail dynamics, I use three-year pooled average for 1990–1993, 2000–2003, and 2010–2013 to construct the retail density variable. By linking NETS to census data, this analysis controls for the commercial environment known to influence gentrification. Since the annual NETS database began in 1990, retail density cannot be calculated for earlier periods.

Measuring Gentrification

Interpretive accounts of gentrification suggest that, more recently, it has become synonymous with an influx of middle-class White residents and concomitant decline in Black and Latino residents (Hyra 2015; Sullivan and Shaw 2011). Yet recent quantitative studies find that the pace of gentrification is inversely related to the share of Blacks and Latinos, and Whites tend to avoid areas of Black concentration, all else equal (Hwang and Sampson 2014; Timberlake and Johns-Wolfe 2017). However, as the number of gentrifiable White neighborhoods declines and the real-estate market remains strong, previously less desirable neighborhoods with a larger share of Black and Latino residents become more attractive. Testing the degree to which we observe significant racial transition in gentrifying neighborhoods, that is not fully determined by household income, requires developing a reliable measure of gentrification. Since there is no consensus definition of gentrification or discrete variable that reasonably reflects the dynamic process of neighborhood change, I develop a gentrification construct that capture crucial social, economic and spatial elements of gentrification yet allows for examining racially patterned changes in the gentrifying neighborhood of New York City over time.

Hammel and Wyly (1996) define gentrification as “the replacement of low-income, inner-city working-class residents by middle- or upper-class households, either through the market for existing housing . . . or new upscale housing construction” (p. 250). Smith (1998) emphasizes neighborhood and housing attributes, defining gentrification as “the process by which central urban neighborhoods that have undergone disinvestment and economic decline experience a reversal, reinvestment, and the in-migration of a relatively well-off, middle- and upper-middle-class population” (p. 198). Other scholars define gentrification succinctly as the production of urban space for progressively more affluent consumers (Hackworth 2002). Across overlapping definitions, at a minimum, gentrification is characterized by geographic specificity such as a neighborhood boundary, class stratification and economic inequality, capital reinvestment in housing and the built environment, and resident mobility.

Dependent variable

To consistently estimate the pace of gentrification across neighborhoods over time, I construct a “gentrification index” that is a unit weighted composite score derived from the sum of percentage point differences each period between 1970 and 2010 for three tract-level attributes: share of college-educated residents; share of neighborhood newcomers; and share of owner-occupied housing units. Across the study period, higher index scores correspond to a greater magnitude of gentrification. The dispersion between gentrifying neighborhoods and economically stagnant or declining neighborhoods is reflected by the low citywide average index score of .068, with a range of 2.27 and variance of .130. There is notably less dispersion among the 10% to 13% of neighborhoods considered to be “gentrifying” in any period. Among gentrifying neighborhoods, the average index score increases to .295 with a range of 1.20 and variance of .015. Neighborhoods are labeled gentrifying if they have index scores that are at least one standard deviation above the citywide average for the period, and the neighborhood was not upper-income in previous periods. Neighborhoods are upper-income if median household income is at least 40% above the citywide average.⁴

The three variables used to construct the gentrification index were selected to represent social, economic, and spatial dimensions of gentrifying processes:

First, when gentrification was introduced into the lexicon of scholarly literature, it was characterized as a change in the social class structure of neighborhoods (Freeman 2005; Glass 1964; Hammel and Wyly 1996; Smith 1996). Since class is notoriously difficult to measure, this study follows prior research and uses four-year college education as a reasonable proxy (Freeman 2005; Pattillo 2007; Keels et al. 2013). Educational attainment is preferable to income, which may overlook highly educated early-career professionals with relatively low current earnings but high earning potential. College attainment is also more likely to exclude blue-collar workers, who, over time, may have higher annual earnings than some early-career professionals but are typically labeled working-class, and thus not perceived as driving gentrification. Other economic measures such as median rent and home values have been used as measures of gentrification (Freeman 2005). However, across the study period, real property values in NYC were volatile and did not necessarily correspond with gentrification (Been et al. 2008). During the 1980s, late-1990s, and 2000s the city experienced steep increases in rents and home values in both gentrifying and nongentrifying neighborhoods, and deep declines during the 1970s and early-1990s. More importantly, though rising property values are commonly attributed to gentrification, it is not clear whether it lags or leads other indicators of gentrification.

The second variable used in the gentrification index is change in resident tenure, a measure of neighborhood newcomers. In gentrifying neighborhoods, we would expect the share of residents with tenure of less than 10 years to increase. Change in residential patterns are central to gentrification, but increases in the share of residents with shorter tenure is not a sufficient indicator of gentrification by itself. We are likely to also observe an influx of newcomers in nongentrifying affordable neighborhoods receiving low-income residents displaced from gentrifying neighborhoods.

The third variable in the index captures the housing ownership structure, specifically, change in the share of renter-occupied housing units relative to total occupied units is considered a measure of neighborhood reinvestment. In global cities such as New York, gentrification is increasingly associated with a decline in renter-occupied units and concomitant increase in owner-occupancy as a result of real property conversions, such as multifamily residences to single-family structures, industrial properties to upscale lofts, multi-unit rentals to cooperative ownership, the development of new luxury condominiums, and eroding protections on affordable housing (Curran 2007; Goetz 2011; Hamnett and Whitelegg 2007). Although most occupied housing units in NYC were still rentals by 2010, the share of rentals declined as owner-occupancy steadily increased citywide.

Independent variables

As previously mentioned, this study includes the share of non-Hispanic Asian, non-Hispanic Black, non-Hispanic White, and Latino residents as the variables of interest for examining racial trends. Since percentage points for Whites are highly correlated with other groups, Whites are excluded in the regression estimates. Median household income is included as an explanatory variable when modeling gentrification. Although income is not a sufficient measure of social status, we would expect significant gains in household income to trend with gentrification, all else equal. In other words, as highly educated newcomers relocate to neighborhoods, possibly as homeowners, neighborhood median income will likely rise in concert. Therefore, change in median income by race is used to further test how racial groups affect the pace of gentrification each period. Other attributes known to affect neighborhood desirability, such as home value, population density, age cohorts, employment status, nativity, and retail density are controlled for in the regression models.

Analytic Approach

This analysis begins by comparing gentrifying and nongentrifying neighborhoods to establish the validity of the gentrification construct and indicate general trends. Table 1 presents the socioeconomic, housing, and commercial composition for the average gentrifying and nongentrifying neighborhood and the city overall, and tests mean differences. Based on previous literature, we would expect gentrifying and nongentrifying neighborhoods to be substantially different with regard to household income, professional occupation, college education, resident mobility, retail amenities, home value, and rent, for instance. To test the overarching hypothesis that racial transition is associated with gentrification, I first use difference of means tests to analyze the significance of racial transition each decade in both gentrifying and nongentrifying neighborhoods (Table 2). Then I use regression analysis to show how racial transition affects the pace of gentrification and how more affluent Blacks and Latinos affect gentrification, over time (Table 3). This study concludes with a discussion of the implications of racial transition and gentrification for policy and planning.

Table 1.

Mean Differences between Gentrifying and Non-Gentrifying Neighborhoods in New York City.

	A				B				C				D
	1980				1990				2000				2010
	All NYC	Non-Gentrifying	Gentrifying		All NYC	Non-Gentrifying	Gentrifying		All NYC	Non-Gentrifying	Gentrifying		All NYC	Non-Gentrifying	Gentrifying
Population	3,354	3,343	3,589		3,462	3,414	4,081	^***	3,788	3,839	3,379	^**	3,833	3,910	3,071	^***
% Asian	2.9%	3.0%	2.8%		6.5%	6.5%	6.7%		10.3%	10.6%	8.7%	^***	12.5%	13.2%	11.0%	^**
% Black	23.8%	25.0%	24.9%		25.4%	28.3%	9.6%	^***	27.0%	29.5%	19.2%	^***	24.8%	26.9%	22.2%
% Latino	18.7%	20.3%	19.6%		22.6%	25.8%	13.2%	^***	24.9%	28.2%	17.0%	^***	26.2%	29.9%	20.8%	^***
% White	53.6%	50.8%	51.6%		45.1%	38.9%	70.2%	^***	35.7%	29.5%	53.3%	^***	34.1%	27.5%	43.1%	^***
Median Income^	$14,829	$13,600	$14,404		$31,397	$27,402	$41,446	^***	$40,775	$35,678	$48,215	^***	$54,763	$46,379	$60,999	^***
Asian	$21,173	$18,723	$25,664	^***	$37,956	$34,903	$42,704	^***	$49,566	$45,205	$56,127	^***	$66,703	$57,121	$80,038	^***
Black	$14,561	$13,962	$13,217	^**	$30,692	$27,668	$37,358	^***	$41,086	$37,867	$43,395		$51,031	$46,691	$51,781
Latino	$14,305	$13,392	$13,399		$29,666	$26,883	$34,903	^***	$39,098	$35,243	$45,531	^***	$51,384	$45,555	$54,588
White	$18,748	$16,947	$22,712	^***	$30,961	$26,931	$42,957	^***	$42,372	$37,128	$51,261	^***	$60,973	$51,446	$71,505	^***
Demographics
% Age18to24	11.8%	11.9%	12.3%		10.6%	11.0%	8.7%	^***	10.0%	10.2%	9.2%		10.3%	10.5%	10.1%
% Age25to34	16.3%	15.7%	22.1%	^***	18.4%	18.1%	21.5%	^***	16.7%	16.3%	18.7%	^***	16.6%	15.5%	21.8%	^***
% Age35to44	11.6%	11.2%	13.5%	^***	15.0%	14.5%	17.5%	^***	15.7%	15.6%	16.5%	^***	14.5%	14.2%	15.3%	^***
% Age45to54	10.8%	10.7%	10.6%		10.6%	10.2%	11.3%	^***	12.7%	12.3%	13.4%	^***	13.8%	13.7%	13.2%	^**
% Age55to64	10.7%	10.7%	9.3%	^***	9.0%	8.8%	9.8%	^***	8.6%	8.2%	9.3%	^***	10.6%	10.4%	10.2%
% Age65up	13.5%	13.7%	11.6%	^***	13.4%	12.9%	16.4%	^***	11.9%	11.4%	14.8%	^***	12.2%	12.0%	11.3%	^**
% Foreign Born	23.0%	24.0%	19.3%	^***	27.7%	28.8%	24.7%	^***	35.5%	37.6%	29.9%	^***	37.0%	39.7%	32.3%	^***
% College Education	14.1%	11.0%	28.5%	^***	19.7%	14.7%	42.2%	^***	24.4%	19.3%	40.4%	^***	30.5%	24.1%	43.9%	^***
% Unemployed	4.7%	4.8%	5.2%	^**	5.8%	6.3%	4.0%	^***	5.9%	6.2%	5.8%		5.7%	6.0%	5.5%
% Professional	20.0%	17.0%	31.8%	^***	24.5%	20.3%	43.2%	^***	30.0%	25.8%	42.5%	^***	32.1%	26.8%	42.8%	^***
% Tenure <10 years	24.0%	23.3%	33.6%	^***	20.8%	20.0%	28.2%	^***	22.0%	21.1%	27.7%	^***	19.9%	18.6%	25.4%	^***
Neighborhood
Median Rent^	$223	$212	$245	^***	$488	$457	$577	^***	$694	$648	$783	^***	$1,043	$968	$1,109	^***
Home Value^	$53,266	$46,892	$81,465	^***	$193,817	$172,312	$300,108	^***	$231,672	$211,417	$283,714	^***	$532,857	$495,670	$580,571	^***
% Renter-Occupied	64.8%	67.8%	71.2%	^***	62.4%	66.6%	57.7%	^***	61.7%	64.9%	59.5%	^**	57.1%	60.6%	55.6%	^**
Retail Density (per acre)	n/a	n/a	n/a		0.662	0.552	1.275	^***	0.855	0.662	1.714	^***	1.285	1.032	1.636	^***
N	2092	1677	218		2103	1661	279		2105	1674	227		2097	1572	281

dollars reported in constant 2010 values.

p <.1. **p < .05. ***p < .01.

Table 2.

Racial Transition Over the Course of a Decade for Gentrifying and Nongentrifying Neighborhoods.

I	A			B			C			D
Gentrifying Neighborhoods	1980 (n = 218)	Period Change (1970–1980)		1990 (n = 279)	Period Change (1980–1990)		2000 (n = 227)	Period Change (1990–2000)		2010 (n = 281)	Period Change (2000–2010)
Population	3,588	−704	***	4,080	167		3,379	194		3,071	−81
% Asian	2.8	0.009	***	6.7	0.033	***	8.7	0.042	***	11.0	0.025	***
% Black	24.9	0.152	***	9.6	−0.026		19.2	0.009		22.2	−0.068	***
% Latino	19.6	n/a		13.2	−0.016		17.0	−0.012		20.8	−0.026
% White	51.6	−0.356	***	70.2	0.017		53.3	−0.052		43.1	0.063	***
II	A			B			C			D
Nongentrifying Neighborhoods	1980 (n = 1,677)	Period Change (1970–1980)		1990 (n = 1,661)	Period Change (1980–1990)		2000 (n = 1,674)	Period Change (1990–2000)		2010 (n = 1,572)	Period Change (2000–2010)
Population	3,343	−498	***	3,414	109		3,839	354	***	3,910	106
% Asian	3.0	0.017	***	6.5	0.035	***	10.6	0.036	***	13.2	0.022	***
% Black	25.0	0.059	***	28.3	0.020	**	29.5	0.027	***	26.9	−0.007
% Latino	20.3	n/a		25.8	0.051	***	28.2	0.037	***	29.9	0.052	***
% White	50.8	−0.286	***	38.9	−0.101	***	29.5	−0.117	***	27.5	−0.071	***

Note. n/a—U.S. Census data not disaggregated for Hispanic/Latino population in 1970.

p <.1. **p < .05. ***p < .01.

Table 3.

Regression Estimates of the Effects of Racial and Income Change for Gentrification Over Time (1980–2010).

	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Prior gentrification	−0.104***	−0.0958***	−0.0953***	−0.0963***	−0.101***	−0.102***
Prior gentrification	(0.00779)	(0.00699)	(0.00693)	(0.00698)	(0.0114)	(0.0107)
Percentage point Asian	0.0553	−0.0522	0.293***	−0.0542	−0.179*	−0.209**
Percentage point Asian	(0.0462)	(0.0494)	(0.0729)	(0.0512)	(0.0934)	(0.0956)
Percentage point Black	−0.229***	−0.240***	−0.143**	−0.123***	−0.332***	−0.350***
Percentage point Black	(0.0436)	(0.0368)	(0.0569)	(0.0333)	(0.0692)	(0.0761)
Percentage point Latino	−0.471***	−0.485***	−0.221***	−0.382***	−0.494***	−0.532***
Percentage point Latino	(0.0519)	(0.0423)	(0.0609)	(0.0439)	(0.0865)	(0.0920)
Percentage change median income		0.0229***	0.0591***
Percentage change median income		(0.00572)	(0.00738)
Large gain median income				0.0420***
Large gain median income				(0.00715)
Percentage change Med Inc Asian					−0.000901	−0.0117***
Percentage change Med Inc Asian					(0.00116)	(0.00250)
Percentage change Med Inc Black					0.00897	0.00215
Percentage change Med Inc Black					(0.00476)	(0.00725)
Percentage change Med Inc Latino					0.00529	0.0318***
Percentage change Med Inc Latino					(0.00441)	(0.00932)
Percentage point foreign		0.0679***	0.0445*	0.0635**	−0.0209	−0.0312
Percentage point foreign		(0.0254)	(0.0268)	(0.0266)	(0.0445)	(0.0451)
Percentage point 25 to 34		0.171**	0.376***	0.503***	0.558***	0.549***
Percentage point 25 to 34		(0.0852)	(0.103)	(0.0986)	(0.137)	(0.161)
Percentage point 35 to 44		0.171**	0.433***	0.573***	0.420***	0.333**
Percentage point 35 to 44		(0.0900)	(0.118)	(0.116)	(0.134)	(0.163)
Percentage point 45 to 54		0.374***	0.417***	0.469***	0.537***	0.459***
Percentage point 45 to 54		(0.0818)	(0.0791)	(0.0803)	(0.194)	(0.173)
Percentage point 65 up		−0.311***	−0.00487	−0.0528	−0.431***	−0.426***
Percentage point 65 up		(0.0828)	(0.0868)	(0.0887)	(0.140)	(0.148)
Percentage point unemployment		−0.369***	−0.317**	−0.324***	−0.426**	−0.396**
Percentage point unemployment		(0.103)	(0.127)	(0.116)	(0.200)	(0.186)
Percentage point population		−0.00777***	−0.00728***	−0.00710***	−0.0846***	−0.101***
Percentage point population		(0.00140)	(0.00173)	(0.00160)	(0.0327)	(0.0306)
Percentage point home value		3.95e-05***	6.47e-05***	4.66e-05***	4.42e-05*	4.54E-05
Percentage point home value		(1.28e-05)	(1.67e-05)	(1.81e-05)	(2.63e-05)	(2.53e-05)
Retail density (per acre)		0.0250***	−0.00122	0.00256	0.00699	0.0162
Retail density (per acre)		(0.00513)	(0.00521)	(0.00576)	(0.00713)	(0.0104)
Year = 2000			0.101***	0.0574***		0.0115
Year = 2000			(0.0150)	(0.00849)		(0.0218)
Year = 2010			0.0984***	0.0801***		−0.00757
Year = 2010			(0.0132)	(0.00942)		(0.0231)
Year2000#c.chAsian			−0.455***
Year2000#c.chAsian			(0.0673)
Year2010#c.chAsian			−0.383***
Year2010#c.chAsian			(0.0722)
Year2000#c.chBlack			0.0383
Year2000#c.chBlack			(0.0506)
Year2010#c.chBlack			−0.214***
Year2010#c.chBlack			(0.0717)
Year2000#c.chLatino			−0.182***
Year2000#c.chLatino			(0.0675)
Year2010#c.chLatino			−0.302***
Year2010#c.chLatino			(0.0743)
LgGain#c.chAsian				0.357***
LgGain#c.chAsian				(0.0883)
LgGain#c.chBlack				−0.191**
LgGain#c.chBlack				(0.0867)
LgGain#c.chLatino				−0.0889
LgGain#c.chLatino				(0.0701)
Year2000#c.chMedIncAsian						0.00985***
Year2000#c.chMedIncAsian						(0.00282)
Year2010#c.chMedIncAsian						0.0198***
Year2010#c.chMedIncAsian						(0.00420)
Year2000#c.chMedIncBlack						0.0125
Year2000#c.chMedIncBlack						(0.0117)
Year2010#c.chMedIncBlack						−0.000835
Year2010#c.chMedIncBlack						(0.00800)
Year2000#c.chMedIncLatino						−0.0322***
Year2000#c.chMedIncLatino						(0.0114)
Year2010#c.chMedIncLatino						−0.0284***
Year2010#c.chMedIncLatino						(0.0104)
Constant	0.0899***	0.0498***	−0.0259**	0.0188**	0.0872***	0.0774***
Constant	(0.00217)	(0.00686)	(0.0114)	(0.00771)	(0.0129)	(0.0171)
Observations	6,289	5,314	5,314	5,314	2,229	2,229
R ²	.130	.172	.236	.207	.274	.297
Number of tracts	2,104	1,842	1,842	1,842	1,215	1,215
Tract Fixed Effects	YES	YES	YES	YES	YES	YES
Year Fixed Effects	YES	YES	YES	YES	YES	YES

Note. Robust standard errors in parentheses.

p < .1. **p < .05. ***p < .01.

Findings

Gentrifying and Nongentrifying Neighborhoods

For neighborhoods labeled “gentrifying” in any decade, we can assume that processes of change were evident during the prior period, although we are unable to precisely identify the beginning (or ending) of the process. Conversely, neighborhoods are considered “nongentrifying” if they qualify for gentrification, meaning they were not upper-income in prior periods, but their index score was not significantly greater than the citywide average for the period. To establish that gentrifying and nongentrifying neighborhoods in this study are meaningfully different from each other each period, in ways that we would assume, Table 1 compares them based along socioeconomic, housing, and commercial attributes. For example, since gentrification is associated with an influx of more affluent residents and investment capital, we would expect attributes associated with affluence, such as household income, the share of residents in professional occupations, and median rend, to name a few, to be consistently higher in gentrifying neighborhoods.

The variables used to construct the gentrification index follow expected patterns. Specifically, the percentage of college-educated residents is nearly twice as high in gentrifying neighborhoods, the share of newcomers is 6 to 10 percentage points higher in gentrifying neighborhoods, and renter occupancy is almost consistently lower in gentrifying areas, suggesting that gentrification indeed signifies neighborhood reinvestment in the form of home ownership. The slightly higher percentage of renter-occupancy in 1980 likely reflects the city’s fiscal crisis and depressed homeownership rates during the 1970s. The city’s fiscal uncertainty made consumers and banks reluctant to make long-term investments in real estate.

Neighborhoods that gentrified during the 1970s, on average, had comparable racial composition and median income as nongentrifying neighborhoods, see column A in Table 1. White-flight, suburbanization, and deindustrialization characterized 1970s New York, and the composition of the city was about 50% White, 25% Black, 20% Latino, and 3% Asian. Although the overall median income was not significantly higher in gentrifying areas, Whites and Asians in those areas had higher median income than those in other neighborhoods, suggesting that more affluent Whites remained in gentrifying areas, resisting White-flight. Surprisingly, median income for Blacks in 1980 was significantly lower in gentrifying neighborhoods. A possible explanation is that some of the tracts that gentrified during the 1970s contained clusters of public housing that low-income Black households disproportionately occupied. While some middle-class Blacks likely entered gentrifying neighborhoods, as Bostic and Martin (2003) contend, public housing occupants likely remained, depressing the median income in those neighborhoods.

Columns B and C in Table 1 show neighborhoods that were gentrifying during the 1980s and 1990s. Differences between gentrifying and nongentrifying neighborhoods are significant for most attributes and generally consistent with expectations based on prior research. For example, Flores and Lobo’s (2012) study of demographic change in NYC describes White-dominated neighborhoods that became more racially mixed communities throughout the 1970–2010 period. They argue that the city was becoming increasingly diverse, but Blacks were less able to access housing in formerly all-White neighborhoods. In this study, I find that during the 1980s nongentrifying neighborhoods became more racially mixed at the same time that gentrifying neighborhoods became disproportionately White (70.2%). Over the next decade, gentrifying neighborhoods were notably less White, although still majority White, and more Black. By 2010, the average gentrifying neighborhood was no longer majority White. This can be interpreted in multiple ways. But it is important to remember that Table 1 depicts the composition different compilations of neighborhoods at the end of each decade. It does not reveal the magnitude of change across each period.

In addition, it is also worth noting that median household income for Blacks in 2000 and 2010 (columns C and D) was not significantly higher in gentrifying neighborhoods relative to nongentrifying, which is somewhat surprising. It is possible that Black households entering gentrifying neighborhoods had substantially lower incomes than existing Black households, but this seems unlikely given high economic barriers to entry in gentrifying neighborhoods. Alternatively, income parity across neighborhood types may reflect an out-migration of working-class and middle-class Blacks from gentrifying neighborhoods, leaving behind lower income Blacks and perhaps more affluent Black newcomers. Debates about displacement tend to focus on vulnerable poor renters, for good reason, but these discussions typically ignore voluntary and involuntary out-migration of slightly better-off middle-class Black households from rapidly gentrifying neighborhoods that are no longer affordable. While inflated rents are a burden for many residents, the confluence of wage stagnation, the racial wealth gap, and discriminatory housing practices exacerbate the burden of affordability and housing precarity for Black and Latino residents (Conley 2010; Darity and Hamilton 2012; Kochhar and Fry 2014; Oliver and Shapiro 1993).

(H1) Racial Transition Increases Over the Study Period

Racial conflict manifest during recent periods of gentrification in cities across the country (Hyra 2017) may in fact reflect processes of change in the racial structure of neighborhoods that differ from previous periods (Freeman and Cai 2015). Table 2 compares racial transition over the course of each decade for gentrifying and nongentrifying neighborhoods, and shows variability in the magnitude of transition across periods. Specifically, Table 2 shows that neighborhoods gentrifying during the latter period begin with a larger share of Black and Latino residents in the beginning of the decade than prior periods, and experience the largest decline in the proportion of Blacks and Latinos. Column A shows that during the 1970s, changes in the population overall and by race were significant across the city. In gentrifying neighborhoods, for instance, the share of Blacks increased by 15 percentage points, and the share of Whites declined by 36 percentage points. This considerable decline in percentage White may reflect racial restructuring in cities across the country as White-dominated neighborhoods temporarily diversified with White-flight and easing of restrictive covenants, but eventually flipped to majority minority neighborhoods (Gotham 2000; Oliver and Shapiro 1993; Satter 2009). The growing share of Blacks in gentrifying neighborhoods corroborates prior research that finds the Black middle-class contributed to gentrification during the 1970s (Bostic and Martin 2003). However, the findings for the 1970s are not fully reliable because the U.S. Census data did not begin disaggregating Hispanic ancestry until the 1980 Census. Therefore, Latinos were most likely classified as White, and to a lesser degree as Black, biasing observed trends during that period.

During the 1980s and 1990s (columns B and C), gentrifying neighborhoods were majority White and remained White, although the White share was declining. During both periods, changes to racial composition for Whites, Blacks, and Latinos was nominal in gentrifying areas. Although gentrifying neighborhoods remained fairly stable during the 1980s and 1990s, racial churn in nongentrifying neighborhoods was significant for all groups. White residents consistently fled nongentrifying areas of the city each decade, while other groups entered these neighborhoods.

During the first three periods of this study, White residents were overrepresented in gentrifying neighborhoods relative to city averages, and the proportion of Black and Latino residents was consistently less than the city average. By 2010 (column D), however, there were notable changes in the racial composition of gentrifying neighborhoods. Most importantly, this was the first period that the average gentrifying neighborhood was majority Black and Latino (52%) in the beginning of the period, and majority White and Asian by the end of the period. Over the course of the decade, there was a significant decline in the share of Blacks, by 6.8 percentage points, and significant growth in the share of Whites, by 6.3 percentage points.

Modeling Gentrification

This section models gentrification to show how it associates with racial transition and household income across multiple periods, all else equal. As discussed previously, prior literature on gentrification emphasizes class conflict that ensues after the entry of more affluent residents and the displacement of low-income, often minority, residents (Glass 1964; Smith 1996). Numerous scholars examine how racial homogeneity, intra-group class conflict, and racial solidarity affect processes of gentrification, as middle-class Blacks and Latinos relocate to historically Black and Latino neighborhoods (Betancur 2002; Boyd 2008; Hyra 2008; Moore 2009; Pattillo 2007; M. M. Taylor 2002). Although it is widely presumed that the entry of middle-class Whites and the exit of less affluent Black and Latino residents is the more likely trajectory of gentrification, few studies examine how racial change affect the magnitude or pace of gentrification (Hwang and Sampson 2014; Timberlake and Johns-Wolfe 2017). Although racial transition, specifically from Black or Latino toward White, has been associated with gentrification in New York City since the 1980s, albeit nominaly, (Mele 2000; Smith 1996; Schaffer and Smith 1986), there is reason to believe that it has become increasingly pronounced. White neighborhoods are more likely to have gentrified during early periods. But as gentrification spreads geographically across the city, non-White neighborhoods also change—initially through intra-group class upgrading (Moore 2009; Pattillo 2007; M. M. Taylor 2002), but then through racial transition.

To investigate how racial transition associates with gentrification across multiple periods, this study uses a series of dynamic panel regression models with a lagged dependent variable and neighborhood fixed effects. Gentrification is the dependent variable of interest, and the primary explanatory variables are percentage point by race—for Asian, Black, and Latino (Whites are excluded because they are highly inversely correlated with change for Blacks and Latinos)—and percentage change in median household income. Dynamic panel models include a lag for gentrification to allow the influence of the primary explanatory variables to be better interpreted for the current period (time t) after accounting for gentrification that may have begun in the previous period (time t − 1) and controlling for unobservable and time-invariant neighborhood factors that may affect the pace of gentrification.⁵

The basic empirical models used are the following:

G α,_{it} = β 0 + G β_{Git - 1} + β R_{it} + μ_{i} + τ_{t} {+ u}_{it},

G α,_{it} = β 0 + G β_{Git - 1} + β R_{it} + β Y_{it} + β X_{it} + μ_{i} + τ_{t} {+ u}_{it},

G α,_{it} = β 0 + G β_{Git - 1} + β R_{it} + β Y_{it} + β X_{it} + β R_{it} \times Y_{it} + μ_{i} + τ_{t} {+ u}_{it},

where i and t are indexes for neighborhood and year, respectively. Model 1 is a simple baseline test of whether change in neighborhood racial composition affects the pace of gentrification overall after accounting for gentrification in prior periods. Gα_it is the gentrification-index-dependent variable; the parameter Gβ is the lagged variable for gentrification, an indicator variable for gentrification in the prior period; and R_it represents the explanatory variables of interest (racial change). Model 2 adds change in median household income (in constant 2010 dollars), Y_it and a vector of neighborhood-level socioeconomic, housing, and commercial characteristics associated with gentrification X_it. Models 3 to 6 include interaction terms to isolate the effects of race and time, and race and income R × Y. This allows for the examination of differences in the pace of gentrification each period and the testing of whether the pace of gentrification is modified differently for higher income Blacks and Latinos, for example. µi and τt are neighborhood and time fixed effects, respectively, and u_it is the error term. The parameter of interest is βR, which measures the direct impact of racial change on gentrification. If changes in racial composition affect gentrification through channels other than growth in neighborhood income, then we expect βR to be significant. Specifically, we would expect growth in the proportion of Blacks or Latinos to reduce gentrification by the value of βR on average. Likewise, the parameter βy represents the elasticity of gentrification to household income. We expect change in income to be positive and significantly greater than zero in all model specifications.

(H2) The Share of Black and Latino Residents Declines With Gentrification

Table 3 presents regression results for all models. The results of the baseline model 1 generally support popular claims that as the share of Black and Latino households increases, the pace of gentrification declines, by 23 and 47 percentage points, respectively.⁶ The baseline model also includes neighborhood and time fixed effects. This controls for unobserved time-invariant factors within neighborhoods—such as transit access, zoning and land-use ordinances, regulatory issues, and aggregate investments—potentially correlated with gentrification. However, this general model obfuscates potential variability across periods. We cannot determine whether effects are greater in latter periods, as hypothesized. Moreover, the coefficient on the lag variable is negative and significant, suggesting that earlier gentrification tempers gentrification in the current period.

Since gentrification is commonly associated with the entry of more affluent households, model 2 adds change in median household income and other population, housing, and retail attributes. The coefficients for Blacks and Latinos remain robust. Both racial groups depress the pace of gentrification, while Asians seem to have no significant effect. The coefficients conform to interpretive accounts about gentrification and racial transition. That is to say, gentrifying neighborhoods are associated with a declining minority population and strong economic growth, after controlling for gentrification during prior periods. Coefficients on other neighborhood attributes, such as higher retail density and inflated home values, follow expected patterns.

A primary concern of this research is understanding how changes in neighborhood racial composition articulate with gentrification, and how this likely differs across periods. Model 3 introduces period effects using a dummy variable for decade and creating interaction terms to segment racial change by decade and test whether the negative effects of the share of Black and Latino residents increase over time. These results show descriptive statistics are robust with respect to racial transition and gentrification. When conditioned on the main effects, growth in the percentage of Blacks has no significant effect in the first period but in subsequent periods significantly reduces gentrification. By the last period, the pace of gentrification declines by 35% in neighborhoods with a growing share of Black residents. The negative effects of a growing share of Latinos in a neighborhood is significant in each period, with the magnitude increasing over time. The coefficients on the interaction terms for Asians are also negative and significant in each period, but the negative effect declines over time. In other words, as the proportion of Asians increases in 2000 and 2010, gentrification declines slightly less each period, by 16% and 9% respectively.

(H3) The Pace of Gentrification Dampened by Affluent Blacks and Latinos

The next three models extend the analysis by using interactions between racial composition and neighborhood economic growth, and income growth by race each period. Model 4 tests the effects of racial change in neighborhoods with the strongest economic gains, excluding upper-income neighborhoods. These large gain (LgGain) neighborhoods are in the top income growth quintile. Approximately 43% of neighborhoods were in the top quaintile and gentrifying in the same period. In 2010, this includes neighborhoods such as Bedford-Stuyvesant, Midtown South, East Harlem, Greenpoint, and Downtown Brooklyn, among others. Whereas neighborhoods such as Baychester, Richmond Hill, the Lower East Side, and East New York were also large gainers but without concurrent gentrification. The coefficient on the interactions suggest that, overall, the magnitude of gentrification is variable in neighborhoods with the largest economic gains and it is tempered by racial transition. For example, as the share of Black residents increases the pace of gentrification is significantly depressed. This might look like working-class and middle-class Blacks relocating from gentrifying neighborhoods such as Bedford-Stuyvesant and Harlem to low income non-gentrifying neighborhoods, such as East New York and Brownsville, causing median income to rise in lieu of additional signs of gentrification, at least in the short term.

Models 5 and 6 examine the mutually constitutive nature of racial identity and income on gentrification. Even if the percentage of Black and Latino residents in gentrifying neighborhoods declines overall, it is not clear whether this holds across the income distribution. To examine how middle-class and working-class Blacks and Latinos affect gentrification, model 5 replaces change in median income overall with change in median income by race. While previous models show that income growth consistently increases the pace of gentrification, the results from model 5 shows no significant effect when income growth is disaggregated by race. It is possible that effects are decade specific, therefore erased across the study period.

Model 6 further addresses the question of how median income by race articulates with gentrification by periodizing effects using Dummy year × Change in median income by race interaction terms. The results disaggregate effects erased in the previous model. For example, the coefficients on the interaction terms in model 6 have the opposite sign as the main effects, thereby muting main effects. For Asians, this means that although the coefficient on Asian household income is negative, when interacted with period dummies, the effects are consistently positive and significant each period. Although the magnitude of the effects are small each period, they become slightly larger by the last period, suggesting that the pace of gentrification increases with Asian median income. Conversely, as Latino household income increases, the rate of gentrification significantly declines both periods. Although growth in Black household income shifts from a positive coefficient to negative by 2010, there is no statistically significant impact on the pace of gentrification in either period. This is somewhat surprising considering the consistent inverse relationship between the pace of gentrification and the share of Black residents. But it concurs with the descriptive statistics showing no meaningful differences in income for Blacks in gentrifying and non-gentrifying neighborhoods in 2000 and 2010. Furthermore, the negative coefficient on Black median income in 2010 (Model 6) may reflect a loss of middle-class Blacks.

This study finds compelling evidence of racial transition in gentrifying neighborhoods. The results presented here indicate that, overall, the pace of gentrification is inversely related to the share of Black and Latino residents. The negative effects seems to increase over time, even in neighborhoods with the strongest economic growth, and as household incomes increase for Latinos. Although much of the gentrification literature focuses on the vulnerability of low-income residents to displacement, the findings here suggest that middle-income Blacks and Latinos have become increasingly susceptible to displacement or voluntary relocation in gentrifying neighborhoods. It is reasonable to infer from well-documented evidence on the racial wage gap, discrimatory practices, and the ever-expanding geographies of gentrification across the city that thwart entry and retention in the average gentrifying neighborhood is increasingly challenging for many middle-income Blacks and Latinos. For example, according to data for 2010 Black median income in gentrifying neighborhoods ($51,781) was comparable with Black median income citywide ($51,031), yet markedly below the citywide median income of Asians ($80,038) and Whites ($71,505). Moreover, by 2010, the average median rent and home value in gentrifying neighborhoods had increased 42% and 105%, respectively, over the previous period. Yet Black and Latino median income in these neighborhoods was up just 19% and 20%, respectively, over the same period. Comparatively, median income for Asians and Whites kept pace with rent inflation, increasing 42% and 39%, respectively.

Implications for Planning and Policy

City leaders around the country routinely extol the influx of affluent households, “creative class” millennials, and investment capital for generating much needed tax revenue and spurring investment in previously forlorn areas. Yet gentrification remains contentious for the way it exacerbates existing inequalities. Theorizing gentrification as solely a spatial manifestation of economic inequality obscures the fact that ethno-racial inequality operates alongside economic inequality. As racial formation theory reminds us, race is experienced and reinforced through societal institutions, ideologies, and norms that police the boundaries of racial categories that are not reducible to class or eroded by income (Omi and Winant 1994). This article provides a nuanced picture of gentrification and the durable racial hierarchy by constructing a periodization of neighborhood change in New York City, between 1970 and 2010, and testing the claim that sizable change in neighborhood racial composition, specifically decline of Black and Latino households, signals gentrification.

The results of this analysis show an inverse relationship between the pace of gentrification and the share of Black and Latino residents, all else equal. And these trends seem to increase over the study period. On the one hand, the declining share of Black and Latino in gentrifying neighborhoods support popular fears that, over time, gentrification likely alters neighborhood character. However, when we consider changing racial composition by income, a slightly different dynamic emerges. That is to say, we would expect the pace of gentrification to increase with growing household income, but it continues to decline for Latinos, albeit at a slower rate, and more affluent Blacks do not advance gentrification by 2000 as commonly theorized, and may slow the pace of gentrification by 2010. More research is needed to further isolate working-class and middle-class Black and Latino households to better understand their ability to remain in place as predominately Black and Latino neighborhoods experience gentrification.

Few would deny that New York’s real-estate market, including the pace and longitudinal effects of gentrification, is an outlier relative to most American cities. Nevertheless, there is growing evidence of racial transition in gentrifying neighborhoods across the country (Hyra 2017; Sullivan and Shaw 2011). Yet most municipal leaders in the United States adopt the supposition of nonracialism, thus adopting universalism and nonracial approaches to planning and policy. As Thompson (2017) forcefully argues, nonracial approaches to planning and policymaking are surely more politically palpable, but adopting such an approach tends to maintain the status quo, which is itself a racial appeal. But “keeping race off the table is not innocent . . . it empowers some people while disempowering others” (Thompson 2017, p. 6). Gentrification is a hot button issue on its own, but the degree to which it entrenches a racial pattern on a city’s landscape, much like urban renewal and redlining, should be confronted and addressed.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biography

Stacey Sutton is an assistant professor of urban planning and policy in the College of Urban Planning and Public Affairs at the University of Illinois at Chicago. Her research interests are in alternative models for equitable development, economic democracy, neighborhood change, and disparate effects of place-based policy. Recent projects include an analysis of municipal support for worker-owned cooperatives, the efficacy of planning tools, and neighborhood retail trends in Chicago and New York City 1990–2014.

References

Adams

Michael Henry.

May 2016. “The End of Black Harlem.” The New York Times. Retrieved June 1, 2016 from https://www.nytimes.com/2016/05/29/opinion/sunday/the-end-of-black-harlem.html.

Anderson

Matthew B.

Sternberg

Carolina

. 2013. “Non-White Gentrification in Chicago’s Bronzeville and Pilsen: Racial Economy and the Intraurban Contingency of Urban Redevelopment.” Urban Affairs Review 49:435–67.

Atkinson

Rowland

. 2003. “Introduction: Misunderstood Savior or Vengeful Wrecker? The Many Meanings and Problems of Gentrification.” Urban Studies 40:2343–50.

Been

Vicki

Ellen

Ingrid

Gedal

Michael

Voicu.

Ioan

2008. “The Impact of Supportive Housing on Surrounding Neighborhoods: Evidence from New York City.” Furman Center for Real Estate and Urban Policy Working Paper Series Accessed on January 25, 2018 from http://furmancenter.org/files/publications/A3ImpactofSHonNeighborhoods_000.pdf.

Betancur

John J.

2002. “The Politics of Gentrification the Case of West Town in Chicago.” Urban Affairs Review 37:780–814.

Betancur

John J.

Smith

Janet L.

. 2016. Claiming Neighborhood: New Ways of Understanding Urban Change. Urbana: Univ. of Illinois Press.

Blundell

Richard

Bond

Stephen

. 1998. “Initial Conditions and Moment Restrictions in Dynamic Panel Data Models.” Journal of Econometrics 87:115–43.

Bostic

Raphael W.

Martin

Richard W.

. 2003. “Black Home-Owners as a Gentrifying Force? Neighbourhood Dynamics in the Context of Minority Home-Ownership.” Urban Studies 40:2427–49.

Boyd

Michelle R.

2008. Jim Crow Nostalgia: Reconstructing Race in Bronzeville. Minneapolis: Univ. of Minnesota Press.

10.

Brooks

Richard

Rose

Carol M.

. 2013. Saving the Neighborhood: Racially Restrictive Covenants, Law, and Social Norms. Cambridge: Harvard Univ. Press.

11.

Carroll

Rebecca

. 2016. “There Goes the Neighborhood: Race & Gentrification.” Take Away, WNYC Public Radio, May 10. Accessed April 4, 2018. http://www.wnyc.org/story/there-goes-neighborhood-explores-gentrification-brooklyn/.

12.

Chapple

Karen

Jacobus

Rick

. 2009. “Retail Trade as a Route to Neighborhood Revitalization” p. 19–68. In Urban and Regional Policy and Its Effects, edited by Edited by Weir, Pindus, Wial, and Wolman. Washington, DC: Brookings Institution Press.

13.

Conley

Dalton

. 2010. Being Black, Living in the Red: Race, Wealth, and Social Policy in America. Berkeley: Univ. of California Press.

14.

Curran

Winifred

. 2007. “‘From the Frying Pan to the Oven’: Gentrification and the Experience of Industrial Displacement in Williamsburg, Brooklyn.” Urban Studies 44:1427–40.

15.

Darity

William

Hamilton

Darrick

. 2012. “Bold Policies for Economic Justice.” The Review of Black Political Economy 39:79–85.

16.

Ellen

Ingrid Gould

. 2000. Sharing America’s Neighborhoods: The Prospects for Stable, Racial Integration. Cambridge: Harvard Univ. Press.

17.

Ellen

Ingrid Gould

O’Regan

Katherine

. 2011. “How Low-Income Neighborhoods Change: Entry, Exit, and Enhancement.” Regional Science & Urban Economics 41:89–97.

18.

Flores

Ronald J. O.

Lobo

Arun Peter

. 2012. “The Reassertion of a Black/Non-Black Color Line: The Rise in Integrated Neighborhoods Without Blacks in New York City, 1970–2010.” Journal of Urban Affairs 35:255–82.

19.

Freeman

Lance

. 2005. “Displacement or Succession? Residential Mobility in Gentrifying Neighborhoods.” Urban Affairs Review 40:463–91.

20.

Freeman

Lance

. 2006. There Goes the Hood: Views of Gentrification from the Ground Up. Philadelphia: Temple Univ. Press.

21.

Freeman

Lance

. 2009. “Neighbourhood Diversity, Metropolitan Segregation and Gentrification: What Are the Links in the US?” Urban Studies 46:2079–101.

22.

Freeman

Lance

Cai

Tiancheng

. 2015. “White Entry into Black Neighborhoods: Advent of a New Era?” The ANNALS of the American Academy of Political and Social Science 660:302–18.

23.

Glass

Ruth L.

1964. London: Aspects of Change. London, UK: MacGibbon & Kee.

24.

Goetz

Edward

. 2011. “Gentrification in Black and White: The Racial Impact of Public Housing Demolition in American Cities.” Urban Studies 48:1581–604.

25.

Gotham

Kevin

. 2000. “Urban Space, Restrictive Covenants and the Origins of Racial Residential Segregation in a US City, 1900–50.” International Journal of Urban and Regional Research 24:616–33.

26.

Hackworth

Jason

. 2002. “Postrecession Gentrification in New York City.” Urban Affairs Review 37:815–43.

27.

Hammel

Daniel J.

Wyly

Elvin K.

. 1996. “A Model for Identifying Gentrified Areas with Census Data.” Urban Geography 17:248–68.

28.

Hamnett

Chris

Whitelegg

Drew

. 2007. “Loft Conversion and Gentrification in London: From Industrial to Postindustrial Land Use.” Environment and Planning A: Economy and Space 39:106–24.

29.

Hwang

Jackelyn

Sampson

Robert J.

. 2014. “Divergent Pathways of Gentrification: Racial Inequality and the Social Order of Renewal in Chicago Neighborhoods.” American Sociological Review 79:726–51.

30.

Hyra

Derek S.

2008. The New Urban Renewal: The Economic Transformation of Harlem and Bronzeville. Chicago: The Univ. of Chicago Press.

31.

Hyra

Derek S.

2015. “The Back-to-the-City Movement: Neighbourhood Redevelopment and Processes of Political and Cultural Displacement.” Urban Studies 52:1753–73.

32.

Hyra

Derek S.

2017. Race, Class, and Politics in the Cappuccino City. Chicago: The Univ. of Chicago Press.

33.

Keele

Luke

Kelly

Nathan J.

. 2005. “Dynamic Models for Dynamic Theories: The Ins and Outs of Lagged Dependent Variables.” Political Analysis 14: 186-205.

34.

Keels

Micere

Burdick-Will

Julia

Keene

Sara

. 2013. “The Effects of Gentrification on Neighborhood Public Schools.” City & Community 12:238–59.

35.

Kirkland

Elizabeth

. 2008. “What’s Race Got to Do with it? Looking for the Racial Dimensions of Gentrification.” Western Journal of Black Studies 32:18–30.

36.

Kochhar

Rakesh

Fry

Richard

. 2014. Wealth Inequality Has Widened Along Racial, Ethnic Lines Since End of Great Recession. Washington, D.C.: Pew Research Center.

37.

Lees

Loretta

. 2000. “A Reappraisal of Gentrification: Towards a ‘Geography of Gentrification’” Progress in Human Geography 24:389–408.

38.

Lees

Loretta

. 2008. “Gentrification and Social Mixing: Towards an Inclusive Urban Renaissance?” Urban Studies 45:2449–70.

39.

Lees

Loretta

Slater

Tom

Wyly

Elvin K.

eds. 2010. The Gentrification Reader. London: Routledge.

40.

Ley

David

. 1994. “Gentrification and the Politics of the New Middle Class.” Environment and Planning D: Society and Space 12:53–74.

41.

Massey

Douglas S.

Denton

Nancy A.

. 1993. American Apartheid: Segregation and the Making of the Underclass. Cambridge: Harvard Univ. Press.

42.

McGee

Henry W.

1991. “Afro-American Resistance to Gentrification and the Demise of Integrationist Ideology in the United States.” The Urban Lawyer 23:25–44.

43.

McKinnish

Terra

Walsh

Randall

, White

T. Kirk

. 2010. “Who Gentrifies Low-Income Neighborhoods?” Journal of Urban Economics 67:180–93.

44.

Mele

Christopher

. 2000. Selling the Lower East Side: Culture, Real Estate, and Resistance in New York City. Minneapolis: Univ. of Minnesota Press.

45.

Meltzer

Rachel

Capperis

Sean

. 2017. “Neighbourhood Differences in Retail Turnover: Evidence from New York City.” Urban Studies 54:3022–57.

46.

Moore

Kesha S.

2009. “Gentrification in Black Face? The Return of the Black Middle Class to Urban Neighborhoods.” Urban Geography 30:118–42.

47.

Oliver

Melvin L.

Shapiro

Thomas M.

1993. Black Wealth, White Wealth: A New Perspective on Racial Inequality. New York, NY: Routledge Press.

48.

Omi

Howard

Winant

Michael

. 1994. Racial Formation in the United States: From the 1960s to the 1990s. New York, NY: Routledge

49.

Pattillo

Mary

. 2007. Black on the Block: The Politics of Race and Class in the City. Chicago: The Univ. of Chicago Press.

50.

Rosen

Andrew

. 2003. The Transformation of British Life 1950-2000: A Social History. Manchester, UK: Manchester University Press

51.

Satter

Beryl

. 2009. Family Properties: Race, Real Estate, and the Exploitation of Black Urban America. New York, NY: Macmillan Publishers.

52.

Schaffer

Richard

Smith

Neil

. 1986. “The Gentrification of Harlem?” Annals of the Association of American Geographers 76:347–65.

53.

Slater

Tom

. 2011. “Gentrification of the City”p. 571-585. In The New Blackwell Companion to the City, edited by Gary Bridge and Sophie Watson. Hoboken, NJ: John Wiley & Sons.

54.

Smith

Neil

. 1979. “Toward a Theory of Gentrification a Back to the City Movement by Capital, Not People.” Journal of the American Planning Association 45:538–48.

55.

Smith

Neil

. 1996. The New Urban Frontier: Gentrification and the Revanchist City. New York, NY: Routledge Press.

56.

Smith

Neil

. 1998. “Gentrification.” In The Encyclopedia of Housing, edited by Vliet

W. Van

, 198–99. London: Taylor & Francis.

57.

Sullivan

Daniel Monroe

Shaw

Samuel C.

. 2011. “Retail Gentrification and Race: The Case of Alberta Street in Portland, Oregon.” Urban Affairs Review 47:413–32.

58.

Sutton

Stacey A.

2010. “Rethinking Commercial Revitalization: A Neighborhood Small Business Perspective.” Economic Development Quarterly 24:352–71.

59.

Taylor

K. H.

2015. “Gentrification of Detroit Leaves Black Owned Business Behind.” NBC News, November 1. Accessed April 4, 2018. http://www.nbcnews.com/news/nbcblk/gentrification-detroit-leaves-black-residents-behind-n412476.

60.

Taylor

Monique M.

2002. Harlem: Between Heaven and Hell. Minneapolis: Univ. of Minnesota Press.

61.

Thompson

J. Phillip

. 2017. “Place Matters, and So Does Race.” Urban Affairs Review 53:210–18.

62.

Timberlake

Jeffrey M.

Johns-Wolfe

Elaina

. 2017. “Neighborhood Ethnoracial Composition and Gentrification in Chicago and New York, 1980 to 2010.” Urban Affairs Review 53:236–72.

63.

Wyly

Elvin K.

Hammel

Daniel J.

. 2004. “Gentrification, Segregation, and Discrimination in the American Urban System.” Environment and Planning A: Economy and Space 36 (7): 1215–41.

64.

Zukin

Sharon

. 1982. Loft Living: Culture and Capital in Urban Change. New Brunswick: Rutgers Univ. Press.

65.

Zukin

Sharon

. 1987. “Gentrification: Culture and Capital in the Urban Core.” Annual Review of Sociology 13:129–47.

66.

Zukin

Sharon

. 2011. Naked City: The Death and Life of Authentic Urban Places. Oxford: Oxford Univ. Press.

67.

Zukin

Sharon

Trujillo

Valerie

Frase

Peter

Jackson

Danielle

Recuber

Tim

Walker

Abraham

. 2009. “New Retail Capital and Neighborhood Change: Boutiques and Gentrification in New York City.” City & Community 8:47–64.

Gentrification and the Increasing Significance of Racial Transition in New York City 1970–2010

Abstract

Keywords

Introduction

Background and Theoretical Framework

Racial Transition, Neighborhood Diversity, Residential Mobility

Data, Measures, and Approach

Measuring Gentrification

Dependent variable

Independent variables

Analytic Approach

Findings

Gentrifying and Nongentrifying Neighborhoods

(H1) Racial Transition Increases Over the Study Period

Modeling Gentrification

(H2) The Share of Black and Latino Residents Declines With Gentrification

(H3) The Pace of Gentrification Dampened by Affluent Blacks and Latinos

Implications for Planning and Policy

Footnotes

Declaration of Conflicting Interests

Funding

Notes

Author Biography

References