No-Majority Communities: Racial Diversity and Change at the Local Level

Abstract

The United States is experiencing a profound increase in racial and ethnic diversity, although its communities are experiencing the trend differently depending on their size and location. Using census data from 1980–2010, we focus on a subset of highly diverse local jurisdictions in which no ethnoracial group makes up more than half of the population. We track the prevalence, emergence, and characteristics of these no-majority places, finding that they are rapidly increasing in number and are home to substantial and growing shares of the Black, Latino, and Asian populations. Transitions in no-majority places varied considerably over time. Older cohorts of places that became no-majority decades ago moved toward Latino or Black majorities while those in recent cohorts tended to persist as no-majority places. Most of these communities continued to diversify in the decades after first becoming no-majority and remain quite diverse today. However, the shift toward no-majority status was often accompanied by large White population declines.

In 2010 Angel Taveras was elected the first Hispanic mayor of Providence, Rhode Island. Taveras, the son of Dominican immigrants, prevailed during a crowded Democratic primary, breaking with a longstanding local tradition of Irish and Italian mayoral regimes (Filindra and Orr 2013). Hit hard by recession, Providence was ripe for political change, but demographic forces helped pave the way. Rapid Latino growth had altered the city’s racial and ethnic make-up so dramatically that, a decade before Taveras’s victory, White representation had fallen below one-half of all residents (down from four-fifths in 1980). At the time of his election, Providence was 38.1% Hispanic, 37.6% White, 13.1% Black, and 6.3% Asian, or what we refer to as a no-majority place. It retained that status in 2015 when Jorge Elorza, of Guatemalan origin, succeeded Taveras as mayor, benefitting from a similar multiethnic coalition of voters (Hill 2015).

The political transition in Providence offers a vivid example of the consequences associated with the shift underway toward greater ethnoracial diversity in the United States. Diversity—how evenly the members of a population are spread across racial-ethnic categories—is on the rise at both large and small spatial scales, as revealed by the emergence of majority-minority states (Franklin 2015; Wright et al. 2014) and multiethnic neighborhoods (Farrell and Lee 2011; Logan and Zhang 2010). Diversity gains have also been documented in the “middle”, not only for cities like Providence but for suburbs, small towns, and other types of communities (Hall et al. 2016; Parisi et al. 2015; Sharp and Lee 2016). We believe that such communities constitute especially salient settings in which to study diversity trends. Many of them have reputations that attract some racial and income groups and repel others, leading to increased residential sorting at the municipal rather than the neighborhood level (Farrell 2008; Fischer et al. 2004; Lichter et al. 2015). They are also jurisdictional sites of localized decision-making and service provision as well as social arenas in which concerns about diversity inspire vigorous debate.

As the Providence case suggests, much research on diversity is prompted by interest in its politically relevant outcomes. One line of survey-based inquiry finds that community and neighborhood ethnoracial diversity may reduce interpersonal trust, but the effect depends upon the targets or objects of trust (e.g., people in general, neighbors), whether respondents interact with those targets, and how diversity is measured (Abascal and Baldassarri 2015; Alesina and Ferrarra 2002; Costa and Kahn 2003; Gijsberts et al. 2012; Hou and Wu 2009; Putnam 2007; Stolle et al. 2008; Uslaner 2012). Similarly, hypothesized relationships between local diversity and spending on public goods (−), civic engagement (−), attitudes toward immigrants and minorities (−), and support for anti-immigrant policies (+) receive varied, contingent support (Alesina et al. 1999; Berg 2009; Costa and Kahn 2003; Gijsberts et al. 2012; Hopkins 2009, 2010; Laurence 2014; Portes and Vickstrom 2011; Tolbert and Hero 2001). Outside of the political realm, some evidence hints at the boundary-blurring implications of diversity. More diverse metropolitan areas, for example, exhibit higher rates of exogamy (intermarriage) and multiracial identification (Lee and Bean 2010). And residential segregation is lower in diverse community contexts, in part because diversity appears to diminish or buffer the impact of racial threat (indexed by minority population size or growth) on segregation (DeFina and Hannon 2009; Parisi et al. 2015; cf. Uslaner 2012).

This range of potential consequences has motivated us to examine a subset of communities with exceptionally high levels of diversity. No-majority places are communities in which no ethnoracial group makes up more than half of the local population. Because the presence of a majority group is a taken-for-granted aspect of community identity and the political landscape, places lacking a majority group tend to get overlooked. Thus, little is known about them. With respect to their prevalence, the demographic prominence of non-Hispanic Whites—who remain three-fifths of all Americans (Hixson et al. 2011)—hints that such places are rare although perhaps increasing in number. By definition, they are very diverse but not identically so. The racial-ethnic structure of no-majority places—i.e., the specific groups present—can differ as well. Thanks to rapid Hispanic and Asian growth, that structure need no longer be mainly Black and White. No-majority places also provide a strategic opportunity to assess the likelihood of stable diversity at a supra-neighborhood scale. Are these places as fragile as many racially mixed neighborhoods, where diversity has often been a temporary state during the transition from one form of homogeneity to another (Bader and Warkentien 2016; Friedman 2007)? Or do they show greater compositional persistence over time? A final unresolved question concerns the sources of no-majority status, which may or may not resemble the community attributes correlated with diversity in general (see, e.g., Allen and Turner 1989; Farrell 2005; Hall and Lee 2010; Lee et al. 2012).

We address these issues for approximately 10,000 incorporated and unincorporated places. A simple “majority rule” criterion allows us to distinguish those places in which no ethnoracial group exceeds half of the population from their White-, Black-, Latino-, Asian-, and Other-majority counterparts. Using decennial census and American Community Survey data, we document the prevalence, diversity magnitude, racial-ethnic structure, stability, and community correlates of no-majority places between 1980 and 2010. Our analysis is framed by two theoretical perspectives—spatial assimilation and ethnic stratification—that offer different predictions about the prospects for no-majority communities. Before these perspectives are developed, however, some additional background is in order regarding the diversification trend nationally, the concept of diversity, and how no-majority places fit into the bigger picture.

Background

Ethnoracial Diversification and No-Majority Places

According to recent projections, the United States is on track to become a majority-minority society by the mid-2040s (Colby and Ortman 2015). Although continued immigration flows from Latin America and Asia make the most visible contribution to diversification, other forces include group differences in natural increase, high rates of interracial coupling (and the subsequent multiracial offspring), and individual shifts in racial-ethnic identity (Hirschman 2005; Lichter 2013). Moreover, the youthful age structures of minority groups ensure that the diversity trend will continue well into the future even if immigration should slow or halt (Johnson and Lichter 2010). These dynamics are altering the composition of neighborhoods, communities, metropolitan areas, and states in similar fashion. Regardless of geographic scale, substantial minority gains and absolute or relative White declines have combined to erode the extent to which Whites dominate demographically, most notably in the South and West. Principal cities now exhibit complex ethnoracial mixes, with Hispanics often supplanting Blacks as the largest population of color, and fewer suburbs and rural towns look like the White bastions of popular stereotype (Berube 2003; Hall et al. 2016; Lichter 2012; Price and Singer 2008).

These kinds of communities, or “places” in Census Bureau parlance, represent the “legal, political, and economic containers that hold U.S. neighborhoods” (Lichter et al. 2015, p. 847). Their diversity can be conceptualized as having two dimensions: 1) the number of groups present in a place, and 2) the size of the groups in relation to each other (White 1986). Consistent with intuition, places which contain more ethnoracial groups of the same size will be very diverse while those inhabited by members of a single group will be very homogeneous (i.e., lacking diversity). Our interest centers on a particular type of place: one where no group comprises more than half of all residents. By definition, such no-majority places exhibit high diversity magnitude since they meet both of the criteria just noted. We also pay attention to their racial-ethnic structure, or the particular groups that live in them. This allows us to identify which ethnoracial groups are most likely to be experiencing high levels of community diversity as well as those for whom jurisdictional no-majority status often remains a remote and unlikely possibility.

The significance of no-majority places can be traced to the assumption that the high diversity embodied by them has consequences for their residents. A number of specific diversity-related consequences are mentioned in the introduction. More generally, no-majority communities may serve as sites of incorporation where residents of different backgrounds come into contact with one another and develop a cosmopolitan orientation (Anderson 2011). A potential downside of high diversity, however, is that it often generates resistance and conflict, especially in places that are changing rapidly (Wilson and Taub 2006). Despite firm indications that diversity and its concomitants benefit the country as a whole (Alba and Nee 2003; Bean and Stevens 2003; Card 2005; Peri 2012; Waters and Pineau 2015), the record strikes us as less clear regarding its social, economic, and political impacts at the local level (Fainstein 2005; Lichter 2013; Putnam 2007). Public perceptions remain divided as well. Hence, it is not unusual to see a community’s unease over high or rising diversity manifested as racialized discourse about schools, zoning, crime, law enforcement, taxes, and immigration.

As we show below, no-majority places are also significant because of their increasing number during the last few decades and the parallel gains in population share residing in them. Many of our biggest cities currently are or once were no-majority places. Some, like New York, have been no-majority for decades. Others, like Cleveland, attained no-majority status briefly while shifting from a White to Black majority. Still others have recently joined the no-majority club (Las Vegas) just as other members appear poised to leave (Los Angeles). These no-majority places are created by many of the same forces driving diversity across the nation. Whether their no-majority status proves temporary or enduring, and why, remain to be determined.

Theoretical Approaches

We are guided by two theoretical approaches—spatial assimilation and ethnic stratification—when examining the prevalence, diversity, emergence, and persistence of no-majority communities. The spatial assimilation perspective has shed light on various aspects of the residential distribution of ethnoracial and nativity groups (Alba and Logan 1991; Iceland 2009; Rosenbaum and Friedman 2007; South et al. 2008). It is commonly used to address neighborhood segregation, locational attainment, and residential mobility patterns, but here we apply it to community diversity. This perspective presumes that residential patterns will reflect the economic and cultural characteristics of racial-ethnic groups. As groups converge in such characteristics, they become more geographically dispersed; this dispersion culminates in shared residential spaces like no-majority places. Ideational shifts (e.g., improved racial attitudes), policy interventions (e.g., anti-discrimination legislation), human capital advancements (e.g., higher education), and economic opportunities (e.g., diversified local labor markets) have helped advance the process of convergence (Alba and Nee 2003; Logan et al. 2004; Waters and Pineau 2015; White and Glick 2009).

What are the implications of spatial assimilation at the community level as it pertains to no-majority communities? First, the spatial assimilation perspective predicts that these communities will become more common, increasing in number across regional, metropolitan, micropolitan, and rural contexts. Second, they will persist over time, retaining high levels of diversity rather than rapidly turning over to a dominant group majority. Third, this persistent diversity will continue to include a substantial White presence; that is, increasing diversity at the local level will not spur White flight across jurisdictional lines.

The ethnic stratification perspective offers an alternative view of the prospects for no-majority communities, emphasizing the various forces aligned against the development and persistence of such communities. Despite changing racial attitudes, for example, members of various ethnoracial groups tend to express preferences for local residential contexts that approach or exceed own-group majorities (Charles 2007; Clark 2002, 2009). This is particularly the case for Whites. Factorial experiments reveal that they rank potential neighborhoods less attractive as Black and Latino representation increases, even when the areas are purported to have high-quality public schools, low crime rates, and increasing property values (Lewis et al. 2011). Asking White respondents about the desirability of actual communities rather than hypothetical neighborhoods, Krysan (2002) finds that predominantly White communities are more desirable than racially mixed ones regardless of whether the latter are more prosperous than the former (also see Bader and Krysan 2015). These preferences can translate into behavior, as Whites tend to exit diverse neighborhoods, especially ones experiencing rapid growth in foreign-born populations (Crowder et al. 2011). Germane to our study, White flight can also be prompted by extralocal racial changes in nearby neighborhoods, underscoring the need to consider diversity dynamics at larger geographic scales (Crowder and South 2008). Indeed, Frey’s (1996) balkanization thesis suggests that Whites may flee metropolitan areas if they feel threatened by minority growth.

In addition to racial preferences and mobility, there are institutional practices that make homogeneity the rule in many municipalities. These include density zoning and other land use controls, which limit construction of types of housing (e.g., multi-unit apartments) more affordable to a diverse array of homeseekers (Pendall 2000; Rothwell and Massey 2009). They also encompass state policies that incentivize the siting of low-income housing in racially segregated communities (Texas Dept. of Housing and Community Affairs v. Inclusive Communities Project, Inc. 2015). Furthermore, subtle forms of housing discrimination on the part of real estate agents and rental property owners continue to steer White and minority homeseekers in different directions, ensuring that most White residents are exposed to token levels of diversity at the local level (Turner et al. 2013).

In comparison to spatial assimilation, the ethnic stratification perspective is less sanguine about the long-term fate of no-majority communities. While acknowledging the increasing emergence of these diverse places, it predicts that their no-majority status will not persist indefinitely. Instead, the absence of a majority group is considered an intermediate way station during the transition toward a more homogeneous (i.e., group-majority) ethnoracial composition. This could take the form of rapid minority (probably Hispanic) population growth via immigration and subsequent enclave consolidation. The process is exacerbated by the constrained housing choices of many minority households due to the institutional practices summarized above. When combined with group-specific (often White) population losses via flight or avoidance, the end state in this scenario would be a homogeneous community that was only temporarily diverse. Any persistent no-majority communities will likely feature multiple minority groups but be largely absent Whites, who have exited as minority/immigrant population shares exceed White tolerance for localized diversity.

The spatial assimilation and ethnic stratification perspectives have less to say regarding the correlates of community diversity. Previous research on community diversity (Allen and Turner 1989; Farrell 2005; Hwang and Murdock 1998; Lee et al. 2012, 2014; Hall and Lee 2010) hints at a few community characteristics that should be conducive (or not) to no-majority status. One is scale: larger places and those in metropolitan settings have greater visibility and thus are able to attract more migrants than smaller or rural communities. By virtue of their scale, such places are also more likely to have a full complement of service-providing organizations, which may be a draw to minorities. For immigrant-rich groups, the context of reception matters as well. Communities located in regions closer to countries of origin (i.e., in the West or South) and with a higher percentage of foreign-born inhabitants should better facilitate access to a range of co-ethnic resources (employment niches, social capital, etc.) and thus function as an immigrant magnet (Portes and Rumbaut 1996). Overall, these characteristics are predicted to promote a more balanced racial-ethnic structure in which no group achieves a numerical majority.

By contrast, another aspect of the context of reception—a large retirement-age population—may be less conducive to localized diversity. This could reflect simple demographics, with older age cohorts less diverse than youthful ones (Frey 2015). It might also be due to residential preferences, as older members of the population tend to express more discomfort with immigration and the nation’s changing racial-ethnic composition (Pew Research Center 2011). Symbolically, a substantial share of senior citizens aging in place could signal a stagnant economy, deterring settlement by people of color in search of work. Whatever the mechanisms involved, we anticipate that the higher the percentage of residents 65 years of age and over, the slimmer their chances are of living in a no-majority community.

Opportunity, broadly defined, should be positively related to diversity and no-majority status. On the economic front, income parity (or inequality) among ethnoracial groups constitutes one indicator of the local opportunity structure, with a higher ratio of minority to White income associated with greater diversity. Local labor market characteristics are also salient. One might expect, for example, greater diversity in communities with many construction and transportation jobs given that these two industries account for the largest shares of Latino and Black workers, respectively (U.S. Bureau of Labor Statistics 2015). Housing represents another dimension of opportunity. In particular, communities with looser markets—more rental units, vacancies, and new housing—could be attractive to a variety of racial and ethnic groups.

Finally, we hypothesize that the presence of one or more institutional hubs increases the likelihood of a place having a highly diverse, no-majority population. State capitals and county seats, base communities, and college towns are anchored by institutions—government, the military, and higher education, respectively—that express a commitment to equal opportunity and provide minorities with employment or other avenues of upward mobility. Correctional facilities, which have proliferated in nonmetropolitan settings (Glasmeier and Farrigan 2007), constitute a less voluntary kind of hub. However, their location inside city or town limits should, like the other institutions of interest, contribute to a more complex ethnoracial composition in their host communities.

Methods

Data

We rely on 1980–2010 U.S. Census summary files and the 2008–2012 estimates from the American Community Survey. Our scope is national, and places serve as our community units of analysis. Places are geographic units that include incorporated cities, suburbs, towns, and villages as well as their unincorporated equivalents, census designated places (CDPs). Incorporated municipalities contain three of every five residents in the United States and, in aggregate, they are growing more rapidly than the nation as a whole (Cohen 2015). CDPs have no local government but are “settled population centers with a definite residential core, a relatively high population density, and a degree of local identity” (U.S. Census Bureau 1994, p. 9–20). In short, they are recognized communities, much as incorporated places are.

In 2010 the Census Bureau identified roughly 29,000 places in the United States, although many of them are quite small in size. We therefore implemented a population threshold for the analysis, excluding all places with less than 1,000 people at any of the four time points.¹ This left us with a subset of 10,815 places that account for 83% of the total 2010 population found in census places.² These places range in size from the minimum threshold of 1,000 (Franklin, NE and White Deer, TX) to over eight million (New York) residents, with a 2010 median and mean population of 4,630 and 17,492, respectively. We allow place boundaries to vary from decade to decade rather than holding them constant. This falls in line with our emphasis on communities as jurisdictional domains. Annexation of adjacent land, for example, reflects municipal decision-making that can affect housing supply, provision of services, and budgetary concerns for the entire community. It also has implications for changes in population size, distribution, and composition (e.g., ethnoracial diversity) at the community level.

Measuring Ethnoracial Composition

We use census cross-tabulations of race by Hispanic origin to create five mutually exclusive panethnic groups for each census year. The three largest groups are non-Hispanic Whites and Blacks along with Hispanics or Latinos of any race.³ Because they are aggregated in summary files prior to 2000, we combine Asian and Pacific Islander populations into a fourth group (hereafter Asian). The fifth group, much smaller than the others, includes a combination of respondents identifying as “other race,” American Indian/Alaska Native, or more than one race (reported from 2000 onward). Of the 189 million residents included in our 10,815 places, 57% are White, 15% Black, 20% Hispanic/Latino, 6% Asian, and 3% fall in the “Other” category as of 2010.

We apply a “majority rule” criterion to all places in each decade, classifying them as either no-majority, White-majority, Black-majority, Latino-majority, Asian-majority, or Other-majority. Given our focus on no-majority communities and the temporal scope involved, we further delineate these places according to the year in which they first became no-majority (if they did so). This quasi-cohort approach (Lee and Hughes 2015) allows us to track changes in communities that emerged as no-majority in different decades during the 1980–2010 time period.⁴

In addition to classifying places according to their no-majority/group-majority status, we measure the magnitude of diversity using the entropy index (White 1986). This index takes into account the number of groups present in a community and their size relative to one another. It is formally defined as:

$$E=\sum _{r=1}^n{Q}_r\text{ln}\frac{1}{Q_r},$$

where Q_r refers to group r’s proportion of a community population. In this formulation, the maximum value of E is equal to the natural log of the total number of groups, or 1.609 in the case of our five panethnic groups. For ease of interpretation, we divide each community diversity score by this maximum and multiply by 100, resulting in a range of possible values from 0 to 100. A diversity score of zero occurs when only one group is present. There were nine such communities in 1980 and 1990, all of which had all-White populations. The maximum score of 100 would occur in the event of a community having identical representation (20%) of each of the five groups. None of our communities reaches the maximum in any of the decennial years, though Suisun City, CA came closest in 2010 with an E score of 95.4.

It is important to note that different racial-ethnic structures can yield the same magnitude of diversity. For example, in 2010 both Brooklyn Park, MN and Duarte, CA had identical E scores of 79.6. Brooklyn Park is a White-majority (50%) community with sizable Black (24%) and Asian (15%) populations. Duarte is a no-majority community characterized by a Hispanic plurality (48%) and substantial White (27%) and Asian (16%) representation. Below we find that no-majority communities account for the lion’s share of high-diversity communities, though a number of group-majority communities also have highly diverse racial structures.

Community Characteristics

We examine a number of community characteristics that may play a role in local diversity.⁵ These characteristics represent four broad conceptual dimensions: scale, context of reception, opportunity, and institutional hub. In addition to population size, we measure scale by identifying whether a community was metropolitan, micropolitan, or rural in character in 2010. Metropolitan areas include at least one densely urbanized population of 50,000 or more combined with the remainder of the core county and any outlying counties (or their equivalents) that are economically tied to the core via commuting patterns. Micropolitan areas are a smaller version of metro areas, consisting of a core urban cluster of 10,000–50,000 residents along with contiguous commuter counties (Frey et al. 2006). We classify communities as “rural” if in 2010 they were located in a county which was not part of a metropolitan or micropolitan area.

To capture the context of reception, we include indicators of region, immigrant presence, and retirement-age population. For region, we rely on the traditional census-based Northeast-Midwest-South-West configuration. Because geography emerges as such an important correlate of community diversity, we present many of the descriptive results below subdivided by region. In addition to region, we measure the context of reception with a place’s percentage foreign-born and the percentage of the local population aged 65 years or over.

Opportunity, both economic and residential, is operationalized with a series of income, industry, and housing variables. Regarding income, we calculate a minority/White income ratio by dividing the average income of nonwhite households by that of White households.⁶ The relative size of the local construction and transportation sectors is captured by calculating the percentage of employed persons aged 16 years and over working in these industries. We use three variables to measure housing opportunities: the percentage of occupied units that are rental units, the vacancy rate among all housing units, and the percentage of units that were built within the previous decade.

Our last set of variables helps pinpoint communities that serve as institutional hubs for government, military, higher education, or corrections. Government hubs are identified by the percentage of employed workers who have federal, state, or local government jobs. Communities with military bases are captured by calculating the percentage of persons in the labor force enlisted in the Armed Forces, and university-oriented towns are reflected in the percentage of the school-age population enrolled in college. Finally, correctional hubs are measured with the percentage of the total population that is in a correctional facility.⁷

Results

A Growing Presence

Our first aim is to establish the number of no-majority communities in each decade. Overall, 724 of these highly diverse places existed in 2010, nearly six times more than in 1980 (125). As revealed in Figure 1, the emergence of no-majority communities was sharply uneven across area types and regions. Metropolitan areas saw a large gain in the number of these communities, increasing from 90 in 1980 to 542 in 2010. The micropolitan and rural increases were less dramatic; nevertheless, when combined they now account for one-quarter of all no-majority places. Regionally, such places were nearly nonexistent in northeastern and midwestern states in 1980 but there have been large gains since then. For example, a tenfold increase occurred in the Midwest—from 5 in 1980 to 50 in 2010—though in absolute terms the 2010 midwestern count fails to exceed the number found in either the southern or western region three decades prior. Overall, this trend toward more no-majority communities tracks with the predictions of the spatial assimilation perspective. However, the trend has been selective, disproportionately concentrated in the metropolitan areas of the nation’s southern and western states.

Figure 1.

No-Majority Places by Type and Region, 1980 and 2010

Despite these gains, no-majority communities remain the exception rather than the rule. In 2010, just one in ten metropolitan places lacked a majority racial-ethnic group, and the figure is closer to one in 20 for micropolitan and rural communities. This is particularly the case among northeastern and midwestern communities, of which less than 5% were no-majority in 2010. However, the modest number of no-majority places understates the changing share of the U.S. population residing in them. The percentage of Americans found in these places quadrupled from 4.1% in 1980 to 17.8% in 2010. Substantial shares of the total Black (29.7%), Latino (32.7%), and Asian (37.6%) populations were living in no-majority settings by 2010, in contrast to one out of ten Whites (10.1%).

Table 1 documents the regional patterns underlying the broader national trends. All four regions experienced growth in the share of the total population residing in no-majority communities between 1980 and 2010. The greatest gains are apparent in the Northeast and West, where over one-quarter of all residents now live in these diverse communities. The Midwest experienced modest increases and its no-majority shares of midwestern Hispanic, Asian, and Other populations actually declined. The South accounts for the largest number of no-majority places but the population share residing in them is half that found in the Northeast and West. The regional differences are particularly stark with regard to the share of the White population located in no-majority communities. Less than 4% of midwestern Whites lived in no-majority communities in 2010, representing only a slight increase over three decades. In the South, the increasing White share was larger, although in 2010 it was still roughly half that of the nation. By contrast, the northeastern share jumped by 12 percentage points and the western share nearly tripled. Regardless of region, people of color are far more likely to be living in a no-majority community than are White residents. In general, the increasing population shares found in no-majority places are consistent with spatial assimilation, but the large White/nonwhite gaps in these shares better align with ethnic stratification.

Table 1.

Group Proportions Residing in No-Majority Places by Region, 1980–2010

	Northeast			Midwest
	1980	2010	Diff	1980	2010	Diff
Total (%)	1.7	26.5	24.8	5.2	7.9	2.7
White	0.9	13.2	12.3	2.6	3.7	1.1
Black	6.1	62.9	56.8	23.0	25.2	2.2
Latino	5.6	56.7	51.1	34.5	27.7	−6.8
Asian	2.7	47.1	44.4	17.7	14.3	−3.4
Other	3.2	34.5	31.3	7.2	7.1	−0.1
	South			West
	1980	2010	Diff.	1980	2010	Diff.
Total (%)	1.1	13.5	12.4	10.7	27.2	16.5
White	0.7	8.6	7.9	6.6	18.4	11.8
Black	1.7	19.1	17.4	37.8	48.5	10.7
Latino	3.5	23.2	19.7	20.4	34.1	13.7
Asian	1.3	23.9	22.6	17.6	45.7	28.1
Other	0.9	12.6	11.7	8.3	24.9	16.6

Diversity Magnitude and Structure

How diverse are no-majority communities, relatively speaking? We explore the magnitude of community diversity in Figure 2, where box plots summarize entropy-based diversity scores for no-majority and group-majority communities in 2010. As a reminder, the 5-group entropy index yields values ranging from zero to 100, with higher values representing the presence of multiple panethnic groups of comparable size. With a median diversity score of 73, no-majority communities tend to be a great deal more diverse than their group-majority counterparts. The median no-majority place scores 43 points higher than the typical White-majority place and more than 20 points higher than the median Black- or Latino-majority place.

Figure 2.

Box Plots of Diversity Magnitude in No-Majority and Group-Majority Places, 2010

Nevertheless, variation in diversity magnitude can be observed across the group-majority categories. A number of group-majority communities—situated in the top “whisker” of the box plots—achieve high levels of diversity. For example, Malden, MA (80.3), Pontiac, MI (74.4), and Bellflower, CA (79.3) all have 2010 diversity scores above the no-majority median despite their respective White, Black, and Latino majorities. Asian-majority places, though few in number, are even more diverse than their group-majority counterparts. At the same time, cases exist in which no-majority communities exhibit somewhat modest levels of diversity. Ulysses, KS (49.8) and Jal, NM (49.3) are no-majority communities with diversity scores slightly below 50. Both have White and Hispanic populations that are just short of majorities combined with a minuscule presence of other groups. On the whole, however, no-majority communities dominate the upper echelons of diversity. In 2010 all of the 100 most diverse places in the United States were no-majority communities.

While the entropy index allows us to assess the magnitude of diversity, it does not provide information about racial and ethnic structure. In other words, which groups are represented in these diverse places? Figure 3 presents the 2010 mean ethnoracial mix of no-majority communities broken out by area type and region. Because the means are weighted by place population size, they represent the local racial-ethnic composition for the average community resident. The diversity scores for the weighted means appear in parentheses next to the type and regional labels.

Figure 3.

Weighted Mean Racial Composition of No-Majority Places by Area Type and Region, 2010

The average no-majority place has a White plurality across all area types and three of the four regions. So, while most Whites do not live in these communities, they do tend to comprise a sizable share of the population. This conforms broadly to spatial assimilation predictions about Whites residing in the same jurisdictional spaces as people of color. The largest White shares— above 40%—are found in micropolitan and rural no-majority places. The second largest segment of the no-majority population in these communities is Black, ranging from 24 to 32% of the population. This stands in contrast to metropolitan communities, where Hispanics make up 30% of the population (compared to an average Black share of 20%). Regional distinctions in nomajority composition are evident, with midwestern and southern communities evincing a predominantly White-Black-Latino structure while western communities include a greater Asian representation. The West is the only region in which Hispanic residents achieve a slight plurality over their White counterparts. Although diversity levels are high across regions, northeastern nomajority places are the most diverse on average.

Persistence or Change?

We next assess the trajectories of change for no-majority communities. We present these results in Table 2 for the 1980, 1990, and 2000 no-majority cohorts by tracking their respective transitions from first decade of no-majority status to their 2010 racial-ethnic structure. We also track the mean diversity levels and mean percentage White for each of these cohorts, providing data for the cohort inception year (T1) and 2010. The top panel in the table shows transitions for all no-majority places. Retention of no-majority status was the most common destination for the two most recent cohorts, but less than one-third of the 1980 cohort remained no-majority three decades later. These 1980 no-majority communities were more likely to transition to a Latino or Black majority even though their mean diversity scores dipped by only two points in the ensuing period. The 1990 and 2000 cohorts became even more diverse over time, although their increasing diversity occurred alongside rapid declines in White representation. By 2010, for example, the mean percentage White in the 1980 and 1990 cohorts had dropped by approximately 20 percentage points. Thus, the mean White pluralities presented in Figure 3 are due largely to the disproportionate influence of the 402 most recent no-majority communities which are not included in Table 2.⁸

Table 2.

Ethnoracial Transitions in No-Majority Places by Cohort, Area Type, and Region

	2010 Racial/Ethnic Structure					Mean Diversity		Mean % White		N of Places
	No Maj.	White Maj.	Latino Maj.	Black Maj.	Asian Maj.	T1	2010	T1	2010
All Places
1980	31.2	4.8	36.0	24.8	3.2	62.3	60.3	44.6	24.2	125
1990	44.0	2.0	33.3	16.7	4.0	64.0	66.3	45.5	27.6	150
2000	61.6	2.8	17.9	15.3	2.0	68.5	70.0	42.9	34.5	352
Metropolitan
1980	33.3	4.4	40.0	17.8	4.4	64.2	60.8	43.8	21.2	90
1990	44.1	0.9	35.1	14.4	5.4	67.5	69.0	44.7	24.2	111
2000	62.1	2.3	20.7	12.3	2.7	70.8	71.8	42.2	32.7	261
Micropolitan
1980	35.0	10.0	10.0	45.0	0.0	59.6	62.5	45.6	32.0	20
1990	57.1	7.1	28.6	7.1	0.0	58.0	64.9	47.4	37.2	14
2000	71.2	1.9	7.7	19.2	0.0	64.0	67.2	43.1	38.3	52
Rural
1980	13.3	0.0	46.7	40.0	0.0	54.4	54.4	47.9	31.9	15
1990	36.0	4.0	28.0	32.0	0.0	51.9	55.2	48.3	37.5	25
2000	46.2	7.7	12.8	30.8	0.0	58.9	62.1	47.3	41.6	39
Northeast
1980	64.3	7.1	14.3	14.3	0.0	64.3	74.8	44.3	23.5	14
1990	60.0	0.0	25.0	15.0	0.0	67.7	73.3	45.4	24.9	20
2000	68.6	2.9	22.9	2.9	2.9	73.5	74.3	41.8	30.9	35
Midwest
1980	20.0	0.0	20.0	60.0	0.0	58.7	53.3	43.2	17.1	5
1990	37.5	0.0	0.0	62.5	0.0	58.1	63.9	48.4	30.2	8
2000	50.0	0.0	25.0	25.0	0.0	65.8	66.7	45.0	32.8	20
South
1980	17.9	7.1	28.6	46.4	0.0	56.6	56.2	46.8	29.4	56
1990	44.3	4.3	27.1	24.3	0.0	59.0	62.9	47.0	32.1	70
2000	59.7	4.7	9.9	25.1	0.0	64.9	67.4	45.0	37.2	191
West
1980	38.0	2.0	52.0	0.0	8.0	68.4	61.5	42.3	19.3	50
1990	38.5	0.0	50.0	0.0	11.5	70.3	68.6	43.1	22.3	52
2000	65.1	0.0	29.2	0.0	5.7	73.9	73.9	39.0	31.1	106

The next three panels of Table 2 document cohort transitions by area type. The metropolitan cohorts largely mirror the overall results given that they contain most of the places. The more recent metro cohorts tended to retain no-majority status and the 1980 cohort was most likely to transition to a Latino majority. Regardless of cohort, metro places lost White population shares at a higher rate than the other area types. There are fewer no-majority cases among the micropolitan and rural cohorts but some interesting patterns emerge. Micropolitan places were more likely to hold on to their no-majority status than those in metropolitan areas, and the remainder were more likely to shift to Black rather than Latino majorities. Rural no-majority communities were least likely among the three area types to retain no-majority status, shifting instead toward Latino or Black majorities. However, they retained larger White shares on average. In all, metro no-majority communities were the most diverse despite the diversity declines experienced by their oldest cohort.

We have already identified large regional differences in the prevalence of no-majority communities, and the bottom four panels of Table 2 investigate whether no-majority transitions differ by region as well. In general, midwestern and southern no-majority places were less diverse and less likely to remain no-majority over time. This was especially the case in the midwestern cohorts, the older of which were far more likely to shift toward Black majorities than retain no-majority status. (Note, however, their very small Ns.) By contrast, the Northeast is distinct for the durability of its no-majority places over both the short and long term. This persistence in northeastern no-majority status did not necessarily translate into stable racial composition, as White shares declined markedly even as diversity increased. Western no-majority communities experienced diversity declines, particularly the 1980 cohort. Over half of this cohort shifted toward a Latino majority, though it remained quite diverse in 2010. White representation decreased across all four regions, most markedly in the older cohorts of the West.

To sum up, no-majority transitions varied by region and area type but some of the most distinctive differences were by cohort. Older cohorts were more likely to transition to either a Hispanic or Black majority than to remain no-majority. This runs counter to the spatial assimilation perspective, which predicts the persistence of no-majority structures over time. Despite shifts in racial-ethnic structure, however, these communities tended to remain very diverse in 2010. In fact, most of the cohorts continued to diversify during the ensuing decades and any mean diversity declines were modest. For example, the 1980 metro cohort experienced diversity decreases on the order of one point per decade. This does not comport with the rapid pace of turnover predicted by the ethnic stratification perspective.

The stratification perspective does receive support when looking specifically at the changes in White population in no-majority cohorts. White population shares diminished rapidly in most no-majority places but especially those found in metropolitan areas. To get a sense of the magnitude of these shifts, we present White population changes in no-majority cohorts in Table 3. Overall, the White population in no-majority communities declined by 7.1 million between 1980 and 2010. Meanwhile, White-majority communities underwent an aggregate increase of 8.6 million during the same period.⁹ For the older no-majority cohorts, average White population losses have been the rule since the 1980s. The more recent 2000 and 2010 no-majority cohorts experienced mean White gains in the 1980s prior to declines in the 1990s and 2000s. These same general patterns hold across regions (not shown), although the western 2010 cohort of nomajority places stands out as an exception because of its net White increase (+439,000) and positive White growth rate (+3.6%) during the 2000s.

Table 3.

White Population Change in No-Majority Cohorts and White-Majority Communities

	Mean % change			Total change	N of places
All Places	1980s	1990s	2000s	1980–2010
NM cohorts
All	+2.7	−13.4	−11.6	−7.1 mil	1,029
1980	−7.3	−21.0	−12.3	−1.4 mil	125
1990	−7.7	−19.3	−16.9	−2.7 mil	150
2000	+2.7	−16.6	−13.0	−2.4 mil	352
2010	+9.6	−6.1	−8.1	−0.6 mil	402
W-majority	+7.5	+7.7	+2.8	+8.6 mil	8,930

Correlates of No-Majority Status

Our final set of analyses explores the correlates of no-majority status and emergence across U.S. communities. Table 4 presents the results of a logistic regression of no-majority status on community characteristics.¹⁰ The data are structured in panel format with dummy variables included for each decade (1980 as the reference). According to the coefficient for population size, larger communities are more likely to lack a majority group regardless of area type. Consistent with the regional differences outlined above, northeastern and midwestern communities are less likely than western communities to have no majority and southern locations are most likely to lack a majority group.

Table 4.

Logistic Regression of No-Majority Status on Community Characteristics

	Coef.	SE
Year
2010	1.03***	.11
2000	.82***	.11
1990	.41***	.11
1980 (ref.)
Population size	.34***	.04
Area Type
Metropolitan	−.22	.14
Micropolitan	−.03	.16
Rural (ref.)
Region
Northeast	−.55***	.13
Midwest	−.90***	.16
South	.54***	.09
West (ref.)
% foreign-born	.87***	.05
% retirement age	−.23***	.06
Minority/white inc.	−.06	.03
% construction	.13*	.06
% transportation	.22***	.06
% renters	.82***	.13
% vacant	.15*	.07
% new units	−.21***	.03
% government	.71***	.10
% military	−.04	.02
% college	−.30***	.06
% incarcerated	.01	.02
Intercept	−2.73***	.70
Pseudo R2	.24
N	34,096

^*p<.05

^**p<.01

^***p<.001

Note: Standard errors are adjusted for clustered observations. All continuous predictors are logarithmically transformed.

Two other context-of-reception variables prove statistically significant, as immigrant destinations tend not to have a local majority group while retirement destinations do. Among our measures of economic and residential opportunity, both construction and transportation employment are related to no-majority status. So are high rates of rental housing, vacant housing, and older housing. A pair of the institutional hub measures also has strong effects: robust government employment and low college enrollment are significantly associated with nomajority status. The latter effect—contrary to our expectations—may simply reflect the fact that Whites have dominated the ranks of college attendees, limiting the extent to which college towns can diversify.

Returning to our cohort-based analyses, how are racial composition and change implicated in the emergence of no-majority communities? We address this issue with a series of logistic regressions summarized in Table 5. Across the three models, the dependent variable is a dummy measure that identifies emergent no-majority cohorts in 1990, 2000, and 2010, respectively. Our focus is on how initial racial structure and subsequent racial change influence the likelihood of a community becoming no-majority during each decade. To capture racial structure, we include a dummy variable identifying whether a community had a White majority at the beginning of the decade. Decadal White, Black, Latino, and Asian growth rates are used to assess the impact of race-specific population dynamics. Since annexation could potentially alter the racial composition of communities, we also construct an areal ratio variable in which each place’s land area (square miles) in the most recent census year is divided by its land area in the previous census year. Ratio values higher than one reflect an increase in land area consistent with annexation. Finally, the models include all of the community controls (measured at the beginning of each decade) that appear in Table 4.¹¹

Table 5.

Logistic Regressions of Emerging No-Majority Status on Racial Structure and Change

	1980-1990		1990-2000		2000-2010
	Coef.	SE	Coef.	SE	Coef.	SE
White-majority	2.67***	.64	2.33***	.41	2.40***	.36
Non-white maj. (ref.)
White growth rate	−2.03***	.49	−5.16***	.48	−4.11***	.51
Black growth rate	.08	.05	.08	.06	.10	.12
Latino growth rate	.85**	.28	.71***	.11	1.29***	.18
Asian growth rate	−.03	.09	−.03	.07	.46*	.21
Land area ratio	.32	.29	1.45***	.25	1.05***	.24
Population size	.55***	.10	.41***	.07	.42***	.06
Area Type
Metropolitan	−.88*	.36	−.41	.26	−.30	.25
Micropolitan	−1.23**	.47	−.18	.26	−.01	.26
Rural (ref.)
Region
Northeast	−1.23***	.30	−.76**	.23	−.72***	.20
Midwest	−1.60***	.39	−1.41***	.31	−.89***	.23
South	.44*	.22	.31	.17	.32*	.15
West (ref.)
% foreign-born	.95***	.18	.74***	.10	1.14***	.11
% retirement age	−.25	.14	−.09	.11	−.24*	.11
Minority/white inc.	.24***	.07	−.42***	.10	−.67***	.15
% construction	.05	.19	.51**	.19	.44**	.14
% transportation	.37	.20	.14	.12	.42***	.13
% renters	.64*	.26	.42	.22	.60***	.16
% vacant	.22	.18	.10	.13	−.03	.11
% new units	−.11	.12	.12	.11	−.11	.06
% government	.91***	.24	.56**	.19	.83***	.19
% military	−.08	.06	.00	.04	−.07	.04
% college	−.43**	.13	−.27	.14	−.33*	.14
% incarcerated	.09	.06	.08*	.04	.40***	.10
Intercept	−5.46*	2.22	−3.52*	1.51	−.86	1.58
Pseudo R2	.26		.30		.27
N	6,439		7,061		7,539

^*p<.05

^**p<.01

^***p<.001

Note: Standard errors are adjusted for clustered observations. All continuous predictors are logarithmically transformed.

Communities with White-majority populations were more likely to relinquish their majorities than were other places, as revealed by the significant positive effects for the White-majority dummy. Thus, the White pluralities that typify no-majority communities (see Figure 3) are in most cases what is left of a recent White majority. The significant negative effects for White population growth in all three decades suggest that no-majority communities are most likely to emerge in the wake of White declines (or more modest White growth vis-à-vis other groups). Population changes for the other groups vary in terms of impact. Black population growth did not have a significant effect while Hispanic growth contributed to the emergence of new no-majority communities in every decade. Asian population dynamics played no significant role in earlier decades but in the 2000s helped usher in the largest cohort of 2010 no-majority places. These effects occur net of the control for annexation, indicating that the emergence of no-majority communities cannot be ascribed to boundary changes alone. On the whole, the results are consistent with the ethnic stratification perspective, which predicts that diversifying places are less likely to retain White populations.

The remaining coefficients in Table 5 indicate that larger cities outside the Northeast and Midwest were most likely to become no-majority in the 1980s, 1990s, and 2000s. A large immigrant presence also heightens the likelihood of emerging no-majority status in every decade. Communities with substantial construction employment sectors and available rental housing tended to shift to a no-majority status, as did government and correctional hubs. University towns were less likely to lose their (typically White) majorities in two of the three decades. Interestingly, racial income parity was associated with the emergence of no-majority communities in the 1980s but this effect was reversed in subsequent decades.¹²

Conclusion

By mid-century the trend toward greater ethnoracial diversity will transform the United States into a no-majority nation. As illustrated above, a number of American communities have already reached this threshold and in the process have taken on much greater demographic significance. Not only are no-majority places increasing in frequency but by 2010 they housed nearly 55 million residents, including substantial shares of the Black, Hispanic, and Asian populations. They are concentrated in metropolitan areas in the western and southern U.S. yet are becoming more prominent in all regions and across the entire rural-to-metro continuum.

The rising prevalence of these highly diverse places is broadly consistent with spatial assimilation in that multiple groups are sharing jurisdictional realms. We also find that older no-majority cohorts tend to remain quite diverse over time, even when they move toward group-majority structures. However, the emergence of no-majority communities has been accompanied by large White population declines, a trend more in keeping with ethnic stratification. No-majority places are more likely than other communities to have an aging housing stock, a slack housing market, and low levels of college enrollment. Additionally, high levels of racial income inequality appear to have become a precursor of emerging no-majority status. These results run contrary to our idea that economic and residential opportunities serve as magnets for diversity. That said, job opportunities in construction and transportation do predict the emergence of no-majority places in recent decades. Immigration has played a major role as well. Concentrated foreign-born populations are characteristic of no-majority places, facilitate their emergence, and in some cases help to propel shifts from no-majority to (still diverse) Latino-majority status. Latino and White population dynamics, usually moving in opposite directions, often portend the emergence of a no-majority community.

We should emphasize that the largest cohort of no-majority places is the most recent one, comprised of 402 communities and over 17 million people. It includes large cities like Austin, Charlotte, Las Vegas, and St. Louis, along with rural communities like Cameron, TX and Marksville, LA. Many of these newly no-majority communities have White pluralities complemented by substantial Hispanic, Black, and Asian population shares. They have typically experienced White population losses (except in the West), although at a more modest rate than those in previous no-majority cohorts. In many respects, the 2010 cohort is a herald of the nation’s transition to no-majority status. How it fares in the coming decades will help determine whether localized diversity becomes a common experience across ethnoracial groups or whether jurisdictional boundaries will serve as the new color lines in a racially fragmented society. Either way, the social, political, and economic consequences of place-level diversity trends promise to be profound.

At present, several more places appear poised to join the no-majority ranks. One of these, Dayton, Ohio—currently 51% White—has proactively embraced the growth of its foreign-born population as a way to counter demographic decline and to invigorate its economy. In 2011 the City of Dayton initiated the Welcome Dayton plan to promote immigrant integration through linguistic outreach and engagement with immigrant communities, local businesses, law enforcement, and service providers. A recent Partnership for a New American Economy (PNAE) report provides some evidence that these efforts have paid off in demographic and economic terms, as Dayton’s foreign-born workforce increased by 23% between 2007 and 2012. Elsewhere in Ohio, current and former no-majority cities like Cincinnati (2010 cohort) and Cleveland (1990 cohort) would have experienced population gains rather than declines during this period if they had attracted immigrants at the same rate as Dayton (PNAE 2016). PNAE estimates that Dayton’s growing immigrant population wields $115 million in spending power, has contributed more than $15 million in local and state taxes, and has added more than $116 million in value to the housing stock in Dayton and surrounding Montgomery county (PNAE 2016).

Dayton’s experience raises the question of how municipal governments will navigate the nation’s “diversity explosion” (Frey 2015). On the one hand, localities may attempt to resist rapid changes through density zoning (Rothwell and Massey 2009) or by imposing a range of restrictive ordinances—often ostensibly aimed at undocumented immigrants—in housing, education, and policing (Singer et al. 2008). On the other hand, there is a growing list of successful and ongoing efforts to engender stable neighborhood and community diversity (Maly 2005; Nyden et al. 1997; White House Task Force on New Americans 2016). Our results indicate that local compositional change differs greatly by region and community size and thus municipal responses will likely differ as well. Trajectories of diversification will also depend in part on race-specific population dynamics that we have not directly addressed. For example, White population declines in diversifying areas could be due to natural decrease, net out-migration, or some combination of the two. These distinctions are important since the dynamics driving White decline (e.g., aging vs. flight) alongside nonwhite (often Latino) growth may affect whether the residents and local governments of no-majority places are more likely to “hunker down” (Putnam 2007) or, instead, to become “self-consciously diverse” communities (Nyden et al. 1997).

Author Biography

Chad R. Farrell is a professor of sociology at the University of Alaska Anchorage. His primary research interests revolve around urban inequality with foci on racial and ethnic segregation, diversity, and community change.

Barrett A. Lee, a professor of sociology and demography at Pennsylvania State University, is currently studying ethnoracial diversity, segregation, neighborhood change, residential displacement, and community attachment. He also has a career-long interest in urban homelessness.